LDDMM: how To Bring Deformations from Subject-Specific Atlases Into the Tangent Space at the Global Atlas#
Load meshes:
Use DenseMaternalMeshLoader to load left hippocampus structure for all three subjects
Create subject-specific pre-pregnancy atlases:
For each subject i:
Get pre-pregnancy meshes of left hippocampus
Compute subject-specific pre-pregnancy atlas \(X_{i, \text{pre-pregnancy}}\).
Create global pre-pregnancy atlas:
Compute global pre-pregnancy atlas from subject-specific pre-pregnancy atlases: \(X^{\text{pre-pregnancy}} = \text{mean}_i ~~X^{\text{pre-pregnancy}}_{i}\).
Extract pregnancy deformations as momenta:
For each subject i and pregnancy timepoint t:
Compute deformation from pre-pregnancy atlas to each pregnancy mesh, using control points of global atlas: \(\text{deformation}_{it} = X^{\text{pre-pregnancy}}_{i} \rightarrow X^{\text{pregnancy}}_{it}\) as momenta at control points
Parallel transport deformation into global deterministic atlas space to get Xit: this is assumed to be the identity
Build prediction model:
Use linear regression to predict gestational week (Yit=t) from transported deformations (Xit)
[1]:
import shutil
from pathlib import Path
import numpy as np
import pandas as pd
import pyvista as pv
import polpo.lddmm as plddmm
import polpo.preprocessing.dict as ppdict
import polpo.preprocessing.pd as ppd
import polpo.utils as putils
from polpo.plot.pyvista import RegisteredMeshesGifPlotter
from polpo.preprocessing.load.deformetrica import LoadMeshFlow
from polpo.preprocessing.load.pregnancy.jacobs import MeshLoader, TabularDataLoader
from polpo.preprocessing.mesh.decimation import PvDecimate
from polpo.preprocessing.mesh.io import DictMeshWriter, PvReader
from polpo.preprocessing.mesh.registration import RigidAlignment
from polpo.preprocessing.str import TryToInt
[2]:
DEBUG = False
RECOMPUTE = True
STATIC_VIZ = True
if STATIC_VIZ:
pv.set_jupyter_backend("static")
[3]:
SUBJECT_IDS = ["01", "1001B", "1004B"]#, "1009B"]#, "2004B"]
RESULTS_DIR = Path("results") / "regression_example"
ATLAS_DIR = RESULTS_DIR / "atlases"
MESHES_DIR = RESULTS_DIR / "meshes"
REGISTRATIONS_DIR = RESULTS_DIR / "registrations"
TRANSPORTS_DIR = RESULTS_DIR / "transports"
PARALLEL_SHOOT_DIR = RESULTS_DIR / "parallel_shoot"
if RESULTS_DIR.exists() and RECOMPUTE:
shutil.rmtree(RESULTS_DIR)
data_dir = "/home/data/maternal" if putils.in_frank() else "~/.herbrain/data/maternal"
[4]:
def _sub_id2name(sub_id):
return f"sub_{sub_id}"
Load, write and filter meshes#
[5]:
path2meshes = MeshLoader(
subject_subset=SUBJECT_IDS,
struct_subset=["L_Hipp"],
derivative="fsl",
as_mesh=True,
data_dir=data_dir,
) + ppdict.ExtractUniqueKey(nested=True)
if DEBUG:
path2meshes += ppdict.DictMap(ppdict.TruncateDict(n_keys=8))
# subject_id, session_id
raw_meshes = path2meshes()
print("Number of meshes for:")
for subject_id, meshes in raw_meshes.items():
print(f" -subject {subject_id}: {len(meshes)}")
Number of meshes for:
-subject 01: 27
-subject 1001B: 14
-subject 1004B: 22
[6]:
csvs = {}
for subject_id in SUBJECT_IDS:
csvs[subject_id] = TabularDataLoader(data_dir=data_dir, subject_id=subject_id)()
for subject_id in SUBJECT_IDS:
print(subject_id)
display(csvs[subject_id])
01
| estro | prog | lh | gestWeek | stage | EndoStatus | trimester | |
|---|---|---|---|---|---|---|---|
| sessionID | |||||||
| 1 | NaN | NaN | NaN | -3.0 | pre | pilot1 | pre |
| 2 | 3.42 | 0.840 | NaN | -0.5 | pre | pilot2 | pre |
| 3 | 386.00 | NaN | NaN | 1.0 | pre | IVF | pre |
| 4 | 1238.00 | NaN | NaN | 1.5 | pre | IVF | pre |
| 5 | 1350.00 | 2.940 | NaN | 2.0 | pre | IVF | first |
| 6 | 241.00 | 8.760 | NaN | 3.0 | preg | Pregnant | first |
| 7 | NaN | NaN | NaN | 9.0 | preg | Pregnant | first |
| 8 | NaN | NaN | NaN | 12.0 | preg | Pregnant | first |
| 9 | NaN | NaN | NaN | 14.0 | preg | Pregnant | second |
| 10 | 4700.00 | 53.900 | 1.45 | 15.0 | preg | Pregnant | second |
| 11 | 4100.00 | 56.800 | 0.87 | 17.0 | preg | Pregnant | second |
| 12 | 6190.00 | 70.600 | 0.93 | 19.0 | preg | Pregnant | second |
| 13 | 9640.00 | 54.700 | 0.62 | 22.0 | preg | Pregnant | second |
| 14 | 8800.00 | 64.100 | 0.73 | 24.0 | preg | Pregnant | second |
| 15 | 8970.00 | 61.400 | 0.73 | 27.0 | preg | Pregnant | third |
| 16 | 10200.00 | 74.200 | 0.69 | 29.0 | preg | Pregnant | third |
| 17 | 9920.00 | 83.000 | 0.77 | 31.0 | preg | Pregnant | third |
| 18 | 9860.00 | 95.300 | 0.83 | 33.0 | preg | Pregnant | third |
| 19 | 12400.00 | 103.000 | 0.59 | 36.0 | preg | Pregnant | third |
| 20 | 9.18 | 0.120 | 0.96 | 43.0 | post | Postpartum | post |
| 21 | 20.70 | 0.043 | 4.01 | 46.0 | post | Postpartum | post |
| 22 | 17.50 | 0.068 | 7.58 | 49.0 | post | Postpartum | post |
| 23 | 11.50 | 0.042 | 4.67 | 51.0 | post | Postpartum | post |
| 24 | NaN | NaN | NaN | 68.0 | post | Postpartum | post |
| 25 | NaN | NaN | NaN | 93.0 | post | Postpartum | post |
| 26 | NaN | NaN | NaN | 162.0 | post | Postpartum | post |
| 27 | NaN | NaN | NaN | 162.0 | post | Postpartum | post |
1001B
| gestWeek | stage | trimester | group | sex | age_at_conception | |
|---|---|---|---|---|---|---|
| sessionID | ||||||
| base1 | -18.0 | pre | pre | pregnant | F | 34.0 |
| base2 | -13.0 | pre | pre | NaN | NaN | NaN |
| gest1 | 8.0 | preg | first | NaN | NaN | NaN |
| gest2 | 13.0 | preg | first | NaN | NaN | NaN |
| gest3 | 17.0 | preg | second | NaN | NaN | NaN |
| gest4 | 22.0 | preg | second | NaN | NaN | NaN |
| gest5 | 26.0 | preg | second | NaN | NaN | NaN |
| gest6 | 30.0 | preg | third | NaN | NaN | NaN |
| gest7 | 35.0 | preg | third | NaN | NaN | NaN |
| post1 | 43.0 | post | post | NaN | NaN | NaN |
| post2 | 48.0 | post | post | NaN | NaN | NaN |
| post3 | 56.0 | post | post | NaN | NaN | NaN |
| post4 | 65.0 | post | post | NaN | NaN | NaN |
| post5 | 93.0 | post | post | NaN | NaN | NaN |
1004B
| gestWeek | stage | trimester | group | sex | age_at_conception | |
|---|---|---|---|---|---|---|
| sessionID | ||||||
| base1 | -2.5 | pre | pre | pregnant | F | 26.0 |
| base2 | 2.0 | preg | first | NaN | NaN | NaN |
| gest1 | 5.0 | preg | first | NaN | NaN | NaN |
| gest2 | 7.0 | preg | first | NaN | NaN | NaN |
| gest3 | 8.5 | preg | first | NaN | NaN | NaN |
| gest4 | 10.5 | preg | first | NaN | NaN | NaN |
| gest5 | 12.5 | preg | first | NaN | NaN | NaN |
| gest6 | 14.5 | preg | second | NaN | NaN | NaN |
| gest7 | 16.5 | preg | second | NaN | NaN | NaN |
| gest8 | 19.0 | preg | second | NaN | NaN | NaN |
| gest9 | 21.0 | preg | second | NaN | NaN | NaN |
| gest10 | 23.0 | preg | second | NaN | NaN | NaN |
| gest11 | 24.5 | preg | second | NaN | NaN | NaN |
| gest12 | 27.0 | preg | third | NaN | NaN | NaN |
| gest13 | 30.0 | preg | third | NaN | NaN | NaN |
| post1 | 43.0 | post | post | NaN | NaN | NaN |
| post2 | 50.0 | post | post | NaN | NaN | NaN |
| post3 | 52.0 | post | post | NaN | NaN | NaN |
| post4 | 54.0 | post | post | NaN | NaN | NaN |
| post5 | 57.0 | post | post | NaN | NaN | NaN |
| post6 | 65.5 | post | post | NaN | NaN | NaN |
| post7 | 93.0 | post | post | NaN | NaN | NaN |
[7]:
# Rigid alignment of all meshes to the first mesh of subject 01
align_pipe = RigidAlignment(
target=putils.get_first(raw_meshes["01"]),
# TODO: update here
max_iterations=10,
)
# Coarsen all meshes to reduce number of vertices
# NB: this steps kills any correspondence between meshes
# TODO: evaluate relevance of this step
decimate_pipe = PvDecimate(target_reduction=0.6)
# TODO: it is better to go the other way around
rigidly_aligned_meshes = ppdict.DictMap(align_pipe)(raw_meshes)
meshes = ppdict.NestedDictMap(decimate_pipe)(rigidly_aligned_meshes)
[8]:
pl = pv.Plotter(shape=(2, 3), border=False, window_size=[600, 300])
for i, subject_id in enumerate(SUBJECT_IDS):
pl.subplot(0, i)
pl.add_mesh(putils.get_first(rigidly_aligned_meshes[subject_id]), show_edges=True)
pl.add_text(
f"First mesh of subject {subject_id}", position="upper_edge", font_size=4
)
pl.subplot(1, i)
pl.add_mesh(putils.get_first(meshes[subject_id]), show_edges=True, color="white")
pl.add_text(
f"Decimated First Mesh of subject {subject_id}",
position="upper_edge",
font_size=4,
)
pl.show()
2025-10-29 09:40:36.802 ( 4.280s) [ 7F8B78B22440]vtkXOpenGLRenderWindow.:1458 WARN| bad X server connection. DISPLAY=
[9]:
def _key2name(key):
sub_id, session_id = key.split("/")
return f"{_sub_id2name(sub_id)}/mesh_{session_id}"
meshes2files = (
ppdict.UnnestDict(sep="/")
+ DictMeshWriter(dirname=MESHES_DIR, ext="vtk", key2name=_key2name)
+ ppdict.NestDict(sep="/")
+ ppdict.DictMap(ppdict.DictMap(key_step=TryToInt()))
)
# subject, session: filename
dataset = meshes2files(meshes)
Estimate deterministic atlas for subject and global#
Compute subject specific deterministic atlas by averaging pre-pregnancy meshes.
First, we select the pre-pregnancy meshes.
[10]:
# TODO: update here
pre_keys_selector = ppd.DfIsInFilter("stage", ["pre"]) + (lambda x: x.index)
pre_keys = ppdict.DictMap(pre_keys_selector)(csvs)
[11]:
pre_dataset = {
subject_id: ppdict.SelectKeySubset(pre_keys[subject_id])(meshes)
for subject_id, meshes in dataset.items()
}
Closely follows LDDMM: how to estimate a deterministic atlas?.
[12]:
kernel_width = 10.0 # NB: influences the number of cp
registration_kwargs = dict(
kernel_width=kernel_width,
regularisation=1,
max_iter=2000,
freeze_control_points=False,
metric="varifold",
attachment_kernel_width=2.0,
tol=1e-10,
)
[13]:
templates = {}
for subject_id, dataset_ in pre_dataset.items():
output_dir = ATLAS_DIR / _sub_id2name(subject_id)
if output_dir.exists():
templates[subject_id] = plddmm.io.load_template(output_dir, as_path=True)
continue
templates[subject_id] = plddmm.learning.estimate_deterministic_atlas(
targets=dataset_,
output_dir=output_dir,
initial_step_size=1e-1,
**registration_kwargs,
)
output_dir = ATLAS_DIR / "global"
if not output_dir.exists():
templates["global"] = plddmm.learning.estimate_deterministic_atlas(
targets=templates,
output_dir=output_dir,
initial_step_size=1e-1,
**registration_kwargs,
)
else:
templates["global"] = plddmm.io.load_template(output_dir, as_path=True)
templates = ppdict.DictMap(PvReader())(templates)
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 10.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/atlases/sub_01/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
{'max_iterations': 2000, 'freeze_template': False, 'freeze_control_points': False, 'freeze_momenta': False, 'use_sobolev_gradient': True, 'sobolev_kernel_width_ratio': 1, 'max_line_search_iterations': 50, 'initial_control_points': None, 'initial_cp_spacing': None, 'initial_momenta': None, 'dense_mode': False, 'number_of_threads': 1, 'print_every_n_iters': 20, 'downsampling_factor': 1, 'dimension': 3, 'optimization_method_type': 'ScipyLBFGS', 'convergence_tolerance': 1e-10, 'initial_step_size': 0.1, 'gpu_mode': <GpuMode.KERNEL: 4>, 'state_file': 'results/regression_example/atlases/sub_01/deformetrica-state.p', 'load_state_file': False, 'memory_length': 10}
instantiating kernel torch with kernel_width 10.0 and gpu_mode GpuMode.KERNEL. addr: 0x7f88d117be00
instantiating kernel torch with kernel_width 2.0 and gpu_mode GpuMode.KERNEL. addr: 0x7f88d10b15b0
>> Set of 24 control points defined.
>> Momenta initialized to zero, for 5 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
------------------------------------- Iteration: 20 -------------------------------------
>> Log-likelihood = -3.250E+03 [ attachment = -3.243E+03 ; regularity = -6.885E+00 ]
>> Gradient at Termination: 945186.1524416754
>> CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH
>> Estimation took: 10 seconds
Deformetrica.__del__()
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 10.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/atlases/sub_1001B/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
{'max_iterations': 2000, 'freeze_template': False, 'freeze_control_points': False, 'freeze_momenta': False, 'use_sobolev_gradient': True, 'sobolev_kernel_width_ratio': 1, 'max_line_search_iterations': 50, 'initial_control_points': None, 'initial_cp_spacing': None, 'initial_momenta': None, 'dense_mode': False, 'number_of_threads': 1, 'print_every_n_iters': 20, 'downsampling_factor': 1, 'dimension': 3, 'optimization_method_type': 'ScipyLBFGS', 'convergence_tolerance': 1e-10, 'initial_step_size': 0.1, 'gpu_mode': <GpuMode.KERNEL: 4>, 'state_file': 'results/regression_example/atlases/sub_1001B/deformetrica-state.p', 'load_state_file': False, 'memory_length': 10}
>> Set of 24 control points defined.
>> Momenta initialized to zero, for 2 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
------------------------------------- Iteration: 20 -------------------------------------
>> Log-likelihood = -7.343E+02 [ attachment = -7.310E+02 ; regularity = -3.316E+00 ]
------------------------------------- Iteration: 40 -------------------------------------
>> Log-likelihood = -6.826E+02 [ attachment = -6.776E+02 ; regularity = -4.912E+00 ]
>> Gradient at Termination: 1497206.5969691253
>> CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH
>> Estimation took: 11 seconds
Deformetrica.__del__()
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 10.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/atlases/sub_1004B/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
{'max_iterations': 2000, 'freeze_template': False, 'freeze_control_points': False, 'freeze_momenta': False, 'use_sobolev_gradient': True, 'sobolev_kernel_width_ratio': 1, 'max_line_search_iterations': 50, 'initial_control_points': None, 'initial_cp_spacing': None, 'initial_momenta': None, 'dense_mode': False, 'number_of_threads': 1, 'print_every_n_iters': 20, 'downsampling_factor': 1, 'dimension': 3, 'optimization_method_type': 'ScipyLBFGS', 'convergence_tolerance': 1e-10, 'initial_step_size': 0.1, 'gpu_mode': <GpuMode.KERNEL: 4>, 'state_file': 'results/regression_example/atlases/sub_1004B/deformetrica-state.p', 'load_state_file': False, 'memory_length': 10}
>> Set of 20 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 5.456944161693467
>> CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH
>> Estimation took: 01 seconds
Deformetrica.__del__()
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 10.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/atlases/global/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
{'max_iterations': 2000, 'freeze_template': False, 'freeze_control_points': False, 'freeze_momenta': False, 'use_sobolev_gradient': True, 'sobolev_kernel_width_ratio': 1, 'max_line_search_iterations': 50, 'initial_control_points': None, 'initial_cp_spacing': None, 'initial_momenta': None, 'dense_mode': False, 'number_of_threads': 1, 'print_every_n_iters': 20, 'downsampling_factor': 1, 'dimension': 3, 'optimization_method_type': 'ScipyLBFGS', 'convergence_tolerance': 1e-10, 'initial_step_size': 0.1, 'gpu_mode': <GpuMode.KERNEL: 4>, 'state_file': 'results/regression_example/atlases/global/deformetrica-state.p', 'load_state_file': False, 'memory_length': 10}
>> Set of 24 control points defined.
>> Momenta initialized to zero, for 3 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
------------------------------------- Iteration: 20 -------------------------------------
>> Log-likelihood = -3.883E+03 [ attachment = -3.803E+03 ; regularity = -8.012E+01 ]
>> Gradient at Termination: 713805.435846427
>> CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH
>> Estimation took: 07 seconds
Deformetrica.__del__()
[14]:
pl = pv.Plotter(shape=(1, 4), border=False, window_size=[800, 150])
for i, subject_id in enumerate(SUBJECT_IDS):
pl.subplot(0, i)
pl.add_mesh(templates[subject_id], show_edges=True)
pl.add_text(
f"pre-pregnancy template of {subject_id}", position="upper_edge", font_size=4
)
pl.subplot(0, 3)
pl.add_mesh(templates["global"], show_edges=True)
pl.add_text("global template", position="upper_edge", font_size=4)
pl.show()
Parallel transport to atlas#
Closely follows LDDMM: parallel transport?.
[15]:
global_template = plddmm.io.load_template(ATLAS_DIR / "global", as_path=True)
for subject_id, meshes in dataset.items():
subject_template = plddmm.io.load_template(
ATLAS_DIR / f"sub_{subject_id}", as_path=True
)
transport_dir = TRANSPORTS_DIR / _sub_id2name(subject_id)
registrations_dir = REGISTRATIONS_DIR / _sub_id2name(subject_id)
parallel_shoot_dir = PARALLEL_SHOOT_DIR / _sub_id2name(subject_id)
for session_id, mesh in meshes.items():
transport_dir_sess = transport_dir / plddmm.io.build_parallel_transport_name(
subject_id, "global", session_id
)
if transport_dir_sess.exists():
continue
meshes = {
"global": global_template,
subject_id: subject_template,
session_id: mesh,
}
plddmm.geometry.parallel_transport_ABC(
meshes,
output_dir=transport_dir,
registration_dir=registrations_dir,
compute_shoot=True,
shoot_dir=parallel_shoot_dir,
use_pole_ladder=True,
**registration_kwargs,
)
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->global/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
instantiating kernel torch with kernel_width 20.0 and gpu_mode GpuMode.KERNEL. addr: 0x7f88d14ca5a0
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 2524.9788413671886
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->1/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 2611.255439229994
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->2/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/in_out/dataset_functions.py:265: UserWarning: Watch out, I did not get a distance type for the object shape, Please make sure you are running shooting or a parallel transport, otherwise distances are required.
warnings.warn(msg)
>> Gradient at Termination: 586.7095643285912
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->3/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 741.8078298634514
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->4/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 578.980751037355
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->5/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 4388.264240921341
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->6/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 3840.8066040628923
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->7/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 372.15071554513673
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->8/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 1267.3577594848834
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->9/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 2628.533763673712
>> ABNORMAL:
>> Estimation took: 03 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->10/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
------------------------------------- Iteration: 20 -------------------------------------
>> Log-likelihood = -1.154E+03 [ attachment = -1.153E+03 ; regularity = -1.583E+00 ]
>> Log-likelihood = -1.155E+03 [ attachment = -1.153E+03 ; regularity = -1.603E+00 ]
>> Log-likelihood = -1.155E+03 [ attachment = -1.153E+03 ; regularity = -1.608E+00 ]
>> Log-likelihood = -1.154E+03 [ attachment = -1.152E+03 ; regularity = -1.608E+00 ]
>> Log-likelihood = -1.154E+03 [ attachment = -1.152E+03 ; regularity = -1.608E+00 ]
>> Log-likelihood = -1.154E+03 [ attachment = -1.152E+03 ; regularity = -1.608E+00 ]
>> Log-likelihood = -1.154E+03 [ attachment = -1.152E+03 ; regularity = -1.608E+00 ]
>> Log-likelihood = -1.154E+03 [ attachment = -1.152E+03 ; regularity = -1.608E+00 ]
>> Log-likelihood = -1.154E+03 [ attachment = -1.152E+03 ; regularity = -1.608E+00 ]
>> Log-likelihood = -1.154E+03 [ attachment = -1.152E+03 ; regularity = -1.608E+00 ]
>> Log-likelihood = -1.974E+05 [ attachment = -1.912E+05 ; regularity = -6.272E+03 ]
>> Log-likelihood = -2.017E+03 [ attachment = -2.015E+03 ; regularity = -2.329E+00 ]
>> Log-likelihood = -1.154E+03 [ attachment = -1.152E+03 ; regularity = -1.610E+00 ]
>> Gradient at Termination: 1143.2033609656164
>> ABNORMAL:
>> Estimation took: 03 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->11/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 704.1044122274275
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->12/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 1187.7760353162182
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->13/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 528.6977062126508
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->14/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 7212.225441086446
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->15/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 457.9186057456068
>> CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH
>> Estimation took: 00 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->16/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 186.9579841697961
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->17/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 544.5751335036966
>> CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH
>> Estimation took: 00 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->18/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 1795.162665559362
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->19/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 1587.3162233801522
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->20/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 4497.185742595348
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->21/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 947.2776748829019
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->22/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 1076.386248107878
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->23/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 523.5326604226665
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->24/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 496.46610662568287
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->25/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 48.06148168368736
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->26/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 225.85338479576814
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_01/01->27/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 5118.26557854517
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->global/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
------------------------------------- Iteration: 20 -------------------------------------
>> Log-likelihood = -2.652E+03 [ attachment = -2.638E+03 ; regularity = -1.345E+01 ]
------------------------------------- Iteration: 40 -------------------------------------
>> Log-likelihood = -2.228E+03 [ attachment = -2.192E+03 ; regularity = -3.606E+01 ]
------------------------------------- Iteration: 60 -------------------------------------
>> Log-likelihood = -2.078E+03 [ attachment = -2.048E+03 ; regularity = -2.970E+01 ]
>> Log-likelihood = -2.061E+03 [ attachment = -2.031E+03 ; regularity = -2.961E+01 ]
>> Gradient at Termination: 2415.7619343136867
>> ABNORMAL:
>> Estimation took: 04 seconds
Deformetrica.__del__()
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->base1/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 3781.6545251118077
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->base2/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 2724.7186639541615
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->gest1/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 1416.7969624118982
>> CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->gest2/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 3501.334625308142
>> ABNORMAL:
>> Estimation took: 03 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->gest3/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 1104.9754617530552
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->gest4/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 2290.249468834107
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->gest5/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 720.5682179038324
>> CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->gest6/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 894.7903269439973
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->gest7/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 2360.9127869479375
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->post1/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 3076.771100177274
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->post2/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
------------------------------------- Iteration: 20 -------------------------------------
>> Log-likelihood = -1.005E+03 [ attachment = -1.003E+03 ; regularity = -2.508E+00 ]
>> Gradient at Termination: 1046.2186813542578
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->post3/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 2598.4134294773376
>> CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH
>> Estimation took: 00 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->post4/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 277.693749893504
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1001B/1001B->post5/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 4 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
------------------------------------- Iteration: 20 -------------------------------------
>> Log-likelihood = -1.040E+03 [ attachment = -1.038E+03 ; regularity = -2.436E+00 ]
>> Gradient at Termination: 2067.2767049664726
>> ABNORMAL:
>> Estimation took: 03 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->global/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
------------------------------------- Iteration: 20 -------------------------------------
>> Log-likelihood = -4.700E+03 [ attachment = -4.660E+03 ; regularity = -4.006E+01 ]
------------------------------------- Iteration: 40 -------------------------------------
>> Log-likelihood = -3.829E+03 [ attachment = -3.702E+03 ; regularity = -1.272E+02 ]
------------------------------------- Iteration: 60 -------------------------------------
>> Log-likelihood = -3.417E+03 [ attachment = -3.362E+03 ; regularity = -5.502E+01 ]
>> Log-likelihood = -3.368E+03 [ attachment = -3.310E+03 ; regularity = -5.812E+01 ]
------------------------------------- Iteration: 80 -------------------------------------
>> Log-likelihood = -3.252E+03 [ attachment = -3.192E+03 ; regularity = -6.033E+01 ]
------------------------------------- Iteration: 100 -------------------------------------
>> Log-likelihood = -3.200E+03 [ attachment = -3.121E+03 ; regularity = -7.843E+01 ]
------------------------------------- Iteration: 120 -------------------------------------
>> Log-likelihood = -3.181E+03 [ attachment = -3.097E+03 ; regularity = -8.403E+01 ]
>> Gradient at Termination: 4096.0738155203535
>> ABNORMAL:
>> Estimation took: 07 seconds
Deformetrica.__del__()
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->base1/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 0.12077687381657846
>> ABNORMAL:
>> Estimation took: 00 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->base2/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 86.64903551017103
>> ABNORMAL:
>> Estimation took: 03 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->gest1/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
------------------------------------- Iteration: 20 -------------------------------------
>> Log-likelihood = -1.153E+03 [ attachment = -1.150E+03 ; regularity = -3.684E+00 ]
>> Gradient at Termination: 298.1602912576601
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->gest2/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 977.2991716241706
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->gest3/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 176.87799272520468
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->gest4/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 1675.4409376316264
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->gest5/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 769.3145985685601
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->gest6/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 462.6108938621614
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->gest7/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 2957.0099796598206
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->gest8/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 277.9189697015974
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->gest9/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 712.9032122409644
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->gest10/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 170.30216691164148
>> CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH
>> Estimation took: 00 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->gest11/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 967.5064616308571
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->gest12/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
------------------------------------- Iteration: 20 -------------------------------------
>> Log-likelihood = -1.724E+03 [ attachment = -1.720E+03 ; regularity = -3.340E+00 ]
>> Gradient at Termination: 547.9788517782683
>> CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->gest13/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 196.4641346278797
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->post1/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
------------------------------------- Iteration: 20 -------------------------------------
>> Log-likelihood = -6.814E+02 [ attachment = -6.793E+02 ; regularity = -2.076E+00 ]
>> Gradient at Termination: 773.4194820145234
>> ABNORMAL:
>> Estimation took: 03 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->post2/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
------------------------------------- Iteration: 20 -------------------------------------
>> Log-likelihood = -1.472E+03 [ attachment = -1.470E+03 ; regularity = -2.114E+00 ]
>> Gradient at Termination: 340.665882254072
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->post3/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 714.79655331099
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->post4/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 727.4278418967888
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->post5/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 1123.9839169489226
>> ABNORMAL:
>> Estimation took: 01 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->post6/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
------------------------------------- Iteration: 20 -------------------------------------
>> Log-likelihood = -1.491E+03 [ attachment = -1.482E+03 ; regularity = -8.886E+00 ]
------------------------------------- Iteration: 40 -------------------------------------
>> Log-likelihood = -1.149E+03 [ attachment = -1.131E+03 ; regularity = -1.801E+01 ]
>> Gradient at Termination: 105.64815598517207
>> ABNORMAL:
>> Estimation took: 03 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Logger has been set to: DEBUG
>> No initial CP spacing given: using diffeo kernel width of 20.0
OMP_NUM_THREADS was not found in environment variables. An automatic value will be set.
OMP_NUM_THREADS will be set to 128
context has already been set
>> No specified state-file. By default, Deformetrica state will by saved in file: results/regression_example/registrations/sub_1004B/1004B->post7/deformetrica-state.p.
>> Using a Sobolev gradient for the template data with the ScipyLBFGS estimator memory length being larger than 1. Beware: that can be tricky.
>> Set of 3 control points defined.
>> Momenta initialized to zero, for 1 subjects.
dtype=float32
>> Started estimator: ScipyOptimize
>> Scipy optimization method: L-BFGS-B
------------------------------------- Iteration: 1 -------------------------------------
>> Gradient at Termination: 602.6440333268713
>> ABNORMAL:
>> Estimation took: 02 seconds
Deformetrica.__del__()
source is ignored when pole ladder is used
In exponential update, I am not flowing because I don't have any template points to flow
In exponential update, I am not flowing because I don't have any template points to flow
[ compute_shooting function ]
Defaulting geodesic t0 to 1.
Defaulting geodesic tmax to 1.
Defaulting geodesic tmin to 0.
Visualize shooting from atlases to subject last mesh#
From subject template.
[16]:
meshes = {}
for subject_id, sessions in dataset.items():
registrations_dir = (
REGISTRATIONS_DIR
/ _sub_id2name(subject_id)
/ plddmm.io.build_registration_name(
subject_id,
list(sessions.keys())[-1],
)
)
print(registrations_dir)
meshes[subject_id] = plddmm.io.load_deterministic_atlas_flow(
registrations_dir, as_pv=True
)
results/regression_example/registrations/sub_01/01->27
results/regression_example/registrations/sub_1001B/1001B->post5
results/regression_example/registrations/sub_1004B/1004B->post7
[17]:
pl = RegisteredMeshesGifPlotter(
subtitle=lambda time, subj: f"{subj}, {time}",
shape=(2, 2),
fps=5,
)
pl.add_meshes(ppdict.DictListSwapper()(meshes))
pl.close()
pl.show()
[17]:
From global template.
[18]:
meshes = {}
for subject_id, sessions in dataset.items():
parallel_shoot_dir = (
PARALLEL_SHOOT_DIR
/ _sub_id2name(subject_id)
/ plddmm.io.build_parallel_shoot_name(
subject_id,
"global",
list(sessions.keys())[-1],
)
)
print(parallel_shoot_dir)
# TODO: clean io
meshes[subject_id] = LoadMeshFlow(as_path=False)(parallel_shoot_dir)
results/regression_example/parallel_shoot/sub_01/global(t01>27)
results/regression_example/parallel_shoot/sub_1001B/global(t1001B>post5)
results/regression_example/parallel_shoot/sub_1004B/global(t1004B>post7)
[19]:
pl = RegisteredMeshesGifPlotter(
subtitle=lambda time, subj: f"{subj}, {time}",
shape=(2, 2),
fps=5,
)
pl.add_meshes(ppdict.NestedDictSwapper()(meshes))
pl.close()
pl.show()
[19]:
[20]:
last_meshes = [list(meshes.values())[-1] for meshes in meshes.values()]
(
np.amax(np.abs(last_meshes[0].points - last_meshes[1].points)),
np.amax(np.abs(last_meshes[0].points - last_meshes[2].points)),
)
[20]:
(np.float64(1.0626754760742188), np.float64(1.0740966796875))
Compute velocities#
[21]:
plddmm.io.load_cp(ATLAS_DIR / "global", as_path=True)
[21]:
PosixPath('results/regression_example/atlases/global/DeterministicAtlas__EstimatedParameters__ControlPoints.txt')
[22]:
cp = {}
momenta = {}
vel = {}
atlas_cp = plddmm.io.load_cp(ATLAS_DIR / "global", as_path=False)
for subject_id, meshes in dataset.items():
transport_dir = TRANSPORTS_DIR / _sub_id2name(subject_id)
cp_subj = cp[subject_id] = {}
momenta_subj = momenta[subject_id] = {}
vel_subj = vel[subject_id] = {}
for session_id, mesh in meshes.items():
transport_dir_sess = transport_dir / plddmm.io.build_parallel_transport_name(
subject_id, "global", session_id
)
cp_ = cp_subj[session_id] = plddmm.io.load_transported_cp(
transport_dir_sess, as_path=False
)
momenta_ = momenta_subj[session_id] = plddmm.io.load_transported_momenta(
transport_dir_sess, as_path=False
)
vel_subj[session_id] = plddmm.geometry.velocity_at_x(
atlas_cp, cp_, momenta_, kernel_width=kernel_width
)
Velocities are comparable quantities.
[23]:
vel_week = {}
for subject_id, subj_dict in vel.items():
session2week = csvs[subject_id]["gestWeek"]
vel_week[subject_id] = ppdict.DictMap(
key_step=lambda session: float(session2week[session])
)(vel[subject_id])
vel_ls = []
for subj_id, subj_vel in vel_week.items():
for week, vels in subj_vel.items():
vel_ls.append(
{
"subject": subj_id,
"week": float(week),
"vel": vels,
}
)
[24]:
df = pd.DataFrame.from_dict(vel_ls)
df.to_csv(RESULTS_DIR / "velocities.csv", index=False)
[25]:
vel_week
[25]:
{'01': {-3.0: array([[ 0.00220195, 0.05142251, -0.04584359],
[ 0.00203441, 0.04553534, -0.0403849 ],
[ 0.01117069, 0.04208901, -0.01480857],
[ 0.01141505, 0.04095323, -0.01308329],
[ 0.02461177, 0.05242203, 0.0074419 ],
[ 0.02654672, 0.05643063, 0.00816343],
[-0.00952978, 0.04259073, -0.04397725],
[-0.00891005, 0.03932984, -0.04048735],
[ 0.00771329, 0.03920534, -0.01333026],
[ 0.0095078 , 0.04198331, -0.01236233],
[ 0.0284091 , 0.05524097, 0.00706629],
[ 0.03281986, 0.06477229, 0.00868083],
[-0.07927734, -0.06910087, 0.02113093],
[-0.07929646, -0.06957772, 0.02156907],
[-0.02039247, -0.01452065, 0.0096149 ],
[-0.01966953, -0.01373022, 0.00965709],
[ 0.03515238, 0.0403994 , 0.00036235],
[ 0.03712611, 0.04236589, 0.00033238],
[-0.07195229, -0.06915385, 0.02503579],
[-0.07260469, -0.0697768 , 0.0252749 ],
[-0.01597055, -0.01454593, 0.00938113],
[-0.01794774, -0.01642972, 0.01009707],
[ 0.03159 , 0.03254084, -0.00080494],
[ 0.03221098, 0.03318913, -0.00083092]], dtype=float32),
-0.5: array([[ 0.28073624, -0.03771748, -0.11910288],
[ 0.24700834, -0.03267367, -0.10466966],
[ 0.05790733, 0.04720761, -0.01190824],
[ 0.04631573, 0.05051434, -0.00656617],
[-0.12378439, 0.14774281, 0.08263253],
[-0.13438693, 0.15963137, 0.08960738],
[ 0.29125878, -0.04414172, -0.10352459],
[ 0.26826987, -0.04019475, -0.09480918],
[ 0.05455592, 0.0418247 , -0.01022123],
[ 0.04329734, 0.05208888, -0.00346208],
[-0.1292791 , 0.13320395, 0.06011926],
[-0.15347496, 0.16215703, 0.07757822],
[ 0.01822277, -0.03674002, 0.12647147],
[ 0.01552985, -0.03624805, 0.12773278],
[-0.02790752, -0.02971402, 0.01014938],
[-0.02981752, -0.02817792, 0.01024408],
[-0.08570763, -0.03363552, -0.11327022],
[-0.09000213, -0.03766919, -0.12163616],
[-0.01983483, -0.0282509 , 0.12986222],
[-0.02010593, -0.02859365, 0.13095057],
[-0.03281767, -0.03236318, 0.00248251],
[-0.03357714, -0.03328279, 0.00589831],
[-0.06701682, -0.05722631, -0.12735455],
[-0.06826331, -0.05816635, -0.12967601]], dtype=float32),
1.0: array([[-0.07460082, 0.04267115, 0.11502326],
[-0.06583485, 0.03740777, 0.10118715],
[-0.03684809, -0.00640993, 0.02233634],
[-0.03444544, -0.00862657, 0.017527 ],
[-0.01831101, -0.05514305, -0.05398384],
[-0.01955613, -0.05956712, -0.05864105],
[-0.0625016 , 0.06022805, 0.10531081],
[-0.05748354, 0.05565019, 0.09669446],
[-0.03049582, -0.00813216, 0.02042631],
[-0.03117988, -0.01185694, 0.0155712 ],
[-0.01779092, -0.07733084, -0.04053027],
[-0.02097839, -0.08658872, -0.05201016],
[ 0.0954111 , 0.10972526, -0.08650919],
[ 0.09609711, 0.10939999, -0.08768099],
[ 0.03997931, -0.01910293, -0.01180631],
[ 0.03941264, -0.02051633, -0.01193943],
[-0.00483584, -0.16722679, 0.06600493],
[-0.00483168, -0.17758048, 0.0709202 ],
[ 0.09587596, 0.09382733, -0.09316549],
[ 0.09679447, 0.09452913, -0.09396603],
[ 0.0363256 , -0.02375211, -0.00786869],
[ 0.03906321, -0.02148369, -0.0103578 ],
[-0.00197174, -0.16016097, 0.07513513],
[-0.00207357, -0.16311151, 0.07651208]], dtype=float32),
1.5: array([[ 0.01198503, -0.07364011, 0.1715321 ],
[ 0.01013277, -0.06562266, 0.15104966],
[-0.04204061, -0.10517716, 0.04975571],
[-0.04393545, -0.10496316, 0.0430996 ],
[-0.11144935, -0.18219419, -0.04126779],
[-0.12032802, -0.19642787, -0.04510638],
[ 0.02328905, -0.03880111, 0.14797136],
[ 0.02121852, -0.03588759, 0.13582045],
[-0.03947716, -0.09426212, 0.04445742],
[-0.046776 , -0.10548521, 0.03986024],
[-0.11318596, -0.19016603, -0.01851068],
[-0.1341956 , -0.22382155, -0.02837811],
[ 0.07397631, 0.24833524, -0.19007659],
[ 0.07382244, 0.24897796, -0.19181496],
[ 0.00963856, 0.05791827, -0.03070158],
[ 0.00814917, 0.05489119, -0.03015872],
[-0.05745832, -0.12817045, 0.13235651],
[-0.05986387, -0.13423876, 0.14141463],
[ 0.06535767, 0.23407815, -0.19415094],
[ 0.06593464, 0.2362006 , -0.19584042],
[ 0.01010049, 0.05275662, -0.02256042],
[ 0.01184233, 0.05916202, -0.02777428],
[-0.04047197, -0.10069698, 0.14010946],
[-0.04128723, -0.10272542, 0.14271699]], dtype=float32),
2.0: array([[-0.51738787, -0.2135003 , 0.19766286],
[-0.45481804, -0.18770301, 0.17375314],
[-0.06450903, -0.02807002, 0.02398442],
[-0.04174575, -0.01875373, 0.01526866],
[ 0.32859564, 0.13223207, -0.12711814],
[ 0.35635057, 0.14331275, -0.13784164],
[-0.48149318, -0.22834088, 0.18608071],
[-0.4419133 , -0.21029516, 0.17082854],
[-0.0575492 , -0.02716387, 0.02165513],
[-0.03025584, -0.0161023 , 0.01110851],
[ 0.26964584, 0.13866726, -0.10621264],
[ 0.33740216, 0.16413175, -0.1322993 ],
[ 0.336564 , -0.05987957, -0.11428053],
[ 0.34186321, -0.05782326, -0.11629946],
[ 0.04808851, 0.01328211, -0.01664611],
[ 0.04978728, 0.01537152, -0.01740722],
[-0.25116906, 0.09572912, 0.08474475],
[-0.27132857, 0.10052381, 0.09173276],
[ 0.37151664, -0.02653396, -0.12899971],
[ 0.37470207, -0.02669662, -0.13010615],
[ 0.02925353, 0.01817899, -0.01032424],
[ 0.03919264, 0.01763636, -0.01377901],
[-0.3052895 , 0.07580696, 0.10556553],
[-0.31083015, 0.07724248, 0.10747546]], dtype=float32),
3.0: array([[-0.4192432 , -0.05281937, 0.3315045 ],
[-0.36919716, -0.04756229, 0.291863 ],
[-0.1220516 , -0.12915832, 0.0897274 ],
[-0.10584736, -0.1306855 , 0.07667641],
[ 0.09993222, -0.25884607, -0.09512315],
[ 0.10908147, -0.27919182, -0.10371657],
[-0.39815384, 0.00069388, 0.29550153],
[-0.36633602, 0.0005689 , 0.2713986 ],
[-0.10834516, -0.11501554, 0.07861448],
[-0.09750479, -0.13152218, 0.06893698],
[ 0.0908639 , -0.2766847 , -0.06455931],
[ 0.11199543, -0.32431507, -0.08572568],
[ 0.22010036, 0.36996928, -0.30349138],
[ 0.22414269, 0.37041986, -0.30679518],
[ 0.08995918, 0.07247755, -0.06893626],
[ 0.09042945, 0.06790786, -0.06850549],
[-0.02028055, -0.22429943, 0.16198938],
[-0.02248711, -0.23573019, 0.17351109],
[ 0.25324807, 0.34139594, -0.3174008 ],
[ 0.25563273, 0.34444383, -0.32022566],
[ 0.08437625, 0.06239503, -0.05583541],
[ 0.09155554, 0.07164239, -0.06449388],
[-0.03618557, -0.18566494, 0.1784844 ],
[-0.03693981, -0.1893155 , 0.18182886]], dtype=float32),
9.0: array([[-0.44219372, -0.16419871, 0.29967394],
[-0.38918525, -0.14469594, 0.26334012],
[-0.10544662, -0.06015127, 0.02864351],
[-0.08760323, -0.05412227, 0.01514607],
[ 0.16086541, 0.00950157, -0.2109893 ],
[ 0.17498945, 0.01095942, -0.22884044],
[-0.40001127, -0.09890924, 0.2375415 ],
[-0.36737618, -0.08989452, 0.21693775],
[-0.08871734, -0.05429434, 0.02028356],
[-0.07394607, -0.05154561, 0.00337569],
[ 0.13115512, -0.06092623, -0.13773616],
[ 0.16514757, -0.05421489, -0.1832539 ],
[ 0.36568934, 0.4684062 , -0.47209734],
[ 0.37005457, 0.47027045, -0.47533312],
[ 0.10902346, 0.02223282, -0.04928144],
[ 0.10906821, 0.01936918, -0.04849917],
[-0.13003267, -0.45641744, 0.39230704],
[-0.14014085, -0.4863716 , 0.42039755],
[ 0.3879004 , 0.44509783, -0.46591085],
[ 0.39142993, 0.44879717, -0.46985182],
[ 0.09131657, -0.00116504, -0.01926715],
[ 0.1020669 , 0.01041489, -0.03157732],
[-0.15653479, -0.46226248, 0.42921263],
[-0.15949887, -0.47080657, 0.43708548]], dtype=float32),
12.0: array([[-0.20018655, 0.18169689, 0.05883455],
[-0.17595837, 0.15914126, 0.05193511],
[-0.02361939, -0.03987725, 0.03058307],
[-0.01474537, -0.0498516 , 0.02873255],
[ 0.13026504, -0.26450303, 0.01807519],
[ 0.14133263, -0.28612733, 0.01933656],
[-0.16478564, 0.17448282, 0.0489029 ],
[-0.15071103, 0.15962686, 0.04498291],
[-0.0199294 , -0.03828522, 0.0264424 ],
[-0.00913369, -0.0575407 , 0.02710991],
[ 0.08392114, -0.23326634, 0.02040865],
[ 0.11215021, -0.285881 , 0.02344488],
[ 0.27439743, -0.08139507, -0.07781948],
[ 0.27654916, -0.08336677, -0.0783737 ],
[ 0.0203211 , 0.00258479, -0.02498581],
[ 0.01996894, 0.00049891, -0.02447159],
[-0.2494637 , 0.09510469, 0.02315623],
[-0.26729196, 0.10448116, 0.02438855],
[ 0.27420673, -0.09723233, -0.07796554],
[ 0.27650094, -0.09799472, -0.07868867],
[ 0.00395231, 0.01398292, -0.02245284],
[ 0.01114682, 0.01148363, -0.02462958],
[-0.27233735, 0.13763542, 0.02051914],
[-0.277321 , 0.1400381 , 0.02094205]], dtype=float32),
14.0: array([[-0.08051723, -0.0327426 , 0.06664983],
[-0.07109749, -0.02871589, 0.05865931],
[-0.04541199, 0.00262588, 0.01622847],
[-0.04296685, 0.00431896, 0.01353475],
[-0.03333549, 0.03673425, -0.02340424],
[-0.03564823, 0.03987168, -0.0255153 ],
[-0.02670941, -0.02034902, 0.04689835],
[-0.02372832, -0.01835904, 0.04276315],
[-0.03229035, -0.00454579, 0.01301796],
[-0.03427308, -0.00092428, 0.01082572],
[-0.06443163, -0.00107994, -0.00457289],
[-0.06851917, 0.00827009, -0.01040453],
[ 0.37637392, 0.08864549, -0.14489254],
[ 0.37723047, 0.0891574 , -0.14561114],
[ 0.07678488, -0.05342679, -0.01783838],
[ 0.0741741 , -0.05382123, -0.01719362],
[-0.21770002, -0.23115733, 0.11361033],
[-0.23141973, -0.24684335, 0.12129052],
[ 0.35075927, 0.08465936, -0.13969956],
[ 0.35386336, 0.0851732 , -0.14089419],
[ 0.05521413, -0.05690382, -0.00964773],
[ 0.06471695, -0.05508107, -0.01335578],
[-0.21444406, -0.23920943, 0.11836243],
[-0.21851356, -0.24352723, 0.12055627]], dtype=float32),
15.0: array([[-1.48053512e-01, 4.94064569e-01, 2.36368701e-02],
[-1.30919874e-01, 4.32769388e-01, 2.10968666e-02],
[-1.00938469e-01, -1.05247520e-01, 3.77376676e-02],
[-9.70665663e-02, -1.32239953e-01, 3.77837084e-02],
[-1.02602869e-01, -7.11719096e-01, 6.80055991e-02],
[-1.10405318e-01, -7.69825280e-01, 7.32934251e-02],
[-1.36905730e-01, 5.01043975e-01, -7.58514914e-04],
[-1.26482651e-01, 4.59131062e-01, -8.82418477e-04],
[-8.76417011e-02, -9.94254798e-02, 3.24931107e-02],
[-9.32826996e-02, -1.51295274e-01, 3.67856584e-02],
[-8.13821107e-02, -6.55298531e-01, 8.14336091e-02],
[-1.01453036e-01, -7.95223713e-01, 9.33788121e-02],
[ 1.03405766e-01, -4.30830829e-02, -1.68358132e-01],
[ 1.04653209e-01, -4.83210571e-02, -1.68603718e-01],
[ 8.40532556e-02, 1.12173809e-02, -2.71299500e-02],
[ 8.29946995e-02, 4.33433568e-03, -2.54598614e-02],
[ 9.46829543e-02, 7.56069273e-02, 1.16076186e-01],
[ 1.02347404e-01, 8.90099928e-02, 1.22811861e-01],
[ 1.12486340e-01, -1.03110261e-01, -1.55322626e-01],
[ 1.13705203e-01, -1.03864819e-01, -1.56682298e-01],
[ 8.53831619e-02, 3.06903161e-02, -2.01442037e-02],
[ 8.88626724e-02, 2.80475747e-02, -2.43086070e-02],
[ 1.11776046e-01, 1.88344926e-01, 1.06459759e-01],
[ 1.13663167e-01, 1.91453889e-01, 1.08491853e-01]], dtype=float32),
17.0: array([[-0.35058346, 0.3329043 , 0.0137412 ],
[-0.3092446 , 0.2921714 , 0.01227193],
[-0.15701692, -0.01040395, 0.02301196],
[-0.14524022, -0.02662137, 0.02306052],
[-0.04746852, -0.33533093, 0.04212404],
[-0.05033231, -0.36284995, 0.04536595],
[-0.31832427, 0.34063938, -0.009092 ],
[-0.293143 , 0.31285244, -0.00864958],
[-0.12934594, -0.02222163, 0.02019953],
[-0.13009594, -0.04646472, 0.02276958],
[-0.02266653, -0.3504065 , 0.06101276],
[-0.0317041 , -0.41423538, 0.06777776],
[ 0.2833245 , -0.01002338, -0.15582904],
[ 0.28647032, -0.01322421, -0.15602012],
[ 0.18073586, -0.09847784, -0.01167197],
[ 0.17957535, -0.10230426, -0.01024948],
[ 0.1391714 , -0.23483022, 0.14082105],
[ 0.14933427, -0.24647588, 0.14952461],
[ 0.3012267 , -0.05189766, -0.14275567],
[ 0.3043043 , -0.05262647, -0.14395837],
[ 0.17195633, -0.0891717 , -0.00439421],
[ 0.18095048, -0.09120911, -0.00813082],
[ 0.14809373, -0.18439347, 0.1355253 ],
[ 0.15054181, -0.1878176 , 0.13805844]], dtype=float32),
19.0: array([[-0.09455644, 0.29851055, 0.13471371],
[-0.08341043, 0.2620168 , 0.1184083 ],
[-0.04296469, -0.00563558, 0.01575014],
[-0.03980618, -0.02010114, 0.00977753],
[-0.01429247, -0.29183292, -0.08800633],
[-0.01517179, -0.31591243, -0.09547284],
[-0.07704781, 0.29885396, 0.110355 ],
[-0.07077732, 0.2743527 , 0.10091425],
[-0.0328183 , -0.00758068, 0.0120823 ],
[-0.03334524, -0.02976166, 0.00492184],
[-0.01081309, -0.27365544, -0.05944307],
[-0.01333531, -0.3311195 , -0.07834792],
[ 0.13557966, -0.0527058 , -0.18830882],
[ 0.13643406, -0.05573613, -0.18974212],
[ 0.06564368, -0.02393635, -0.02206587],
[ 0.06493635, -0.02680261, -0.02182807],
[ 0.0158176 , -0.00071073, 0.15069307],
[ 0.0171787 , 0.00261365, 0.1615537 ],
[ 0.13472126, -0.08633238, -0.1878023 ],
[ 0.13603874, -0.08710409, -0.18939947],
[ 0.05892523, -0.01645526, -0.01040817],
[ 0.06282997, -0.0188802 , -0.01538836],
[ 0.01827607, 0.04520613, 0.1659651 ],
[ 0.01851483, 0.04593372, 0.16901092]], dtype=float32),
22.0: array([[-0.49775335, 0.16868335, 0.22451405],
[-0.43836907, 0.14742331, 0.1977226 ],
[-0.14792642, -0.07227184, 0.06679279],
[-0.12879969, -0.08262144, 0.05814494],
[ 0.11150835, -0.32970542, -0.05005807],
[ 0.12174367, -0.35639048, -0.05473327],
[-0.4922614 , 0.19464125, 0.19946386],
[-0.4533762 , 0.1785835 , 0.18324076],
[-0.13398881, -0.0653277 , 0.05875889],
[-0.12144299, -0.08880357, 0.05314834],
[ 0.11832652, -0.3219459 , -0.03052359],
[ 0.1410015 , -0.38541147, -0.04207325],
[ 0.13025254, 0.14377578, -0.21008189],
[ 0.13497889, 0.1419929 , -0.21230777],
[ 0.09474143, 0.02856167, -0.04937828],
[ 0.09615233, 0.02439209, -0.04893807],
[ 0.08781689, -0.08532841, 0.10807145],
[ 0.09260722, -0.08695165, 0.11558823],
[ 0.18211369, 0.10812669, -0.21906142],
[ 0.18394724, 0.10912353, -0.22101876],
[ 0.09709886, 0.02895462, -0.0408667 ],
[ 0.10251088, 0.03188912, -0.04685448],
[ 0.07109953, -0.0342305 , 0.1169661 ],
[ 0.07231593, -0.03506058, 0.11917011]], dtype=float32),
24.0: array([[ 0.05878175, 0.37557152, -0.08889071],
[ 0.05120701, 0.32911158, -0.07792394],
[-0.0424213 , -0.06519391, 0.01253306],
[-0.04660216, -0.08526184, 0.01718658],
[-0.15592939, -0.50566816, 0.11279979],
[-0.16858485, -0.5470773 , 0.12203307],
[ 0.05151994, 0.36940014, -0.09652848],
[ 0.04669544, 0.33835867, -0.08867995],
[-0.03741305, -0.06034703, 0.00947263],
[-0.0485412 , -0.09770565, 0.01803279],
[-0.12457195, -0.4490391 , 0.10542872],
[-0.15617272, -0.54955035, 0.1275503 ],
[-0.05888567, -0.10972722, -0.03507132],
[-0.05965445, -0.11375775, -0.03414084],
[ 0.03057634, 0.01647656, -0.01564473],
[ 0.02961452, 0.01212402, -0.01426446],
[ 0.14005718, 0.16166703, -0.00164282],
[ 0.1511655 , 0.17827578, -0.00310589],
[-0.06101912, -0.14806971, -0.02014938],
[-0.0613985 , -0.14918828, -0.0203681 ],
[ 0.03859929, 0.03372693, -0.01506035],
[ 0.03723516, 0.02998379, -0.01566284],
[ 0.16595218, 0.24281025, -0.01943559],
[ 0.16888982, 0.24701627, -0.01971926]], dtype=float32),
27.0: array([[-0.45981106, 0.20138167, 0.06745882],
[-0.405632 , 0.17592987, 0.05979374],
[-0.2097947 , -0.09454937, 0.06246585],
[-0.19446287, -0.10711583, 0.0611812 ],
[-0.07142456, -0.4134102 , 0.08616267],
[-0.07590351, -0.44670045, 0.0927152 ],
[-0.43315375, 0.2491603 , 0.02042717],
[-0.3992492 , 0.22893393, 0.01829192],
[-0.18116358, -0.09404217, 0.05252147],
[-0.18198797, -0.12232984, 0.05767327],
[-0.04075605, -0.44540685, 0.11520242],
[-0.05423898, -0.52243495, 0.12945703],
[ 0.2663341 , 0.2837588 , -0.32852072],
[ 0.27048662, 0.28182608, -0.32923368],
[ 0.18559058, -0.0272872 , -0.05311282],
[ 0.18461129, -0.03340921, -0.05030164],
[ 0.16708317, -0.37868547, 0.22585447],
[ 0.17942746, -0.39864632, 0.23951109],
[ 0.30001038, 0.23076457, -0.30585024],
[ 0.3031139 , 0.23258543, -0.30852175],
[ 0.18346733, -0.02970218, -0.03787874],
[ 0.19249558, -0.02399703, -0.04607868],
[ 0.18000631, -0.31728536, 0.21440358],
[ 0.18301687, -0.32329664, 0.21846868]], dtype=float32),
29.0: array([[-0.4005102 , 0.30894297, -0.10763498],
[-0.3531988 , 0.27064303, -0.09417138],
[-0.17050989, -0.0643676 , 0.03592921],
[-0.15674254, -0.08112403, 0.04219102],
[-0.03312076, -0.441771 , 0.18617041],
[-0.03462328, -0.47760978, 0.20119672],
[-0.3592955 , 0.36005738, -0.1500509 ],
[-0.33068544, 0.33114526, -0.13840093],
[-0.14660364, -0.06512623, 0.02802902],
[-0.14527501, -0.09654945, 0.04170308],
[-0.03232286, -0.4757285 , 0.20400454],
[-0.03700366, -0.55843127, 0.23854446],
[ 0.35240394, 0.28602147, -0.26506445],
[ 0.35614777, 0.2830631 , -0.26402545],
[ 0.15112086, -0.03662864, -0.03771642],
[ 0.14961915, -0.04328157, -0.03406437],
[-0.01142525, -0.4027103 , 0.19328536],
[-0.01134351, -0.42396784, 0.20369492],
[ 0.37032467, 0.21888869, -0.2257654 ],
[ 0.37387243, 0.22057217, -0.22772957],
[ 0.14084224, -0.04171887, -0.02512621],
[ 0.15142366, -0.03640699, -0.03113514],
[-0.00624459, -0.3370596 , 0.16577455],
[-0.00659518, -0.34343156, 0.16896673]], dtype=float32),
31.0: array([[-2.5403216e-01, 2.2741334e-01, 2.2207912e-02],
[-2.2462361e-01, 1.9943006e-01, 1.9789523e-02],
[-1.7316157e-01, -2.4185536e-02, 3.2316271e-02],
[-1.6649672e-01, -3.5779927e-02, 3.2248255e-02],
[-1.7608373e-01, -2.6981342e-01, 5.6437414e-02],
[-1.8938933e-01, -2.9179415e-01, 6.0773864e-02],
[-1.9640869e-01, 2.3403136e-01, -9.2913043e-03],
[-1.8077186e-01, 2.1480183e-01, -8.9406027e-03],
[-1.4228705e-01, -3.8025040e-02, 2.7693383e-02],
[-1.5258121e-01, -5.6477647e-02, 3.1169735e-02],
[-1.6414134e-01, -2.9613772e-01, 8.0633931e-02],
[-1.9709951e-01, -3.4648207e-01, 8.9748479e-02],
[ 4.3578359e-01, 1.8969189e-03, -2.1561533e-01],
[ 4.3800837e-01, -2.1760106e-04, -2.1590033e-01],
[ 1.9207144e-01, -1.1127279e-01, -2.1466233e-02],
[ 1.8851870e-01, -1.1441709e-01, -1.9535454e-02],
[ 2.3465948e-03, -2.8141376e-01, 1.8149480e-01],
[ 5.0679259e-03, -2.9688439e-01, 1.9267374e-01],
[ 4.2662501e-01, -2.7377088e-02, -1.9792126e-01],
[ 4.3076158e-01, -2.7943321e-02, -1.9960576e-01],
[ 1.7624626e-01, -1.0117667e-01, -1.1177621e-02],
[ 1.8850681e-01, -1.0264599e-01, -1.6393466e-02],
[ 3.0229732e-02, -2.4176723e-01, 1.7432448e-01],
[ 3.0431338e-02, -2.4620293e-01, 1.7759261e-01]], dtype=float32),
33.0: array([[-0.30939412, 0.25602823, -0.04024093],
[-0.2729689 , 0.22444367, -0.03505757],
[-0.14540924, -0.03632286, 0.03015728],
[-0.13519996, -0.04967344, 0.03301162],
[-0.05827967, -0.32549065, 0.10954671],
[-0.06200996, -0.35197398, 0.11828617],
[-0.25822553, 0.2849875 , -0.07913747],
[-0.2373937 , 0.261985 , -0.07326637],
[-0.12168919, -0.04142176, 0.02295251],
[-0.12313439, -0.06454325, 0.03085142],
[-0.06531238, -0.35134014, 0.13087037],
[-0.07464114, -0.41234308, 0.1504453 ],
[ 0.40212965, 0.14688215, -0.2538089 ],
[ 0.40502876, 0.14444757, -0.25345838],
[ 0.14765052, -0.05676161, -0.03645246],
[ 0.14535622, -0.0613478 , -0.0336886 ],
[-0.07254177, -0.30174014, 0.18461415],
[-0.07621259, -0.3177934 , 0.19535702],
[ 0.4035117 , 0.09991341, -0.22432698],
[ 0.40730777, 0.10055739, -0.22627276],
[ 0.13202561, -0.05636821, -0.02333602],
[ 0.14340957, -0.05422735, -0.02931234],
[-0.06396261, -0.25345033, 0.16908593],
[-0.06537932, -0.25819907, 0.17230119]], dtype=float32),
36.0: array([[-0.50396615, 0.29274705, 0.11235292],
[-0.44456264, 0.25620574, 0.09924819],
[-0.22879232, -0.0870034 , 0.06738545],
[-0.21191655, -0.10371231, 0.06408966],
[-0.07566422, -0.48056102, 0.05607919],
[-0.08024184, -0.51949763, 0.06004917],
[-0.43611053, 0.31565544, 0.04795878],
[-0.4011911 , 0.28948408, 0.04321256],
[-0.19339654, -0.0939235 , 0.05577299],
[-0.19430675, -0.12704915, 0.0584758 ],
[-0.07802735, -0.49644494, 0.10468064],
[-0.09013319, -0.58751464, 0.11188557],
[ 0.5509915 , 0.09985324, -0.45323592],
[ 0.55569 , 0.09697405, -0.45446956],
[ 0.22231162, -0.07597005, -0.06129523],
[ 0.21944131, -0.08167987, -0.0579662 ],
[-0.05070086, -0.30051842, 0.34129518],
[-0.05251679, -0.31465408, 0.3628589 ],
[ 0.5631992 , 0.05266851, -0.4245024 ],
[ 0.5685592 , 0.05288879, -0.42816284],
[ 0.20307161, -0.0650299 , -0.0383005 ],
[ 0.21908702, -0.06422012, -0.04959998],
[-0.03913989, -0.22786357, 0.33567 ],
[-0.04021633, -0.23219548, 0.34196517]], dtype=float32),
43.0: array([[-0.19446045, 0.14111139, -0.07030037],
[-0.17218906, 0.12368871, -0.06144791],
[-0.15839593, -0.02104995, 0.02986392],
[-0.15411961, -0.02843486, 0.03414929],
[-0.19640754, -0.18180351, 0.13685781],
[-0.21151847, -0.19659571, 0.14786866],
[-0.14698116, 0.1345232 , -0.09998616],
[-0.13547467, 0.123155 , -0.09219711],
[-0.1307178 , -0.03346084, 0.02571187],
[-0.14284621, -0.04546097, 0.03546733],
[-0.17770374, -0.19517069, 0.15311405],
[-0.2151542 , -0.22931235, 0.17828181],
[ 0.35657182, -0.07064203, -0.18639596],
[ 0.35819957, -0.0719614 , -0.18573496],
[ 0.17184988, -0.09450629, -0.01604349],
[ 0.16850391, -0.09613059, -0.01340861],
[ 0.0384743 , -0.1609671 , 0.1625384 ],
[ 0.04375331, -0.16953689, 0.17154731],
[ 0.34728843, -0.08188462, -0.15949845],
[ 0.35071227, -0.08284059, -0.16085397],
[ 0.16043657, -0.08185948, -0.00823229],
[ 0.1705197 , -0.08459689, -0.01241025],
[ 0.06988467, -0.13401261, 0.14252807],
[ 0.07083529, -0.13644493, 0.14524056]], dtype=float32),
46.0: array([[-0.07004075, 0.14823915, -0.12062981],
[-0.06191396, 0.12997559, -0.10586517],
[-0.04629326, -0.01873393, 0.00459833],
[-0.04439182, -0.02639409, 0.0105104 ],
[-0.04514502, -0.1830036 , 0.1235109 ],
[-0.04848257, -0.1979015 , 0.1336539 ],
[-0.03757638, 0.17531031, -0.13760066],
[-0.03421605, 0.16141215, -0.12669265],
[-0.03685435, -0.017296 , 0.00126569],
[-0.03946661, -0.0307527 , 0.010909 ],
[-0.0595881 , -0.1985355 , 0.1249932 ],
[-0.06695981, -0.2327856 , 0.14873178],
[ 0.23125271, 0.15430318, -0.09171642],
[ 0.23194242, 0.15288556, -0.09051689],
[ 0.0554679 , -0.00371557, -0.0152194 ],
[ 0.05372222, -0.00674571, -0.01326459],
[-0.11368837, -0.17641218, 0.06042663],
[-0.12047824, -0.1858384 , 0.06292184],
[ 0.21815486, 0.12041678, -0.06729179],
[ 0.22011712, 0.12138867, -0.06788805],
[ 0.04399684, -0.00941188, -0.01043033],
[ 0.04996548, -0.00638132, -0.01222807],
[-0.10761664, -0.14928453, 0.0415298 ],
[-0.10969734, -0.15211557, 0.04237182]], dtype=float32),
49.0: array([[-0.20135659, 0.173211 , 0.09515937],
[-0.17838092, 0.151635 , 0.08385899],
[-0.17266777, -0.04569894, 0.03492574],
[-0.16853195, -0.05543169, 0.03144316],
[-0.22395536, -0.27045548, -0.00536828],
[-0.24130216, -0.2924906 , -0.00617419],
[-0.1683 , 0.15366586, 0.06207301],
[-0.15553586, 0.14017004, 0.05657336],
[-0.144473 , -0.05364803, 0.03064693],
[-0.1583723 , -0.07219444, 0.02919251],
[-0.1854651 , -0.2508234 , 0.02720371],
[-0.22928987, -0.30384007, 0.02383701],
[ 0.26134488, -0.16293277, -0.23967549],
[ 0.26294056, -0.16474764, -0.24071723],
[ 0.17769045, -0.06324313, -0.02047354],
[ 0.17482771, -0.06494926, -0.01906266],
[ 0.15610212, 0.01751273, 0.21085319],
[ 0.16944753, 0.02176814, 0.2246778 ],
[ 0.26199064, -0.1696758 , -0.22916172],
[ 0.26474783, -0.17123303, -0.2310967 ],
[ 0.17359264, -0.04303854, -0.0085063 ],
[ 0.18153597, -0.04781896, -0.01453036],
[ 0.19296354, 0.06014663, 0.21369907],
[ 0.19615784, 0.06120408, 0.21766268]], dtype=float32),
51.0: array([[ 0.05423057, 0.01247072, 0.08354215],
[ 0.04663565, 0.0101702 , 0.0738216 ],
[-0.10592296, -0.08253632, 0.05219272],
[-0.11187241, -0.08576963, 0.04982777],
[-0.3032679 , -0.2083361 , 0.04662673],
[-0.32753304, -0.22505388, 0.05003305],
[ 0.11284761, -0.01568013, 0.04937433],
[ 0.10389626, -0.01590446, 0.04506857],
[-0.08209607, -0.08362238, 0.04298121],
[-0.10391498, -0.0968257 , 0.04545613],
[-0.2942024 , -0.16834351, 0.06427979],
[-0.35227153, -0.21019517, 0.07168412],
[ 0.38482615, -0.19035782, -0.24518085],
[ 0.3840928 , -0.19072337, -0.24601373],
[ 0.14046565, -0.01154285, -0.05303598],
[ 0.13534622, -0.01201138, -0.05120148],
[-0.06735972, 0.17494848, 0.13527331],
[-0.06789238, 0.18893746, 0.1434592 ],
[ 0.34070194, -0.17382467, -0.2325198 ],
[ 0.3439528 , -0.1752367 , -0.23459291],
[ 0.12412901, 0.01127877, -0.04147127],
[ 0.13377658, 0.0067709 , -0.04779555],
[-0.02012438, 0.20933525, 0.12918134],
[-0.02083206, 0.21309596, 0.13165748]], dtype=float32),
68.0: array([[-0.34548748, 0.25186995, 0.22617714],
[-0.305383 , 0.22048324, 0.19919659],
[-0.22268176, -0.06886853, 0.06893239],
[-0.21323173, -0.08307813, 0.06025151],
[-0.20878968, -0.39914298, -0.04644113],
[-0.22452854, -0.43157032, -0.05090456],
[-0.31031954, 0.2612819 , 0.1801766 ],
[-0.28639907, 0.23942372, 0.16506036],
[-0.19090584, -0.07227596, 0.05918448],
[-0.20218311, -0.10023706, 0.05381618],
[-0.16954535, -0.39377213, -0.00624312],
[-0.20996827, -0.47056535, -0.01836516],
[ 0.30257937, 0.01701579, -0.35086057],
[ 0.30554414, 0.01445311, -0.3531867 ],
[ 0.20078875, -0.04359638, -0.05599025],
[ 0.19815093, -0.04783039, -0.05452442],
[ 0.16652963, -0.12863263, 0.24445714],
[ 0.18047899, -0.13240786, 0.26086834],
[ 0.31792435, -0.01711002, -0.34664682],
[ 0.32122737, -0.017359 , -0.34965956],
[ 0.19827957, -0.0317658 , -0.03880182],
[ 0.2078451 , -0.0325252 , -0.04809804],
[ 0.20109268, -0.067109 , 0.25406846],
[ 0.20440647, -0.06848276, 0.25880724]], dtype=float32),
93.0: array([[-0.14368315, 0.2932776 , -0.0936086 ],
[-0.12734419, 0.25726846, -0.08201108],
[-0.12981503, -0.02027437, 0.01871139],
[-0.12705499, -0.03500349, 0.02377617],
[-0.17561604, -0.32166493, 0.13196208],
[-0.18919094, -0.3482897 , 0.14270103],
[-0.09987956, 0.23017864, -0.11107448],
[-0.09201773, 0.2096558 , -0.10220388],
[-0.10721786, -0.03038035, 0.01478375],
[-0.11824836, -0.05389083, 0.02462836],
[-0.16524684, -0.24797955, 0.13318977],
[-0.19847237, -0.3144232 , 0.15843274],
[ 0.32211155, -0.47713566, -0.10014185],
[ 0.32331243, -0.48034674, -0.09919468],
[ 0.13845916, -0.06773003, -0.01835995],
[ 0.13533346, -0.06789365, -0.01634242],
[-0.00680299, 0.34951344, 0.0623918 ],
[-0.00482513, 0.37611473, 0.06489561],
[ 0.30963567, -0.4695673 , -0.07841181],
[ 0.3126368 , -0.47357744, -0.07911362],
[ 0.12751502, -0.03034689, -0.0144863 ],
[ 0.13639992, -0.04285953, -0.01659761],
[ 0.02087965, 0.40337712, 0.04218986],
[ 0.02098497, 0.4107398 , 0.04305407]], dtype=float32),
162.0: array([[-1.9727017e-01, -4.4334605e-02, 1.4260431e-01],
[-1.7515738e-01, -3.8771003e-02, 1.2592117e-01],
[-2.1244104e-01, 1.4815328e-02, 7.8050502e-02],
[-2.0974469e-01, 1.7525828e-02, 7.3688120e-02],
[-3.2267377e-01, 7.6507017e-02, 5.3131707e-02],
[-3.4783560e-01, 8.3056070e-02, 5.6940284e-02],
[-1.3740374e-01, -1.4423163e-02, 1.3233699e-01],
[-1.2694649e-01, -1.2368227e-02, 1.2206805e-01],
[-1.7771706e-01, -7.4426387e-03, 7.3126167e-02],
[-1.9833694e-01, 5.8927335e-04, 7.4860401e-02],
[-2.9461169e-01, -2.6569694e-02, 5.5812616e-02],
[-3.5647836e-01, -5.7319026e-03, 6.5119587e-02],
[ 4.4086254e-01, 2.0722334e-01, -9.5875286e-02],
[ 4.4241121e-01, 2.0818883e-01, -9.7216718e-02],
[ 2.1230680e-01, -1.6446747e-01, -2.6930336e-02],
[ 2.0738715e-01, -1.6565020e-01, -2.6183102e-02],
[ 4.6934202e-02, -6.3952762e-01, 4.0254384e-02],
[ 5.4369818e-02, -6.8266791e-01, 4.2140432e-02],
[ 4.2378816e-01, 1.9387028e-01, -1.0517057e-01],
[ 4.2798233e-01, 1.9490911e-01, -1.0613864e-01],
[ 2.0066445e-01, -1.7155507e-01, -2.8574029e-02],
[ 2.1298161e-01, -1.6766459e-01, -3.1486735e-02],
[ 9.8323628e-02, -6.5786541e-01, 3.2294940e-02],
[ 9.9679306e-02, -6.6972762e-01, 3.2960348e-02]], dtype=float32)},
'1001B': {-18.0: array([[ 1.18133366e-01, -1.88309908e-01, -3.48824933e-02],
[ 1.74951598e-01, -2.79627442e-01, -5.14742956e-02],
[-2.78264703e-03, 6.94553275e-03, 2.22006142e-02],
[-4.83069709e-03, 1.12034138e-02, 3.46923731e-02],
[-3.36618125e-02, 5.56390733e-02, 2.77447663e-02],
[-5.52991852e-02, 9.13749784e-02, 4.55625392e-02],
[ 8.42830688e-02, -7.50173628e-02, -2.25346535e-02],
[ 1.26282990e-01, -1.54178485e-01, -3.59191559e-02],
[ 1.01538850e-02, 1.34188905e-02, 7.22976029e-03],
[ 1.31846424e-02, 2.56407983e-03, 1.23187602e-02],
[-1.24517800e-02, 2.24618837e-02, 1.06134443e-02],
[-2.29785517e-02, 3.95101309e-02, 1.93641372e-02],
[ 4.21158187e-02, 3.71672571e-01, 1.04648471e-02],
[ 4.08236310e-02, 3.08882892e-01, 7.57816853e-03],
[ 1.82827413e-02, 1.77068159e-01, 4.55470337e-03],
[ 1.58165749e-02, 1.48515105e-01, 4.98660048e-03],
[ 2.89575837e-04, 1.06156133e-02, 2.87179049e-04],
[-3.21495609e-04, 1.09609328e-02, 1.10675488e-03],
[ 3.77708003e-02, 3.75395179e-01, 5.11166919e-03],
[ 3.09147388e-02, 3.26939762e-01, 9.02234297e-03],
[ 3.00402120e-02, 1.38398051e-01, -3.41300704e-02],
[ 1.73018761e-02, 1.42095968e-01, -4.77384729e-03],
[ 9.30709019e-03, -1.82230351e-03, -2.12248545e-02],
[ 3.00935120e-03, 6.00902364e-03, -5.30830165e-03]], dtype=float32),
-13.0: array([[-0.0112452 , 0.14295611, 0.10049051],
[-0.01686104, 0.21182632, 0.14850718],
[-0.03546895, -0.06755801, -0.03952306],
[-0.05508869, -0.10537651, -0.06192107],
[-0.02654133, -0.09409943, -0.0595783 ],
[-0.04357116, -0.15451375, -0.09784607],
[-0.01985774, 0.04342896, 0.06710176],
[-0.02059358, 0.10775694, 0.10464401],
[-0.0218706 , -0.04322306, -0.00940886],
[-0.03125443, -0.05163033, -0.01715003],
[-0.01065568, -0.03748884, -0.02268458],
[-0.01902706, -0.06680565, -0.0414253 ],
[-0.10092748, -0.39926842, -0.00871534],
[-0.08649848, -0.33391133, -0.00280779],
[-0.04714739, -0.19204253, -0.00826428],
[-0.04108718, -0.16178189, -0.00603913],
[-0.00276733, -0.01302059, -0.00278414],
[-0.00325748, -0.01331649, -0.00288727],
[-0.0949869 , -0.40667695, -0.02190183],
[-0.08499439, -0.3501074 , -0.01436343],
[-0.01744868, -0.18215896, -0.04467044],
[-0.03268857, -0.15996172, -0.01512256],
[ 0.01080004, -0.01700355, -0.02143635],
[ 0.00110853, -0.01135189, -0.00585271]], dtype=float32),
8.0: array([[-0.07475785, 0.29503447, 0.00611002],
[-0.1107536 , 0.4374371 , 0.00860781],
[-0.01946294, -0.06711604, -0.02355612],
[-0.02989203, -0.10514888, -0.03680401],
[ 0.00364675, -0.13403197, -0.0212734 ],
[ 0.00601186, -0.2201054 , -0.03493803],
[-0.06677445, 0.12526144, 0.0200495 ],
[-0.09010354, 0.24902599, 0.0195592 ],
[-0.02191887, -0.04126787, -0.00191023],
[-0.02851316, -0.04155755, -0.01037002],
[ 0.00065557, -0.05304032, -0.00769964],
[ 0.00178967, -0.09466433, -0.01453274],
[-0.13136095, -0.54102904, 0.11125209],
[-0.11519007, -0.4482959 , 0.0953928 ],
[-0.05993262, -0.26113248, 0.04997608],
[-0.05202586, -0.21893172, 0.04272253],
[-0.0025961 , -0.01782043, 0.0013432 ],
[-0.00273245, -0.0183415 , 0.00150852],
[-0.12127904, -0.5558006 , 0.10212881],
[-0.10782851, -0.4790861 , 0.09309475],
[-0.02796945, -0.24385269, 0.00533589],
[-0.04279139, -0.21765903, 0.03262716],
[ 0.01043928, -0.02023044, -0.01951924],
[ 0.00057348, -0.01475031, -0.00321994]], dtype=float32),
13.0: array([[-1.50526792e-01, 1.46242410e-01, 1.16960280e-01],
[-2.22728580e-01, 2.16727033e-01, 1.72913417e-01],
[ 1.39365597e-02, -7.42558986e-02, -2.36520730e-02],
[ 2.23793797e-02, -1.15764685e-01, -3.73613946e-02],
[ 5.15612774e-02, -1.00536592e-01, -5.07483408e-02],
[ 8.46999511e-02, -1.65080488e-01, -8.33569467e-02],
[-1.14425756e-01, 3.80315855e-02, 9.03270394e-02],
[-1.66747645e-01, 1.05269030e-01, 1.30992979e-01],
[-1.18839527e-02, -4.94832993e-02, 4.44862200e-03],
[-1.19188186e-02, -5.88586777e-02, 5.73510479e-04],
[ 1.90001745e-02, -4.01986726e-02, -1.88462473e-02],
[ 3.52013744e-02, -7.15074241e-02, -3.48312333e-02],
[-1.02355830e-01, -4.57062095e-01, 8.62820968e-02],
[-9.37937200e-02, -3.83010656e-01, 7.88099021e-02],
[-4.42633778e-02, -2.19350085e-01, 3.63511518e-02],
[-3.86833064e-02, -1.84928447e-01, 3.20412964e-02],
[-4.77059686e-04, -1.46224741e-02, -2.33387516e-04],
[-1.13505623e-04, -1.49272261e-02, -3.28588649e-04],
[-8.83236453e-02, -4.63942707e-01, 7.08840489e-02],
[-7.91515037e-02, -3.99770170e-01, 6.61816299e-02],
[-1.58829130e-02, -2.05032870e-01, -8.13293178e-03],
[-3.03206444e-02, -1.81968302e-01, 2.04377957e-02],
[ 1.02425264e-02, -1.77637208e-02, -2.05123164e-02],
[ 1.10441656e-03, -1.25322584e-02, -4.01904294e-03]], dtype=float32),
17.0: array([[ 1.05999485e-01, 6.96798787e-02, -1.01033643e-01],
[ 1.57700494e-01, 1.02432035e-01, -1.49876952e-01],
[-7.93236215e-03, -1.07838206e-01, 2.65160333e-02],
[-1.23913139e-02, -1.68009445e-01, 4.14474383e-02],
[-3.46218832e-02, -1.08276501e-01, 4.88251932e-02],
[-5.68460785e-02, -1.77789241e-01, 8.01754072e-02],
[ 1.61692277e-02, 2.57226117e-02, -3.41605805e-02],
[ 6.61601052e-02, 5.79037070e-02, -7.84138814e-02],
[-2.17959434e-02, -5.07823415e-02, 1.98256988e-02],
[-1.38959214e-02, -7.72777349e-02, 1.99718047e-02],
[-1.47277778e-02, -4.22130972e-02, 1.95557214e-02],
[-2.52160095e-02, -7.63386488e-02, 3.46926451e-02],
[-4.02615130e-01, -1.81526199e-01, 2.50307024e-01],
[-3.38747233e-01, -1.51283994e-01, 2.09195167e-01],
[-1.90156579e-01, -8.96021500e-02, 1.18234165e-01],
[-1.60192698e-01, -7.67306015e-02, 1.00490339e-01],
[-1.06764352e-02, -7.92888645e-03, 6.85321912e-03],
[-1.07887499e-02, -8.69159587e-03, 7.61635881e-03],
[-4.00986612e-01, -1.86940730e-01, 2.45923862e-01],
[-3.49304706e-01, -1.60080716e-01, 2.17678264e-01],
[-1.47623882e-01, -9.04135182e-02, 6.32960424e-02],
[-1.51744977e-01, -7.47964755e-02, 8.79914761e-02],
[ 2.06660177e-03, -1.17530441e-02, -1.73030347e-02],
[-6.38063345e-03, -6.23983797e-03, -1.61702977e-04]], dtype=float32),
22.0: array([[-9.11770388e-02, -1.31094322e-01, 2.97883272e-01],
[-1.35447577e-01, -1.94972143e-01, 4.41338480e-01],
[-5.88189922e-02, -6.13654740e-02, -2.41629463e-02],
[-9.10781845e-02, -9.50692296e-02, -3.86870056e-02],
[-2.47784369e-02, -1.63619202e-02, -9.91136879e-02],
[-4.06556614e-02, -2.68393904e-02, -1.62798315e-01],
[-8.03407282e-02, -8.01510140e-02, 1.85120404e-01],
[-1.08194910e-01, -1.27985865e-01, 2.97044903e-01],
[-4.11466137e-02, -3.22834998e-02, 5.10238810e-03],
[-5.95600829e-02, -5.59967756e-02, 1.08728055e-02],
[-1.04441987e-02, -6.49153395e-03, -3.77581008e-02],
[-1.83432028e-02, -1.19411359e-02, -6.86419681e-02],
[-1.61743000e-01, 3.44373807e-02, -1.01534091e-01],
[-1.41626939e-01, 2.41151210e-02, -7.36141205e-02],
[-7.49408528e-02, 1.62507873e-02, -5.47885559e-02],
[-6.55113608e-02, 1.19890636e-02, -4.35573868e-02],
[-4.16387571e-03, 2.04114811e-04, -6.40431885e-03],
[-4.56530415e-03, 1.57867529e-04, -6.77879760e-03],
[-1.50832579e-01, 3.84181365e-02, -1.25137389e-01],
[-1.33271307e-01, 3.50215882e-02, -1.03597917e-01],
[-4.13053520e-02, 1.45732774e-03, -8.74355808e-02],
[-5.44436425e-02, 1.21557973e-02, -5.50708920e-02],
[ 9.14254412e-03, -7.67410686e-03, -2.36991588e-02],
[-2.73344951e-04, -1.29791780e-03, -8.21637362e-03]], dtype=float32),
26.0: array([[-0.1368274 , -0.11533888, 0.18143742],
[-0.20252408, -0.17194901, 0.26889113],
[-0.01198337, -0.07009092, 0.00532349],
[-0.01793809, -0.10883739, 0.00756914],
[ 0.02635919, -0.02784307, -0.04369224],
[ 0.04332242, -0.04570227, -0.07178073],
[-0.11809674, -0.05156573, 0.12243323],
[-0.16219324, -0.09715518, 0.18815291],
[-0.02807509, -0.02614654, 0.01629342],
[-0.03370347, -0.05387031, 0.02456341],
[ 0.00891865, -0.01031007, -0.01626516],
[ 0.0172211 , -0.01948772, -0.02983904],
[-0.20377831, 0.16581263, 0.01494089],
[-0.17954312, 0.13678491, 0.0206255 ],
[-0.09285554, 0.07743537, 0.00251964],
[-0.07996275, 0.06349955, 0.00452129],
[-0.00378551, 0.00313708, -0.00186021],
[-0.00352625, 0.00278937, -0.0017744 ],
[-0.19023848, 0.16763484, -0.00324981],
[-0.1669738 , 0.14677817, 0.00201439],
[-0.06096201, 0.05556365, -0.03848111],
[-0.07030387, 0.06227095, -0.0081778 ],
[ 0.00631598, -0.00449238, -0.02192551],
[-0.00165565, 0.0017312 , -0.00565192]], dtype=float32),
30.0: array([[-2.00018778e-01, -1.15553133e-01, 1.87804133e-01],
[-2.95991033e-01, -1.71273038e-01, 2.77993649e-01],
[ 1.00621553e-02, -7.97187630e-03, -2.98362672e-02],
[ 1.66010354e-02, -1.19405407e-02, -4.71787006e-02],
[ 6.14772812e-02, 2.40133628e-02, -7.46669099e-02],
[ 1.00996301e-01, 3.94552462e-02, -1.22639738e-01],
[-1.56228051e-01, -7.97872990e-02, 1.24851100e-01],
[-2.24698022e-01, -1.21162534e-01, 1.94091484e-01],
[-2.14052405e-02, -1.32097369e-02, 9.08556045e-04],
[-2.34439895e-02, -1.96001232e-02, -6.42296567e-04],
[ 2.24003009e-02, 8.82335287e-03, -2.82538682e-02],
[ 4.17214036e-02, 1.62750315e-02, -5.16378433e-02],
[-1.69260755e-01, -2.45188251e-02, -8.04692227e-03],
[-1.52892932e-01, -2.56773662e-02, 1.43011182e-03],
[-7.49621987e-02, -1.03285955e-02, -9.23031569e-03],
[-6.49583191e-02, -9.30300076e-03, -6.02099299e-03],
[-1.83337741e-03, -1.23536389e-04, -3.30191175e-03],
[-1.26854237e-03, 3.54641874e-04, -3.53847677e-03],
[-1.51948512e-01, -2.05066130e-02, -2.64336392e-02],
[-1.34089008e-01, -1.60035528e-02, -1.81214716e-02],
[-4.34023738e-02, -2.26475149e-02, -4.74739708e-02],
[-5.51630445e-02, -1.05781769e-02, -1.70457605e-02],
[ 8.15709960e-03, -8.78625363e-03, -2.20778622e-02],
[-5.17860579e-04, -2.58265692e-03, -6.09858287e-03]], dtype=float32),
35.0: array([[-2.27499679e-01, -2.42018163e-01, 3.66758138e-01],
[-3.37255239e-01, -3.59272927e-01, 5.43432415e-01],
[-2.46942863e-02, -8.23185593e-02, -2.23686732e-02],
[-3.74993905e-02, -1.27108961e-01, -3.61541621e-02],
[ 3.97855900e-02, -4.35548229e-03, -1.15884945e-01],
[ 6.53794631e-02, -7.09141046e-03, -1.90350354e-01],
[-1.59854680e-01, -1.75580293e-01, 2.29799435e-01],
[-2.40536645e-01, -2.58962631e-01, 3.67044568e-01],
[-3.11594624e-02, -5.96841238e-02, 1.03301387e-02],
[-4.60647717e-02, -9.12753791e-02, 1.93753764e-02],
[ 1.44043900e-02, -2.86675547e-03, -4.40571979e-02],
[ 2.67062392e-02, -4.55724029e-03, -8.01439285e-02],
[-7.07146674e-02, -1.31669119e-01, -1.09063499e-01],
[-6.99403509e-02, -1.22092299e-01, -7.71167427e-02],
[-2.90846340e-02, -6.02537841e-02, -5.94733059e-02],
[-2.71634590e-02, -5.37455864e-02, -4.69634272e-02],
[ 4.56977177e-05, -3.42764333e-03, -7.14857318e-03],
[ 2.21787443e-04, -3.18200979e-03, -7.67358905e-03],
[-5.27907759e-02, -1.19319119e-01, -1.36697754e-01],
[-4.84120175e-02, -1.01781793e-01, -1.13485537e-01],
[-1.39163539e-03, -6.19780906e-02, -9.28540230e-02],
[-1.66221950e-02, -4.85427603e-02, -5.96612468e-02],
[ 1.08844656e-02, -1.00248484e-02, -2.43217088e-02],
[ 1.87564362e-03, -4.60565602e-03, -8.57876893e-03]], dtype=float32),
43.0: array([[-0.03422698, 0.0219276 , 0.26586667],
[-0.05132531, 0.03292428, 0.3939278 ],
[-0.05826478, -0.0135453 , -0.02735849],
[-0.09058046, -0.02087476, -0.04350309],
[-0.03946101, -0.01701864, -0.09327448],
[-0.06478548, -0.02792886, -0.15320142],
[-0.02141931, -0.02907741, 0.15883563],
[-0.03281641, -0.0118154 , 0.26017493],
[-0.02664969, -0.02517635, -0.00098631],
[-0.04512969, -0.02332332, 0.00320028],
[-0.0152822 , -0.00786365, -0.03575705],
[-0.02787559, -0.01296202, -0.06480987],
[-0.00593319, -0.32611504, -0.13941337],
[-0.00642949, -0.27687377, -0.10723824],
[-0.00269838, -0.15442222, -0.07213566],
[-0.00424764, -0.12983485, -0.05844666],
[-0.00072422, -0.00896054, -0.0071668 ],
[-0.0015052 , -0.00848231, -0.00742251],
[ 0.00053042, -0.32753608, -0.16097912],
[-0.0025122 , -0.28133264, -0.13451897],
[ 0.02341778, -0.15215296, -0.10316559],
[ 0.00451439, -0.12983732, -0.06910806],
[ 0.01366246, -0.01689863, -0.02500806],
[ 0.00347639, -0.00994114, -0.00916432]], dtype=float32),
48.0: array([[ 0.11825339, 0.01286125, 0.15518437],
[ 0.1752402 , 0.02020713, 0.2304122 ],
[-0.02440295, 0.0214928 , 0.03128562],
[-0.03831647, 0.03389681, 0.0481727 ],
[-0.05165716, 0.01462595, -0.01507856],
[-0.08483513, 0.02404398, -0.02478757],
[ 0.0590561 , -0.06302807, 0.09257875],
[ 0.10678815, -0.04418084, 0.15064746],
[-0.01119609, -0.02432105, 0.01948496],
[-0.01161104, -0.00941616, 0.03574726],
[-0.02016772, 0.00346453, -0.00568719],
[-0.03626468, 0.00855273, -0.01014906],
[-0.14991498, -0.5220059 , -0.06947006],
[-0.12320855, -0.44479254, -0.05260213],
[-0.06930507, -0.24227755, -0.03548918],
[-0.06041632, -0.20526552, -0.02754301],
[-0.00337124, -0.01131709, -0.00285257],
[-0.00502101, -0.01131174, -0.00252493],
[-0.13741295, -0.50586635, -0.0826842 ],
[-0.13009161, -0.44497445, -0.06828923],
[ 0.0313036 , -0.15293588, -0.05923861],
[-0.03673863, -0.18521914, -0.03664908],
[ 0.0489165 , 0.02220794, -0.01652002],
[ 0.01007678, -0.00304122, -0.00563322]], dtype=float32),
56.0: array([[-1.39564812e-01, 1.48426682e-01, 1.10185228e-01],
[-2.06970513e-01, 2.20243320e-01, 1.63048610e-01],
[-8.09035543e-03, -8.75890329e-02, -7.01760314e-03],
[-1.20916991e-02, -1.36312827e-01, -1.14356754e-02],
[ 3.02670095e-02, -1.12163804e-01, -3.50933038e-02],
[ 4.97246869e-02, -1.84157908e-01, -5.76506667e-02],
[-8.69229287e-02, 7.64358556e-03, 8.53211358e-02],
[-1.38867468e-01, 8.25122818e-02, 1.23226777e-01],
[-1.07478080e-02, -7.03140274e-02, 1.07999137e-02],
[-1.91006400e-02, -7.98767805e-02, 1.15828430e-02],
[ 1.14464341e-02, -4.56688143e-02, -1.28420079e-02],
[ 2.07842086e-02, -8.04724395e-02, -2.38659717e-02],
[ 4.09063436e-02, -6.95723534e-01, 8.70965943e-02],
[ 2.88544670e-02, -5.86317599e-01, 7.91614801e-02],
[ 2.22455040e-02, -3.31475437e-01, 3.72198150e-02],
[ 1.70306042e-02, -2.79840142e-01, 3.30695696e-02],
[ 2.47649965e-03, -2.07547396e-02, 2.13530933e-04],
[ 2.41362699e-03, -2.09644586e-02, 3.05745605e-04],
[ 5.28594181e-02, -6.99248195e-01, 7.19475076e-02],
[ 4.29646038e-02, -6.04120195e-01, 6.72362223e-02],
[ 4.31329347e-02, -2.96833515e-01, -8.77996627e-03],
[ 2.44133063e-02, -2.71981925e-01, 2.06509605e-02],
[ 1.36583345e-02, -1.95983667e-02, -2.11282000e-02],
[ 4.39605815e-03, -1.70498081e-02, -4.15294897e-03]], dtype=float32),
65.0: array([[ 1.00193754e-01, -1.09461159e-01, 1.35065779e-01],
[ 1.48223922e-01, -1.62042603e-01, 1.99826688e-01],
[-2.66734101e-02, -3.98326442e-02, -1.40413204e-02],
[-4.19021510e-02, -6.12889901e-02, -2.24771462e-02],
[-4.87915725e-02, -4.06869641e-03, -4.75463383e-02],
[-8.01339820e-02, -6.63701445e-03, -7.81039223e-02],
[ 6.48655593e-02, -1.16098583e-01, 1.03387475e-01],
[ 1.02413535e-01, -1.46236002e-01, 1.50225535e-01],
[-4.92355460e-03, -4.57473509e-02, 1.03479307e-02],
[-8.70311167e-03, -5.66163659e-02, 9.85320006e-03],
[-1.85753889e-02, -3.25081102e-03, -1.75244194e-02],
[-3.38895880e-02, -4.45767585e-03, -3.24695297e-02],
[-1.97089948e-02, -3.31211746e-01, 9.65207741e-02],
[-1.25784911e-02, -2.86693931e-01, 8.82637054e-02],
[-1.07662855e-02, -1.55180678e-01, 4.11147326e-02],
[-9.41807218e-03, -1.31949276e-01, 3.63514833e-02],
[-1.52185000e-03, -8.55916273e-03, 1.07492306e-04],
[-2.31644535e-03, -7.98362400e-03, 2.99270469e-05],
[-2.12671701e-02, -3.24824035e-01, 8.01502913e-02],
[-2.09669601e-02, -2.79233545e-01, 7.41291270e-02],
[ 1.17288632e-02, -1.49619520e-01, -3.66265094e-03],
[-4.40787384e-03, -1.28533006e-01, 2.41445098e-02],
[ 1.16240820e-02, -1.60149038e-02, -1.99354403e-02],
[ 2.57951068e-03, -9.66545008e-03, -3.71187087e-03]], dtype=float32),
93.0: array([[ 0.03387733, -0.03269347, 0.20284744],
[ 0.05009072, -0.04739994, 0.30098483],
[-0.03208399, 0.01771432, 0.0275381 ],
[-0.04996682, 0.02815826, 0.04214362],
[-0.03568201, 0.02355189, -0.03084492],
[-0.05858772, 0.03870765, -0.05068607],
[ 0.00605299, -0.08838265, 0.12567939],
[ 0.02253845, -0.08708252, 0.20094107],
[-0.01909234, -0.02685613, 0.02202544],
[-0.02532668, -0.01535899, 0.03884961],
[-0.0141524 , 0.00693186, -0.01158481],
[-0.02534322, 0.0147431 , -0.02097993],
[-0.12901762, -0.4864478 , -0.05999541],
[-0.10918865, -0.41659138, -0.04245479],
[-0.05657459, -0.2236111 , -0.03221293],
[-0.05152459, -0.19077094, -0.02433904],
[-0.0014283 , -0.00944967, -0.00325171],
[-0.00354453, -0.00986327, -0.00300564],
[-0.10362079, -0.46166003, -0.07738842],
[-0.10605905, -0.41073835, -0.06286491],
[ 0.08585218, -0.10326439, -0.0634507 ],
[-0.01610331, -0.16226546, -0.03579432],
[ 0.07352694, 0.04163834, -0.02017326],
[ 0.01692932, 0.00266931, -0.00647708]], dtype=float32)},
'1004B': {-2.5: array([[-4.9269088e-02, 7.0253588e-02, 1.5979638e-02],
[-2.2227876e-02, 7.7585109e-02, 5.9659332e-03],
[ 2.0059975e-02, -6.4141909e-03, -5.7245540e-03],
[ 1.9896487e-02, -5.7877542e-04, -6.3721365e-03],
[ 1.5116483e-01, -3.0431725e-02, -4.8283156e-02],
[ 9.4835952e-02, -1.8845901e-02, -3.0321309e-02],
[-1.6076533e-01, 7.1748425e-03, 5.8123033e-02],
[-2.0262763e-02, 3.4287576e-02, 6.4535746e-03],
[-2.3573207e-02, -2.3675224e-02, 1.1320283e-02],
[ 3.1886149e-02, -6.8608490e-03, -9.9371914e-03],
[ 2.2914739e-01, -4.7698654e-02, -7.2994366e-02],
[ 1.6145971e-01, -3.2227099e-02, -5.1608320e-02],
[-1.0788523e-01, -1.6615471e-02, 3.9557178e-02],
[-8.4514720e-03, 7.2295417e-04, 3.0529252e-03],
[-4.0272888e-02, -1.1276133e-02, 1.5345521e-02],
[ 4.1297274e-03, -2.2328070e-03, -1.1297760e-03],
[ 5.4203216e-02, -1.2067673e-02, -1.7166398e-02],
[ 3.4290820e-02, -6.9121635e-03, -1.0952126e-02],
[-8.7242862e-03, -1.4849324e-03, 3.2017203e-03],
[-7.4087409e-04, -1.1080279e-04, 2.7173405e-04],
[-3.4837236e-03, -7.0597418e-04, 1.2922820e-03],
[-1.2910664e-04, -1.0474942e-04, 5.7992504e-05],
[ 1.1447904e-03, -3.4757238e-04, -3.5074141e-04],
[ 1.0151347e-03, -2.1127274e-04, -3.2337700e-04]], dtype=float32),
2.0: array([[-4.75066118e-02, 7.18677565e-02, 1.52170612e-02],
[-2.65195742e-02, 7.61082619e-02, 8.53245705e-03],
[ 5.01976758e-02, 1.24573931e-02, -2.60474011e-02],
[ 3.61963548e-02, 9.85481869e-03, -1.75942592e-02],
[ 2.64017045e-01, 4.15130332e-02, -1.26017049e-01],
[ 1.65046066e-01, 2.59303935e-02, -7.87013397e-02],
[-1.31244317e-01, 2.25897823e-02, 4.21097055e-02],
[-1.94653720e-02, 3.51067223e-02, 6.03849301e-03],
[ 3.47587056e-02, 1.14631364e-02, -2.65104156e-02],
[ 6.14138767e-02, 1.18438387e-02, -3.01160309e-02],
[ 4.04114068e-01, 6.37374818e-02, -1.93391785e-01],
[ 2.81258851e-01, 4.41663414e-02, -1.34150177e-01],
[-8.63789916e-02, -5.66156069e-03, 2.79017612e-02],
[-6.77520363e-03, 1.61726680e-03, 2.12541176e-03],
[-2.10426953e-02, -3.65976739e-04, 3.63380439e-03],
[ 1.10138580e-02, 2.05377955e-03, -5.75426454e-03],
[ 9.75721404e-02, 1.55008165e-02, -4.69489023e-02],
[ 5.99006489e-02, 9.41405632e-03, -2.85921060e-02],
[-6.99245324e-03, -6.06534421e-04, 2.26541818e-03],
[-5.89341798e-04, -3.31359188e-05, 1.89117854e-04],
[-2.52112770e-03, -1.89970102e-04, 7.40200223e-04],
[ 1.07333100e-04, 3.68764777e-05, -9.45125212e-05],
[ 2.29527173e-03, 3.77760531e-04, -1.13399071e-03],
[ 1.79006776e-03, 2.82285531e-04, -8.56626080e-04]], dtype=float32),
5.0: array([[-1.7156062e-01, -3.0334314e-02, 1.0692491e-01],
[-1.5186127e-01, -2.5210133e-02, 1.0100596e-01],
[ 3.7026159e-02, -1.9346474e-02, -1.0590827e-02],
[ 2.6104443e-02, -1.0485310e-02, -6.7428360e-03],
[ 2.3924948e-01, -6.2245902e-02, -8.4258355e-02],
[ 1.4964461e-01, -3.8671780e-02, -5.2703563e-02],
[-1.9956408e-01, -4.2549726e-02, 9.3369596e-02],
[-7.9246514e-02, -1.4356094e-02, 5.0327171e-02],
[ 1.5753334e-02, -4.0227223e-02, -3.3104648e-03],
[ 5.1724706e-02, -1.7468456e-02, -1.7015718e-02],
[ 3.6557236e-01, -9.6882634e-02, -1.2869804e-01],
[ 2.5502604e-01, -6.5999456e-02, -8.9835882e-02],
[-9.9214040e-02, -2.2638733e-02, 3.7864558e-02],
[-1.1036652e-02, -2.4512447e-03, 5.3423196e-03],
[-2.7787842e-02, -1.5519940e-02, 1.0673237e-02],
[ 9.1581056e-03, -4.2074234e-03, -3.1003356e-03],
[ 8.7928019e-02, -2.4205614e-02, -3.0922597e-02],
[ 5.4287616e-02, -1.4124919e-02, -1.9121738e-02],
[-7.7959974e-03, -1.7772205e-03, 2.8943757e-03],
[-6.8627921e-04, -1.5959526e-04, 2.6475795e-04],
[-2.9049809e-03, -8.8400947e-04, 1.0834557e-03],
[ 3.8187558e-05, -1.6362818e-04, -7.3817982e-06],
[ 2.0288876e-03, -6.6379114e-04, -7.0968125e-04],
[ 1.6194314e-03, -4.2906689e-04, -5.7013793e-04]], dtype=float32),
7.0: array([[-4.66910601e-02, 6.33070469e-02, 1.73988529e-02],
[-2.37727500e-02, 7.48846084e-02, 7.53558148e-03],
[ 4.48971651e-02, -3.56182940e-02, -6.62765186e-03],
[ 3.36174853e-02, -1.69562064e-02, -6.84961025e-03],
[ 2.45622307e-01, -1.41680077e-01, -5.21154702e-02],
[ 1.53621867e-01, -8.80928636e-02, -3.27075198e-02],
[-1.39939412e-01, -1.67711750e-02, 5.82763776e-02],
[-1.89734641e-02, 3.08327153e-02, 7.13091157e-03],
[ 2.27869041e-02, -7.70312026e-02, 9.65003110e-03],
[ 5.64687289e-02, -3.58227231e-02, -1.08864475e-02],
[ 3.75464886e-01, -2.19955161e-01, -7.89220035e-02],
[ 2.61755735e-01, -1.50364026e-01, -5.56795523e-02],
[-9.32000577e-02, -3.22292782e-02, 3.92418876e-02],
[-7.24425865e-03, -6.90057641e-04, 3.06420657e-03],
[-2.58060824e-02, -2.74495501e-02, 1.48776779e-02],
[ 9.76908859e-03, -8.81359074e-03, -1.34762551e-03],
[ 9.04066861e-02, -5.46549149e-02, -1.86284930e-02],
[ 5.57257943e-02, -3.21569666e-02, -1.18218530e-02],
[-7.54715595e-03, -2.72667664e-03, 3.17389611e-03],
[-6.36785815e-04, -2.22069299e-04, 2.69537821e-04],
[-2.79596797e-03, -1.45308417e-03, 1.27287128e-03],
[ 5.78002728e-05, -3.19048733e-04, 5.12159131e-05],
[ 2.09794473e-03, -1.46492186e-03, -3.88712739e-04],
[ 1.66318426e-03, -9.74208699e-04, -3.49631970e-04]], dtype=float32),
8.5: array([[-4.92597111e-02, 7.42106661e-02, 1.43896183e-02],
[-2.76343767e-02, 7.91683048e-02, 6.94074295e-03],
[ 4.50091958e-02, 1.16754863e-02, -2.28559189e-02],
[ 3.32429558e-02, 9.64136329e-03, -1.59210823e-02],
[ 2.44154751e-01, 3.90866175e-02, -1.14047796e-01],
[ 1.52680367e-01, 2.44323239e-02, -7.12568760e-02],
[-1.35251611e-01, 2.05621570e-02, 4.52183038e-02],
[-2.03400794e-02, 3.62897180e-02, 5.62427472e-03],
[ 2.55047288e-02, 8.86086561e-03, -2.00991463e-02],
[ 5.62487245e-02, 1.11770760e-02, -2.69930214e-02],
[ 3.73374283e-01, 5.99018931e-02, -1.74821392e-01],
[ 2.60163993e-01, 4.16037068e-02, -1.21446311e-01],
[-8.87996405e-02, -7.94139132e-03, 3.04865502e-02],
[-7.01636309e-03, 1.52187748e-03, 2.29646591e-03],
[-2.37454586e-02, -1.69286935e-03, 5.83320996e-03],
[ 9.85236466e-03, 1.83915382e-03, -5.01320697e-03],
[ 8.99796113e-02, 1.45139731e-02, -4.23395112e-02],
[ 5.53935133e-02, 8.86303559e-03, -2.58757826e-02],
[-7.18324538e-03, -7.96514621e-04, 2.47581885e-03],
[-6.06714282e-04, -4.87486868e-05, 2.07215737e-04],
[-2.64110183e-03, -2.76946055e-04, 8.53940146e-04],
[ 7.03188925e-05, 2.56310814e-05, -6.83582912e-05],
[ 2.09689583e-03, 3.47510446e-04, -1.01099804e-03],
[ 1.65391876e-03, 2.65299343e-04, -7.74377258e-04]], dtype=float32),
10.5: array([[-4.8757210e-02, 6.9497101e-02, 1.5730826e-02],
[-2.4018072e-02, 7.8169808e-02, 7.4425046e-03],
[ 3.1567704e-02, -1.3658900e-02, -1.5672835e-02],
[ 2.6121164e-02, -4.5375745e-03, -1.1833688e-02],
[ 1.9431807e-01, -5.7702754e-02, -8.6174287e-02],
[ 1.2168377e-01, -3.5814829e-02, -5.3901821e-02],
[-1.4963685e-01, 4.2855527e-04, 4.9897294e-02],
[-2.0034360e-02, 3.3924285e-02, 6.3202917e-03],
[-1.3207954e-03, -3.7473816e-02, -7.3875422e-03],
[ 4.3170106e-02, -1.3984833e-02, -1.9787570e-02],
[ 2.9604992e-01, -8.9963824e-02, -1.3169578e-01],
[ 2.0726989e-01, -6.1179761e-02, -9.1840461e-02],
[-9.9732459e-02, -2.1411840e-02, 3.3515956e-02],
[-7.8196945e-03, 3.3621307e-04, 2.5796464e-03],
[-3.2952208e-02, -1.5724033e-02, 9.4598690e-03],
[ 6.7594410e-03, -3.8816207e-03, -3.3972787e-03],
[ 7.0785038e-02, -2.2536065e-02, -3.1694245e-02],
[ 4.4083696e-02, -1.3100771e-02, -1.9550726e-02],
[-8.0675259e-03, -1.8701344e-03, 2.7157525e-03],
[-6.8342849e-04, -1.4468492e-04, 2.2899490e-04],
[-3.1179921e-03, -9.2415448e-04, 1.0103692e-03],
[-3.8916140e-05, -1.6054211e-04, -1.7538892e-05],
[ 1.5845250e-03, -6.2438904e-04, -7.3359744e-04],
[ 1.3114482e-03, -3.9846703e-04, -5.8336934e-04]], dtype=float32),
12.5: array([[-4.9026076e-02, 7.5596273e-02, 1.3509233e-02],
[-2.7129848e-02, 7.7647403e-02, 6.5284329e-03],
[ 4.2978253e-02, 2.7397834e-02, -2.5928102e-02],
[ 3.2106433e-02, 1.8230369e-02, -1.7639322e-02],
[ 2.3618002e-01, 9.8345540e-02, -1.2566811e-01],
[ 1.4771535e-01, 6.1305471e-02, -7.8489378e-02],
[-1.3636662e-01, 3.5028521e-02, 4.2505633e-02],
[-2.0239921e-02, 3.6955986e-02, 5.1915227e-03],
[ 2.1850543e-02, 3.8795412e-02, -2.5756272e-02],
[ 5.4200146e-02, 2.6648207e-02, -3.0027075e-02],
[ 3.6103404e-01, 1.5174034e-01, -1.9281840e-01],
[ 2.5169381e-01, 1.0451723e-01, -1.3378537e-01],
[-8.9690827e-02, 2.4151439e-03, 2.8743092e-02],
[-7.0823664e-03, 2.3507976e-03, 2.1370309e-03],
[-2.4802126e-02, 7.9431664e-03, 4.0907701e-03],
[ 9.3893502e-03, 5.4194313e-03, -5.7048267e-03],
[ 8.6932711e-02, 3.7259728e-02, -4.6791013e-02],
[ 5.3583842e-02, 2.2310656e-02, -2.8512705e-02],
[-7.2550736e-03, 3.5648263e-05, 2.3372378e-03],
[-6.1307452e-04, 2.4406820e-05, 1.9480936e-04],
[-2.6872866e-03, 1.9506851e-04, 7.7207515e-04],
[ 5.5651617e-05, 1.4666974e-04, -9.1101145e-05],
[ 2.0174093e-03, 9.4885059e-04, -1.1280269e-03],
[ 1.5992624e-03, 6.7205599e-04, -8.5408695e-04]], dtype=float32),
14.5: array([[-5.62312156e-02, 6.66079819e-02, 1.78768020e-02],
[-2.88238302e-02, 7.87258521e-02, 1.03787519e-02],
[ 1.06634526e-02, -3.46612073e-02, -1.50409574e-02],
[ 1.42465457e-02, -1.61629319e-02, -1.11742923e-02],
[ 1.14596710e-01, -1.37472108e-01, -8.25117081e-02],
[ 7.20540658e-02, -8.54599476e-02, -5.16028441e-02],
[-1.66496143e-01, -1.74111500e-02, 4.73619476e-02],
[-2.37447545e-02, 3.24901976e-02, 7.44536752e-03],
[-3.88014987e-02, -7.66016692e-02, -7.96739012e-03],
[ 2.24072114e-02, -3.47595327e-02, -1.89395919e-02],
[ 1.72652587e-01, -2.13526696e-01, -1.26149103e-01],
[ 1.22604765e-01, -1.45879999e-01, -8.79283473e-02],
[-1.09902002e-01, -3.37808765e-02, 3.08841653e-02],
[-8.83033779e-03, -7.00296834e-04, 2.46389676e-03],
[-4.40164134e-02, -2.79171262e-02, 8.43676087e-03],
[ 2.08055787e-03, -8.64310376e-03, -3.29670077e-03],
[ 4.02982272e-02, -5.31083420e-02, -3.03843711e-02],
[ 2.59933881e-02, -3.12025528e-02, -1.87201258e-02],
[-8.86945799e-03, -2.86019314e-03, 2.49648746e-03],
[-7.56306516e-04, -2.32408333e-04, 2.11133258e-04],
[-3.61513323e-03, -1.50316081e-03, 9.22265346e-04],
[-1.88939666e-04, -3.18223378e-04, -2.09808259e-05],
[ 7.87009834e-04, -1.42917549e-03, -7.06179359e-04],
[ 7.64912576e-04, -9.45734559e-04, -5.58798725e-04]], dtype=float32),
16.5: array([[-5.21655194e-02, 6.40237182e-02, 1.93627309e-02],
[-3.16902809e-02, 7.26106390e-02, 1.19688921e-02],
[ 4.96370718e-02, -1.94649082e-02, -1.71494689e-02],
[ 3.56320590e-02, -8.14305898e-03, -1.23923291e-02],
[ 2.61876225e-01, -8.06191713e-02, -9.14763287e-02],
[ 1.63702965e-01, -5.00979945e-02, -5.71891367e-02],
[-1.31571516e-01, -2.31336569e-03, 4.78092469e-02],
[-2.17617769e-02, 3.11784036e-02, 8.12539179e-03],
[ 3.48662585e-02, -4.62748185e-02, -1.13952067e-02],
[ 6.08260110e-02, -1.99029520e-02, -2.11712308e-02],
[ 4.00864840e-01, -1.25327602e-01, -1.39987513e-01],
[ 2.78974354e-01, -8.55365321e-02, -9.74565297e-02],
[-8.51455703e-02, -2.16754340e-02, 3.07945814e-02],
[-6.82028104e-03, 1.36852919e-04, 2.48482870e-03],
[-2.05539037e-02, -1.75696015e-02, 7.67942611e-03],
[ 1.09425485e-02, -5.12905791e-03, -3.78289842e-03],
[ 9.68006775e-02, -3.12202554e-02, -3.37842852e-02],
[ 5.94153404e-02, -1.83003880e-02, -2.07544006e-02],
[-6.88259117e-03, -1.87742861e-03, 2.48775375e-03],
[-5.81215310e-04, -1.47543673e-04, 2.10338039e-04],
[-2.47725518e-03, -9.70211113e-04, 9.01394291e-04],
[ 1.08601773e-04, -1.94662134e-04, -3.45606895e-05],
[ 2.27862410e-03, -8.45583505e-04, -7.92958133e-04],
[ 1.77567638e-03, -5.55098406e-04, -6.20092615e-04]], dtype=float32),
19.0: array([[-4.59347032e-02, 6.55248985e-02, 1.53686032e-02],
[-2.22758073e-02, 7.61762410e-02, 7.08056707e-03],
[ 4.11518738e-02, -2.83000842e-02, -1.68732218e-02],
[ 3.16018686e-02, -1.28022414e-02, -1.25626447e-02],
[ 2.31316939e-01, -1.13650046e-01, -9.10786539e-02],
[ 1.44720778e-01, -7.06425831e-02, -5.69585972e-02],
[-1.42962530e-01, -1.11476332e-02, 4.97077629e-02],
[-1.86069719e-02, 3.19416933e-02, 6.12278795e-03],
[ 1.55719062e-02, -6.39298782e-02, -9.23464354e-03],
[ 5.27594611e-02, -2.85367202e-02, -2.10333858e-02],
[ 3.53293657e-01, -1.76572070e-01, -1.39263317e-01],
[ 2.46568039e-01, -1.20595016e-01, -9.70534384e-02],
[-9.56869349e-02, -2.87563242e-02, 3.35070267e-02],
[-7.41398940e-03, -3.52778676e-04, 2.56207446e-03],
[-2.81384345e-02, -2.36046724e-02, 9.07868054e-03],
[ 8.90567061e-03, -7.17244949e-03, -3.66219599e-03],
[ 8.49151909e-02, -4.39383462e-02, -3.35520916e-02],
[ 5.24794050e-02, -2.57963967e-02, -2.06635520e-02],
[-7.74906436e-03, -2.45229690e-03, 2.71650450e-03],
[-6.54285599e-04, -1.97177767e-04, 2.28761128e-04],
[-2.91044964e-03, -1.28153292e-03, 1.00144395e-03],
[ 2.85412298e-05, -2.66606890e-04, -2.47465268e-05],
[ 1.95273012e-03, -1.18478341e-03, -7.80861068e-04],
[ 1.56498631e-03, -7.82061426e-04, -6.16887526e-04]], dtype=float32),
21.0: array([[-5.34694009e-02, 6.47808835e-02, 1.88612193e-02],
[-3.00062969e-02, 7.37764686e-02, 1.05850603e-02],
[ 3.36500965e-02, -2.10885629e-02, -1.21319611e-02],
[ 2.69101467e-02, -8.98332428e-03, -9.65036079e-03],
[ 2.01657698e-01, -8.67570490e-02, -7.22247586e-02],
[ 1.26233414e-01, -5.39153367e-02, -4.52082418e-02],
[-1.46374479e-01, -3.78774083e-03, 5.16380034e-02],
[-2.23854743e-02, 3.15522030e-02, 7.89120048e-03],
[ 4.36447095e-03, -4.95723784e-02, -1.94246427e-03],
[ 4.51014526e-02, -2.14974973e-02, -1.61841102e-02],
[ 3.07534784e-01, -1.34851545e-01, -1.10164531e-01],
[ 2.15042859e-01, -9.20510143e-02, -7.70150423e-02],
[-9.57809612e-02, -2.30147745e-02, 3.37948576e-02],
[-7.66737340e-03, 5.71863129e-05, 2.70362059e-03],
[-3.04013602e-02, -1.87080652e-02, 1.06433565e-02],
[ 7.30020553e-03, -5.50947059e-03, -2.63362238e-03],
[ 7.36833066e-02, -3.35846543e-02, -2.64044777e-02],
[ 4.57499400e-02, -1.96933504e-02, -1.63856242e-02],
[-7.73749081e-03, -1.98709266e-03, 2.73023639e-03],
[-6.56313612e-04, -1.56898532e-04, 2.31537328e-04],
[-2.96092848e-03, -1.02933147e-03, 1.04276650e-03],
[-1.47705596e-05, -2.08145095e-04, 3.63551680e-06],
[ 1.66721246e-03, -9.08721529e-04, -5.98590472e-04],
[ 1.36231352e-03, -5.97280334e-04, -4.88004356e-04]], dtype=float32),
23.0: array([[-4.74029854e-02, 7.53735080e-02, 1.37492567e-02],
[-2.56043747e-02, 8.14283863e-02, 5.38585661e-03],
[ 4.71761115e-02, 7.36717833e-03, -1.90913789e-02],
[ 3.46303321e-02, 7.40389153e-03, -1.39468461e-02],
[ 2.53041953e-01, 2.30490603e-02, -1.00044124e-01],
[ 1.58222675e-01, 1.44590922e-02, -6.25477061e-02],
[-1.35022387e-01, 1.60772149e-02, 4.91505004e-02],
[-1.93890817e-02, 3.68778966e-02, 5.29493392e-03],
[ 2.85062119e-02, 7.61373376e-05, -1.25071043e-02],
[ 5.85115030e-02, 6.98140077e-03, -2.33259350e-02],
[ 3.87067616e-01, 3.50093320e-02, -1.53090373e-01],
[ 2.69613773e-01, 2.45837457e-02, -1.06584072e-01],
[-8.92505273e-02, -1.16661489e-02, 3.36325280e-02],
[-6.98387204e-03, 1.27635687e-03, 2.51374068e-03],
[-2.32962575e-02, -4.76788869e-03, 8.46205000e-03],
[ 1.03155095e-02, 8.36531573e-04, -4.14155424e-03],
[ 9.33317393e-02, 8.33029673e-03, -3.69434394e-02],
[ 5.74099235e-02, 5.22341300e-03, -2.26977803e-02],
[-7.22499518e-03, -1.09984435e-03, 2.73128413e-03],
[-6.09428389e-04, -7.47932936e-05, 2.29248981e-04],
[-2.64250208e-03, -4.38614137e-04, 9.91008594e-04],
[ 8.17513646e-05, -1.02771046e-05, -3.73827461e-05],
[ 2.18107807e-03, 1.81943236e-04, -8.66769929e-04],
[ 1.71456963e-03, 1.55049827e-04, -6.78130251e-04]], dtype=float32),
24.5: array([[-5.21849580e-02, 6.94204792e-02, 1.71629451e-02],
[-3.21539268e-02, 7.88262188e-02, 9.97947436e-03],
[ 5.13067357e-02, -1.69938095e-02, -2.10437737e-02],
[ 3.64992805e-02, -6.31976733e-03, -1.47058554e-02],
[ 2.67939478e-01, -7.00942576e-02, -1.06694013e-01],
[ 1.67472646e-01, -4.35226820e-02, -6.66688383e-02],
[-1.29481524e-01, -3.07125528e-03, 4.55650426e-02],
[-2.17853710e-02, 3.38995270e-02, 7.01678824e-03],
[ 3.83132249e-02, -4.40529101e-02, -1.78156458e-02],
[ 6.24298416e-02, -1.72355156e-02, -2.51053050e-02],
[ 4.10281867e-01, -1.09184913e-01, -1.63502961e-01],
[ 2.85407811e-01, -7.43323565e-02, -1.13625064e-01],
[-8.35690200e-02, -2.40076222e-02, 2.98420731e-02],
[-6.70464896e-03, 1.39366050e-04, 2.33830395e-03],
[-1.93141419e-02, -1.79511402e-02, 6.05830597e-03],
[ 1.13280043e-02, -4.64617414e-03, -4.63960925e-03],
[ 9.91430134e-02, -2.73049455e-02, -3.95746604e-02],
[ 6.07913546e-02, -1.59128401e-02, -2.42074095e-02],
[-6.75499719e-03, -2.07962329e-03, 2.41676695e-03],
[-5.70183678e-04, -1.62928642e-04, 2.03180927e-04],
[-2.41058297e-03, -1.03822164e-03, 8.42074689e-04],
[ 1.22699304e-04, -1.87283847e-04, -5.98858605e-05],
[ 2.34167930e-03, -7.51410960e-04, -9.42234590e-04],
[ 1.81738404e-03, -4.83597483e-04, -7.24245561e-04]], dtype=float32),
27.0: array([[-2.57173806e-01, -1.87496886e-01, 1.99893862e-01],
[-2.37589449e-01, -1.88666657e-01, 1.96282163e-01],
[ 1.76368896e-02, -1.64031535e-02, 7.41399417e-04],
[ 1.31711755e-02, -1.28956996e-02, 2.12298054e-03],
[ 1.80815890e-01, -2.12673601e-02, -5.68992868e-02],
[ 1.13264076e-01, -1.31841954e-02, -3.56484838e-02],
[-2.50186354e-01, -1.07096143e-01, 1.38054252e-01],
[-1.20627031e-01, -9.01997611e-02, 9.53026861e-02],
[-1.47120915e-02, -2.48913001e-02, 1.05036059e-02],
[ 3.44923511e-02, -1.07087567e-02, -7.77327176e-03],
[ 2.75086939e-01, -3.34226526e-02, -8.64328891e-02],
[ 1.92997232e-01, -2.24427860e-02, -6.07887171e-02],
[-1.10545851e-01, -2.24002711e-02, 4.24674936e-02],
[-1.42064542e-02, -6.62138034e-03, 8.20749160e-03],
[-3.68605964e-02, -1.04015870e-02, 1.39667550e-02],
[ 5.55636175e-03, -2.04444537e-03, -1.42135913e-03],
[ 6.55401051e-02, -8.53826851e-03, -2.05066241e-02],
[ 4.10329551e-02, -4.81561199e-03, -1.29195396e-02],
[-8.54563341e-03, -1.46173069e-03, 3.08625586e-03],
[-7.71460822e-04, -1.66671278e-04, 3.03158478e-04],
[-3.33251990e-03, -6.56519784e-04, 1.21027674e-03],
[-7.84134245e-05, -9.06288915e-05, 4.08598135e-05],
[ 1.43937476e-03, -2.55599472e-04, -4.39950789e-04],
[ 1.21870544e-03, -1.47920626e-04, -3.82983300e-04]], dtype=float32),
30.0: array([[-5.28024249e-02, 7.60688484e-02, 1.64442603e-02],
[-3.38067524e-02, 7.70589113e-02, 9.41110589e-03],
[ 5.30809574e-02, 3.37698162e-02, -2.27727070e-02],
[ 3.72355320e-02, 2.17467919e-02, -1.57114826e-02],
[ 2.73402005e-01, 1.22621648e-01, -1.13403164e-01],
[ 1.70855373e-01, 7.64138550e-02, -7.08471164e-02],
[-1.24901772e-01, 4.02735248e-02, 4.44496274e-02],
[-2.21568402e-02, 3.71960141e-02, 6.65420666e-03],
[ 4.30909172e-02, 5.06936684e-02, -2.07765941e-02],
[ 6.39678538e-02, 3.29606943e-02, -2.68423595e-02],
[ 4.18853760e-01, 1.89344183e-01, -1.73877582e-01],
[ 2.91187584e-01, 1.30294040e-01, -1.20752156e-01],
[-7.98985586e-02, 6.16739737e-03, 2.92464122e-02],
[-6.46356540e-03, 2.65443302e-03, 2.26853252e-03],
[-1.70934908e-02, 1.16527323e-02, 5.25502907e-03],
[ 1.17585678e-02, 6.86789863e-03, -5.02374768e-03],
[ 1.01318538e-01, 4.65637632e-02, -4.21327874e-02],
[ 6.20313548e-02, 2.78195776e-02, -2.57298071e-02],
[-6.45545637e-03, 3.36756289e-04, 2.37066159e-03],
[-5.44800190e-04, 5.09952224e-05, 1.98832306e-04],
[-2.27061426e-03, 3.71227128e-04, 8.09494290e-04],
[ 1.42849429e-04, 1.94648281e-04, -7.16485447e-05],
[ 2.40512588e-03, 1.19376765e-03, -1.00858230e-03],
[ 1.85534544e-03, 8.38605920e-04, -7.70195853e-04]], dtype=float32),
43.0: array([[-8.31093043e-02, 7.33238608e-02, 2.66411006e-02],
[-5.87848537e-02, 8.07350576e-02, 1.65618379e-02],
[ 7.32064098e-02, 2.20983401e-02, -3.86417955e-02],
[ 5.06204143e-02, 1.66853424e-02, -2.60952786e-02],
[ 3.75989616e-01, 8.76939967e-02, -1.88348755e-01],
[ 2.34931186e-01, 5.47837503e-02, -1.17633246e-01],
[-1.69423997e-01, 7.91978277e-03, 6.57091364e-02],
[-3.56833003e-02, 3.63395847e-02, 1.08869895e-02],
[ 6.30458519e-02, 1.99661292e-02, -3.89608741e-02],
[ 8.79732370e-02, 2.29629278e-02, -4.48813774e-02],
[ 5.76231122e-01, 1.34535596e-01, -2.89020211e-01],
[ 4.00411844e-01, 9.33396742e-02, -2.00511187e-01],
[-1.04767330e-01, -1.75298117e-02, 4.25864235e-02],
[-8.82458221e-03, 9.16570076e-04, 3.34213162e-03],
[-2.08945833e-02, -2.57997052e-03, 5.81681542e-03],
[ 1.63566824e-02, 4.10402101e-03, -8.56631808e-03],
[ 1.39492139e-01, 3.26525830e-02, -7.01491609e-02],
[ 8.53084996e-02, 1.98913999e-02, -4.27347533e-02],
[-8.44082795e-03, -1.58423057e-03, 3.45049263e-03],
[-7.14718481e-04, -1.13302092e-04, 2.88988434e-04],
[-2.93271802e-03, -5.25636307e-04, 1.13390712e-03],
[ 2.14069485e-04, 6.55682597e-05, -1.38651434e-04],
[ 3.32340202e-03, 7.87954370e-04, -1.69261871e-03],
[ 2.55245715e-03, 5.95884223e-04, -1.28021394e-03]], dtype=float32),
50.0: array([[-1.50551707e-01, -2.43838802e-02, 1.08773768e-01],
[-1.28343940e-01, -2.02106275e-02, 1.02698557e-01],
[ 2.40156706e-02, -1.00590494e-02, -7.73886545e-03],
[ 1.93407573e-02, -5.09899762e-03, -5.04588755e-03],
[ 1.84579939e-01, -2.70465538e-02, -7.31809214e-02],
[ 1.15611807e-01, -1.67552438e-02, -4.58031893e-02],
[-1.98662132e-01, -3.46042514e-02, 9.51672420e-02],
[-6.91822022e-02, -1.14235841e-02, 5.12526631e-02],
[-8.98227189e-03, -2.40133740e-02, 1.39843882e-03],
[ 3.81425358e-02, -8.26211832e-03, -1.41441021e-02],
[ 2.80963778e-01, -4.24203761e-02, -1.11578934e-01],
[ 1.96957052e-01, -2.86140051e-02, -7.80666545e-02],
[-1.04935400e-01, -1.86039899e-02, 3.86048928e-02],
[-1.09383631e-02, -1.96439447e-03, 5.45886531e-03],
[-3.50670815e-02, -1.08874179e-02, 1.18733337e-02],
[ 6.00524805e-03, -2.15752306e-03, -2.47487845e-03],
[ 6.70347661e-02, -1.07601946e-02, -2.67064683e-02],
[ 4.18803990e-02, -6.13775104e-03, -1.66081954e-02],
[-8.29413999e-03, -1.46537297e-03, 2.94940220e-03],
[-7.25980557e-04, -1.30383080e-04, 2.70306773e-04],
[-3.22573446e-03, -6.83001708e-04, 1.12806552e-03],
[-6.25675675e-05, -9.90892368e-05, 1.11888103e-05],
[ 1.48356601e-03, -3.13210301e-04, -6.00900443e-04],
[ 1.24468061e-03, -1.87850033e-04, -4.94315405e-04]], dtype=float32),
52.0: array([[-4.91133891e-02, 8.06416720e-02, 1.11895762e-02],
[-2.32583582e-02, 8.30661356e-02, 3.09890253e-03],
[ 2.30978876e-02, 3.18676680e-02, -2.17700694e-02],
[ 2.13583875e-02, 2.11142786e-02, -1.55127123e-02],
[ 1.61383122e-01, 1.16389431e-01, -1.09977141e-01],
[ 1.01179324e-01, 7.25579783e-02, -6.87318444e-02],
[-1.55023068e-01, 3.61581333e-02, 4.63572629e-02],
[-2.02549938e-02, 3.95011269e-02, 4.04526154e-03],
[-1.65471025e-02, 4.49683331e-02, -1.71896312e-02],
[ 3.46710049e-02, 3.12764384e-02, -2.59507596e-02],
[ 2.45080695e-01, 1.79544538e-01, -1.68463528e-01],
[ 1.72290042e-01, 1.23702593e-01, -1.17132947e-01],
[-1.03591882e-01, 1.60352362e-03, 3.23515460e-02],
[-8.14015046e-03, 2.45829555e-03, 2.34287395e-03],
[-3.73880826e-02, 8.89103487e-03, 7.09555717e-03],
[ 4.84012673e-03, 6.36407034e-03, -4.72561875e-03],
[ 5.81976064e-02, 4.40674983e-02, -4.07409742e-02],
[ 3.66099738e-02, 2.64045410e-02, -2.49516610e-02],
[-8.37621372e-03, -4.31097687e-05, 2.63336883e-03],
[-7.11010653e-04, 1.96915262e-05, 2.19969777e-04],
[-3.31382640e-03, 1.94715147e-04, 9.30072332e-04],
[-9.99785334e-05, 1.70125524e-04, -5.59964938e-05],
[ 1.25583389e-03, 1.11997011e-03, -9.66048159e-04],
[ 1.08570023e-03, 7.95206113e-04, -7.46218371e-04]], dtype=float32),
54.0: array([[-5.5085484e-02, 7.2300032e-02, 1.5592797e-02],
[-3.3175237e-02, 7.7256247e-02, 9.8006269e-03],
[ 3.7426513e-02, 9.2738131e-03, -2.9302945e-02],
[ 2.8677396e-02, 8.1656072e-03, -1.9272178e-02],
[ 2.1467881e-01, 2.9582797e-02, -1.3787527e-01],
[ 1.3431680e-01, 1.8509412e-02, -8.6073555e-02],
[-1.3995038e-01, 1.9476853e-02, 3.8111515e-02],
[-2.3250423e-02, 3.5326723e-02, 6.2490362e-03],
[ 1.3234792e-02, 5.1563489e-03, -3.3280060e-02],
[ 4.8603624e-02, 8.7274192e-03, -3.3246834e-02],
[ 3.2783690e-01, 4.5232575e-02, -2.1181096e-01],
[ 2.2884470e-01, 3.1503078e-02, -1.4673212e-01],
[-9.0418272e-02, -8.1247929e-03, 2.4779795e-02],
[-7.3270537e-03, 1.4438650e-03, 1.9057610e-03],
[-2.6757060e-02, -2.4985259e-03, 1.1864156e-03],
[ 8.2008038e-03, 1.3191983e-03, -6.5094307e-03],
[ 7.8771815e-02, 1.0910049e-02, -5.1531296e-02],
[ 4.8705336e-02, 6.7066327e-03, -3.1283259e-02],
[-7.2992793e-03, -8.0627087e-04, 2.0120780e-03],
[-6.1911909e-04, -5.0421251e-05, 1.6729841e-04],
[-2.7465820e-03, -2.9898912e-04, 6.0813391e-04],
[ 2.2129523e-05, 1.1209732e-05, -1.2223655e-04],
[ 1.8085424e-03, 2.5553169e-04, -1.2574503e-03],
[ 1.4522284e-03, 2.0032671e-04, -9.3820412e-04]], dtype=float32),
57.0: array([[-4.74663042e-02, 6.55223355e-02, 1.67945623e-02],
[-2.67220959e-02, 7.58264884e-02, 9.49087087e-03],
[ 5.36576807e-02, -2.60515325e-02, -2.12841835e-02],
[ 3.82074006e-02, -1.15551874e-02, -1.48937749e-02],
[ 2.77929246e-01, -1.04983807e-01, -1.07841186e-01],
[ 1.73712447e-01, -6.52483255e-02, -6.73862994e-02],
[-1.30135700e-01, -9.53046046e-03, 4.59290966e-02],
[-1.94026232e-02, 3.19482908e-02, 6.82266522e-03],
[ 4.05535959e-02, -5.97872846e-02, -1.79493669e-02],
[ 6.49608597e-02, -2.62938403e-02, -2.53841225e-02],
[ 4.25611168e-01, -1.63153216e-01, -1.65257126e-01],
[ 2.96041012e-01, -1.11392029e-01, -1.14847325e-01],
[-8.56051072e-02, -2.75576115e-02, 3.02417297e-02],
[-6.70109084e-03, -2.58593063e-04, 2.35386821e-03],
[-1.95850823e-02, -2.23470535e-02, 6.16824906e-03],
[ 1.17927799e-02, -6.66065235e-03, -4.68676537e-03],
[ 1.02864698e-01, -4.06208336e-02, -3.99974138e-02],
[ 6.30575567e-02, -2.38297917e-02, -2.44676396e-02],
[-6.93166209e-03, -2.35595321e-03, 2.45023984e-03],
[-5.83590474e-04, -1.88665697e-04, 2.05872653e-04],
[-2.46850424e-03, -1.22345448e-03, 8.54415179e-04],
[ 1.30317349e-04, -2.49929377e-04, -6.02282453e-05],
[ 2.43161572e-03, -1.09775004e-03, -9.52090544e-04],
[ 1.88528211e-03, -7.22628203e-04, -7.32015527e-04]], dtype=float32),
65.5: array([[-5.45864515e-02, 7.19663873e-02, 1.68590024e-02],
[-2.76965071e-02, 7.94319510e-02, 6.84002042e-03],
[ 1.89041682e-02, -6.13191212e-03, -5.49137872e-03],
[ 1.90771762e-02, -3.42367013e-04, -6.22080313e-03],
[ 1.47446305e-01, -2.94366311e-02, -4.75542843e-02],
[ 9.25176144e-02, -1.82236247e-02, -2.98675653e-02],
[-1.63215026e-01, 7.55440770e-03, 5.86733110e-02],
[-2.28474382e-02, 3.51275243e-02, 6.87859813e-03],
[-2.51041297e-02, -2.34901085e-02, 1.16982860e-02],
[ 3.08185369e-02, -6.58289297e-03, -9.72666033e-03],
[ 2.23410234e-01, -4.61754799e-02, -7.18655288e-02],
[ 1.57508448e-01, -3.11675314e-02, -5.08345887e-02],
[-1.08058140e-01, -1.68717541e-02, 3.96961719e-02],
[-8.61234870e-03, 7.53248343e-04, 3.08720977e-03],
[-4.05845456e-02, -1.13435620e-02, 1.54597256e-02],
[ 3.92091693e-03, -2.18805484e-03, -1.08500372e-03],
[ 5.27941845e-02, -1.16995703e-02, -1.68870464e-02],
[ 3.34473997e-02, -6.68651704e-03, -1.07867820e-02],
[-8.72755144e-03, -1.50928146e-03, 3.21124215e-03],
[-7.42514268e-04, -1.12427348e-04, 2.72768433e-04],
[-3.49171809e-03, -7.14770751e-04, 1.29770511e-03],
[-1.34732982e-04, -1.04261424e-04, 5.94535886e-05],
[ 1.10887107e-03, -3.38869577e-04, -3.43379943e-04],
[ 9.89726163e-04, -2.04526586e-04, -3.18377715e-04]], dtype=float32),
93.0: array([[-5.47065213e-02, 8.76432136e-02, 1.10226702e-02],
[-2.82389317e-02, 8.52111652e-02, 3.18982336e-03],
[ 1.20213795e-02, 6.32044673e-02, -2.21188050e-02],
[ 1.47918509e-02, 3.85870300e-02, -1.56444255e-02],
[ 1.17953427e-01, 2.35274523e-01, -1.10865153e-01],
[ 7.41251633e-02, 1.46551743e-01, -6.92796782e-02],
[-1.61074236e-01, 6.30297214e-02, 4.50209081e-02],
[-2.30889767e-02, 4.29594107e-02, 3.98615375e-03],
[-3.48297171e-02, 1.02813758e-01, -1.82399936e-02],
[ 2.34617442e-02, 6.22695982e-02, -2.62243338e-02],
[ 1.77971140e-01, 3.63669395e-01, -1.69870868e-01],
[ 1.26146346e-01, 2.49941081e-01, -1.18069865e-01],
[-1.06256828e-01, 1.94018632e-02, 3.13872769e-02],
[-8.53617862e-03, 4.03262582e-03, 2.27144687e-03],
[-4.19964790e-02, 2.67219469e-02, 6.55847788e-03],
[ 2.39738869e-03, 1.34377452e-02, -4.80935024e-03],
[ 4.16734256e-02, 8.96105692e-02, -4.11049873e-02],
[ 2.67554075e-02, 5.33824638e-02, -2.51532663e-02],
[-8.57576728e-03, 1.37539778e-03, 2.55522272e-03],
[-7.31025357e-04, 1.46225808e-04, 2.13297055e-04],
[-3.48186400e-03, 1.03273895e-03, 8.94600875e-04],
[-1.72130458e-04, 4.02828999e-04, -6.04612251e-05],
[ 8.29916156e-04, 2.31731404e-03, -9.77417687e-04],
[ 7.88463338e-04, 1.61070633e-03, -7.52450200e-04]], dtype=float32)}}
[26]:
vel_week
[26]:
{'01': {-3.0: array([[ 0.00220195, 0.05142251, -0.04584359],
[ 0.00203441, 0.04553534, -0.0403849 ],
[ 0.01117069, 0.04208901, -0.01480857],
[ 0.01141505, 0.04095323, -0.01308329],
[ 0.02461177, 0.05242203, 0.0074419 ],
[ 0.02654672, 0.05643063, 0.00816343],
[-0.00952978, 0.04259073, -0.04397725],
[-0.00891005, 0.03932984, -0.04048735],
[ 0.00771329, 0.03920534, -0.01333026],
[ 0.0095078 , 0.04198331, -0.01236233],
[ 0.0284091 , 0.05524097, 0.00706629],
[ 0.03281986, 0.06477229, 0.00868083],
[-0.07927734, -0.06910087, 0.02113093],
[-0.07929646, -0.06957772, 0.02156907],
[-0.02039247, -0.01452065, 0.0096149 ],
[-0.01966953, -0.01373022, 0.00965709],
[ 0.03515238, 0.0403994 , 0.00036235],
[ 0.03712611, 0.04236589, 0.00033238],
[-0.07195229, -0.06915385, 0.02503579],
[-0.07260469, -0.0697768 , 0.0252749 ],
[-0.01597055, -0.01454593, 0.00938113],
[-0.01794774, -0.01642972, 0.01009707],
[ 0.03159 , 0.03254084, -0.00080494],
[ 0.03221098, 0.03318913, -0.00083092]], dtype=float32),
-0.5: array([[ 0.28073624, -0.03771748, -0.11910288],
[ 0.24700834, -0.03267367, -0.10466966],
[ 0.05790733, 0.04720761, -0.01190824],
[ 0.04631573, 0.05051434, -0.00656617],
[-0.12378439, 0.14774281, 0.08263253],
[-0.13438693, 0.15963137, 0.08960738],
[ 0.29125878, -0.04414172, -0.10352459],
[ 0.26826987, -0.04019475, -0.09480918],
[ 0.05455592, 0.0418247 , -0.01022123],
[ 0.04329734, 0.05208888, -0.00346208],
[-0.1292791 , 0.13320395, 0.06011926],
[-0.15347496, 0.16215703, 0.07757822],
[ 0.01822277, -0.03674002, 0.12647147],
[ 0.01552985, -0.03624805, 0.12773278],
[-0.02790752, -0.02971402, 0.01014938],
[-0.02981752, -0.02817792, 0.01024408],
[-0.08570763, -0.03363552, -0.11327022],
[-0.09000213, -0.03766919, -0.12163616],
[-0.01983483, -0.0282509 , 0.12986222],
[-0.02010593, -0.02859365, 0.13095057],
[-0.03281767, -0.03236318, 0.00248251],
[-0.03357714, -0.03328279, 0.00589831],
[-0.06701682, -0.05722631, -0.12735455],
[-0.06826331, -0.05816635, -0.12967601]], dtype=float32),
1.0: array([[-0.07460082, 0.04267115, 0.11502326],
[-0.06583485, 0.03740777, 0.10118715],
[-0.03684809, -0.00640993, 0.02233634],
[-0.03444544, -0.00862657, 0.017527 ],
[-0.01831101, -0.05514305, -0.05398384],
[-0.01955613, -0.05956712, -0.05864105],
[-0.0625016 , 0.06022805, 0.10531081],
[-0.05748354, 0.05565019, 0.09669446],
[-0.03049582, -0.00813216, 0.02042631],
[-0.03117988, -0.01185694, 0.0155712 ],
[-0.01779092, -0.07733084, -0.04053027],
[-0.02097839, -0.08658872, -0.05201016],
[ 0.0954111 , 0.10972526, -0.08650919],
[ 0.09609711, 0.10939999, -0.08768099],
[ 0.03997931, -0.01910293, -0.01180631],
[ 0.03941264, -0.02051633, -0.01193943],
[-0.00483584, -0.16722679, 0.06600493],
[-0.00483168, -0.17758048, 0.0709202 ],
[ 0.09587596, 0.09382733, -0.09316549],
[ 0.09679447, 0.09452913, -0.09396603],
[ 0.0363256 , -0.02375211, -0.00786869],
[ 0.03906321, -0.02148369, -0.0103578 ],
[-0.00197174, -0.16016097, 0.07513513],
[-0.00207357, -0.16311151, 0.07651208]], dtype=float32),
1.5: array([[ 0.01198503, -0.07364011, 0.1715321 ],
[ 0.01013277, -0.06562266, 0.15104966],
[-0.04204061, -0.10517716, 0.04975571],
[-0.04393545, -0.10496316, 0.0430996 ],
[-0.11144935, -0.18219419, -0.04126779],
[-0.12032802, -0.19642787, -0.04510638],
[ 0.02328905, -0.03880111, 0.14797136],
[ 0.02121852, -0.03588759, 0.13582045],
[-0.03947716, -0.09426212, 0.04445742],
[-0.046776 , -0.10548521, 0.03986024],
[-0.11318596, -0.19016603, -0.01851068],
[-0.1341956 , -0.22382155, -0.02837811],
[ 0.07397631, 0.24833524, -0.19007659],
[ 0.07382244, 0.24897796, -0.19181496],
[ 0.00963856, 0.05791827, -0.03070158],
[ 0.00814917, 0.05489119, -0.03015872],
[-0.05745832, -0.12817045, 0.13235651],
[-0.05986387, -0.13423876, 0.14141463],
[ 0.06535767, 0.23407815, -0.19415094],
[ 0.06593464, 0.2362006 , -0.19584042],
[ 0.01010049, 0.05275662, -0.02256042],
[ 0.01184233, 0.05916202, -0.02777428],
[-0.04047197, -0.10069698, 0.14010946],
[-0.04128723, -0.10272542, 0.14271699]], dtype=float32),
2.0: array([[-0.51738787, -0.2135003 , 0.19766286],
[-0.45481804, -0.18770301, 0.17375314],
[-0.06450903, -0.02807002, 0.02398442],
[-0.04174575, -0.01875373, 0.01526866],
[ 0.32859564, 0.13223207, -0.12711814],
[ 0.35635057, 0.14331275, -0.13784164],
[-0.48149318, -0.22834088, 0.18608071],
[-0.4419133 , -0.21029516, 0.17082854],
[-0.0575492 , -0.02716387, 0.02165513],
[-0.03025584, -0.0161023 , 0.01110851],
[ 0.26964584, 0.13866726, -0.10621264],
[ 0.33740216, 0.16413175, -0.1322993 ],
[ 0.336564 , -0.05987957, -0.11428053],
[ 0.34186321, -0.05782326, -0.11629946],
[ 0.04808851, 0.01328211, -0.01664611],
[ 0.04978728, 0.01537152, -0.01740722],
[-0.25116906, 0.09572912, 0.08474475],
[-0.27132857, 0.10052381, 0.09173276],
[ 0.37151664, -0.02653396, -0.12899971],
[ 0.37470207, -0.02669662, -0.13010615],
[ 0.02925353, 0.01817899, -0.01032424],
[ 0.03919264, 0.01763636, -0.01377901],
[-0.3052895 , 0.07580696, 0.10556553],
[-0.31083015, 0.07724248, 0.10747546]], dtype=float32),
3.0: array([[-0.4192432 , -0.05281937, 0.3315045 ],
[-0.36919716, -0.04756229, 0.291863 ],
[-0.1220516 , -0.12915832, 0.0897274 ],
[-0.10584736, -0.1306855 , 0.07667641],
[ 0.09993222, -0.25884607, -0.09512315],
[ 0.10908147, -0.27919182, -0.10371657],
[-0.39815384, 0.00069388, 0.29550153],
[-0.36633602, 0.0005689 , 0.2713986 ],
[-0.10834516, -0.11501554, 0.07861448],
[-0.09750479, -0.13152218, 0.06893698],
[ 0.0908639 , -0.2766847 , -0.06455931],
[ 0.11199543, -0.32431507, -0.08572568],
[ 0.22010036, 0.36996928, -0.30349138],
[ 0.22414269, 0.37041986, -0.30679518],
[ 0.08995918, 0.07247755, -0.06893626],
[ 0.09042945, 0.06790786, -0.06850549],
[-0.02028055, -0.22429943, 0.16198938],
[-0.02248711, -0.23573019, 0.17351109],
[ 0.25324807, 0.34139594, -0.3174008 ],
[ 0.25563273, 0.34444383, -0.32022566],
[ 0.08437625, 0.06239503, -0.05583541],
[ 0.09155554, 0.07164239, -0.06449388],
[-0.03618557, -0.18566494, 0.1784844 ],
[-0.03693981, -0.1893155 , 0.18182886]], dtype=float32),
9.0: array([[-0.44219372, -0.16419871, 0.29967394],
[-0.38918525, -0.14469594, 0.26334012],
[-0.10544662, -0.06015127, 0.02864351],
[-0.08760323, -0.05412227, 0.01514607],
[ 0.16086541, 0.00950157, -0.2109893 ],
[ 0.17498945, 0.01095942, -0.22884044],
[-0.40001127, -0.09890924, 0.2375415 ],
[-0.36737618, -0.08989452, 0.21693775],
[-0.08871734, -0.05429434, 0.02028356],
[-0.07394607, -0.05154561, 0.00337569],
[ 0.13115512, -0.06092623, -0.13773616],
[ 0.16514757, -0.05421489, -0.1832539 ],
[ 0.36568934, 0.4684062 , -0.47209734],
[ 0.37005457, 0.47027045, -0.47533312],
[ 0.10902346, 0.02223282, -0.04928144],
[ 0.10906821, 0.01936918, -0.04849917],
[-0.13003267, -0.45641744, 0.39230704],
[-0.14014085, -0.4863716 , 0.42039755],
[ 0.3879004 , 0.44509783, -0.46591085],
[ 0.39142993, 0.44879717, -0.46985182],
[ 0.09131657, -0.00116504, -0.01926715],
[ 0.1020669 , 0.01041489, -0.03157732],
[-0.15653479, -0.46226248, 0.42921263],
[-0.15949887, -0.47080657, 0.43708548]], dtype=float32),
12.0: array([[-0.20018655, 0.18169689, 0.05883455],
[-0.17595837, 0.15914126, 0.05193511],
[-0.02361939, -0.03987725, 0.03058307],
[-0.01474537, -0.0498516 , 0.02873255],
[ 0.13026504, -0.26450303, 0.01807519],
[ 0.14133263, -0.28612733, 0.01933656],
[-0.16478564, 0.17448282, 0.0489029 ],
[-0.15071103, 0.15962686, 0.04498291],
[-0.0199294 , -0.03828522, 0.0264424 ],
[-0.00913369, -0.0575407 , 0.02710991],
[ 0.08392114, -0.23326634, 0.02040865],
[ 0.11215021, -0.285881 , 0.02344488],
[ 0.27439743, -0.08139507, -0.07781948],
[ 0.27654916, -0.08336677, -0.0783737 ],
[ 0.0203211 , 0.00258479, -0.02498581],
[ 0.01996894, 0.00049891, -0.02447159],
[-0.2494637 , 0.09510469, 0.02315623],
[-0.26729196, 0.10448116, 0.02438855],
[ 0.27420673, -0.09723233, -0.07796554],
[ 0.27650094, -0.09799472, -0.07868867],
[ 0.00395231, 0.01398292, -0.02245284],
[ 0.01114682, 0.01148363, -0.02462958],
[-0.27233735, 0.13763542, 0.02051914],
[-0.277321 , 0.1400381 , 0.02094205]], dtype=float32),
14.0: array([[-0.08051723, -0.0327426 , 0.06664983],
[-0.07109749, -0.02871589, 0.05865931],
[-0.04541199, 0.00262588, 0.01622847],
[-0.04296685, 0.00431896, 0.01353475],
[-0.03333549, 0.03673425, -0.02340424],
[-0.03564823, 0.03987168, -0.0255153 ],
[-0.02670941, -0.02034902, 0.04689835],
[-0.02372832, -0.01835904, 0.04276315],
[-0.03229035, -0.00454579, 0.01301796],
[-0.03427308, -0.00092428, 0.01082572],
[-0.06443163, -0.00107994, -0.00457289],
[-0.06851917, 0.00827009, -0.01040453],
[ 0.37637392, 0.08864549, -0.14489254],
[ 0.37723047, 0.0891574 , -0.14561114],
[ 0.07678488, -0.05342679, -0.01783838],
[ 0.0741741 , -0.05382123, -0.01719362],
[-0.21770002, -0.23115733, 0.11361033],
[-0.23141973, -0.24684335, 0.12129052],
[ 0.35075927, 0.08465936, -0.13969956],
[ 0.35386336, 0.0851732 , -0.14089419],
[ 0.05521413, -0.05690382, -0.00964773],
[ 0.06471695, -0.05508107, -0.01335578],
[-0.21444406, -0.23920943, 0.11836243],
[-0.21851356, -0.24352723, 0.12055627]], dtype=float32),
15.0: array([[-1.48053512e-01, 4.94064569e-01, 2.36368701e-02],
[-1.30919874e-01, 4.32769388e-01, 2.10968666e-02],
[-1.00938469e-01, -1.05247520e-01, 3.77376676e-02],
[-9.70665663e-02, -1.32239953e-01, 3.77837084e-02],
[-1.02602869e-01, -7.11719096e-01, 6.80055991e-02],
[-1.10405318e-01, -7.69825280e-01, 7.32934251e-02],
[-1.36905730e-01, 5.01043975e-01, -7.58514914e-04],
[-1.26482651e-01, 4.59131062e-01, -8.82418477e-04],
[-8.76417011e-02, -9.94254798e-02, 3.24931107e-02],
[-9.32826996e-02, -1.51295274e-01, 3.67856584e-02],
[-8.13821107e-02, -6.55298531e-01, 8.14336091e-02],
[-1.01453036e-01, -7.95223713e-01, 9.33788121e-02],
[ 1.03405766e-01, -4.30830829e-02, -1.68358132e-01],
[ 1.04653209e-01, -4.83210571e-02, -1.68603718e-01],
[ 8.40532556e-02, 1.12173809e-02, -2.71299500e-02],
[ 8.29946995e-02, 4.33433568e-03, -2.54598614e-02],
[ 9.46829543e-02, 7.56069273e-02, 1.16076186e-01],
[ 1.02347404e-01, 8.90099928e-02, 1.22811861e-01],
[ 1.12486340e-01, -1.03110261e-01, -1.55322626e-01],
[ 1.13705203e-01, -1.03864819e-01, -1.56682298e-01],
[ 8.53831619e-02, 3.06903161e-02, -2.01442037e-02],
[ 8.88626724e-02, 2.80475747e-02, -2.43086070e-02],
[ 1.11776046e-01, 1.88344926e-01, 1.06459759e-01],
[ 1.13663167e-01, 1.91453889e-01, 1.08491853e-01]], dtype=float32),
17.0: array([[-0.35058346, 0.3329043 , 0.0137412 ],
[-0.3092446 , 0.2921714 , 0.01227193],
[-0.15701692, -0.01040395, 0.02301196],
[-0.14524022, -0.02662137, 0.02306052],
[-0.04746852, -0.33533093, 0.04212404],
[-0.05033231, -0.36284995, 0.04536595],
[-0.31832427, 0.34063938, -0.009092 ],
[-0.293143 , 0.31285244, -0.00864958],
[-0.12934594, -0.02222163, 0.02019953],
[-0.13009594, -0.04646472, 0.02276958],
[-0.02266653, -0.3504065 , 0.06101276],
[-0.0317041 , -0.41423538, 0.06777776],
[ 0.2833245 , -0.01002338, -0.15582904],
[ 0.28647032, -0.01322421, -0.15602012],
[ 0.18073586, -0.09847784, -0.01167197],
[ 0.17957535, -0.10230426, -0.01024948],
[ 0.1391714 , -0.23483022, 0.14082105],
[ 0.14933427, -0.24647588, 0.14952461],
[ 0.3012267 , -0.05189766, -0.14275567],
[ 0.3043043 , -0.05262647, -0.14395837],
[ 0.17195633, -0.0891717 , -0.00439421],
[ 0.18095048, -0.09120911, -0.00813082],
[ 0.14809373, -0.18439347, 0.1355253 ],
[ 0.15054181, -0.1878176 , 0.13805844]], dtype=float32),
19.0: array([[-0.09455644, 0.29851055, 0.13471371],
[-0.08341043, 0.2620168 , 0.1184083 ],
[-0.04296469, -0.00563558, 0.01575014],
[-0.03980618, -0.02010114, 0.00977753],
[-0.01429247, -0.29183292, -0.08800633],
[-0.01517179, -0.31591243, -0.09547284],
[-0.07704781, 0.29885396, 0.110355 ],
[-0.07077732, 0.2743527 , 0.10091425],
[-0.0328183 , -0.00758068, 0.0120823 ],
[-0.03334524, -0.02976166, 0.00492184],
[-0.01081309, -0.27365544, -0.05944307],
[-0.01333531, -0.3311195 , -0.07834792],
[ 0.13557966, -0.0527058 , -0.18830882],
[ 0.13643406, -0.05573613, -0.18974212],
[ 0.06564368, -0.02393635, -0.02206587],
[ 0.06493635, -0.02680261, -0.02182807],
[ 0.0158176 , -0.00071073, 0.15069307],
[ 0.0171787 , 0.00261365, 0.1615537 ],
[ 0.13472126, -0.08633238, -0.1878023 ],
[ 0.13603874, -0.08710409, -0.18939947],
[ 0.05892523, -0.01645526, -0.01040817],
[ 0.06282997, -0.0188802 , -0.01538836],
[ 0.01827607, 0.04520613, 0.1659651 ],
[ 0.01851483, 0.04593372, 0.16901092]], dtype=float32),
22.0: array([[-0.49775335, 0.16868335, 0.22451405],
[-0.43836907, 0.14742331, 0.1977226 ],
[-0.14792642, -0.07227184, 0.06679279],
[-0.12879969, -0.08262144, 0.05814494],
[ 0.11150835, -0.32970542, -0.05005807],
[ 0.12174367, -0.35639048, -0.05473327],
[-0.4922614 , 0.19464125, 0.19946386],
[-0.4533762 , 0.1785835 , 0.18324076],
[-0.13398881, -0.0653277 , 0.05875889],
[-0.12144299, -0.08880357, 0.05314834],
[ 0.11832652, -0.3219459 , -0.03052359],
[ 0.1410015 , -0.38541147, -0.04207325],
[ 0.13025254, 0.14377578, -0.21008189],
[ 0.13497889, 0.1419929 , -0.21230777],
[ 0.09474143, 0.02856167, -0.04937828],
[ 0.09615233, 0.02439209, -0.04893807],
[ 0.08781689, -0.08532841, 0.10807145],
[ 0.09260722, -0.08695165, 0.11558823],
[ 0.18211369, 0.10812669, -0.21906142],
[ 0.18394724, 0.10912353, -0.22101876],
[ 0.09709886, 0.02895462, -0.0408667 ],
[ 0.10251088, 0.03188912, -0.04685448],
[ 0.07109953, -0.0342305 , 0.1169661 ],
[ 0.07231593, -0.03506058, 0.11917011]], dtype=float32),
24.0: array([[ 0.05878175, 0.37557152, -0.08889071],
[ 0.05120701, 0.32911158, -0.07792394],
[-0.0424213 , -0.06519391, 0.01253306],
[-0.04660216, -0.08526184, 0.01718658],
[-0.15592939, -0.50566816, 0.11279979],
[-0.16858485, -0.5470773 , 0.12203307],
[ 0.05151994, 0.36940014, -0.09652848],
[ 0.04669544, 0.33835867, -0.08867995],
[-0.03741305, -0.06034703, 0.00947263],
[-0.0485412 , -0.09770565, 0.01803279],
[-0.12457195, -0.4490391 , 0.10542872],
[-0.15617272, -0.54955035, 0.1275503 ],
[-0.05888567, -0.10972722, -0.03507132],
[-0.05965445, -0.11375775, -0.03414084],
[ 0.03057634, 0.01647656, -0.01564473],
[ 0.02961452, 0.01212402, -0.01426446],
[ 0.14005718, 0.16166703, -0.00164282],
[ 0.1511655 , 0.17827578, -0.00310589],
[-0.06101912, -0.14806971, -0.02014938],
[-0.0613985 , -0.14918828, -0.0203681 ],
[ 0.03859929, 0.03372693, -0.01506035],
[ 0.03723516, 0.02998379, -0.01566284],
[ 0.16595218, 0.24281025, -0.01943559],
[ 0.16888982, 0.24701627, -0.01971926]], dtype=float32),
27.0: array([[-0.45981106, 0.20138167, 0.06745882],
[-0.405632 , 0.17592987, 0.05979374],
[-0.2097947 , -0.09454937, 0.06246585],
[-0.19446287, -0.10711583, 0.0611812 ],
[-0.07142456, -0.4134102 , 0.08616267],
[-0.07590351, -0.44670045, 0.0927152 ],
[-0.43315375, 0.2491603 , 0.02042717],
[-0.3992492 , 0.22893393, 0.01829192],
[-0.18116358, -0.09404217, 0.05252147],
[-0.18198797, -0.12232984, 0.05767327],
[-0.04075605, -0.44540685, 0.11520242],
[-0.05423898, -0.52243495, 0.12945703],
[ 0.2663341 , 0.2837588 , -0.32852072],
[ 0.27048662, 0.28182608, -0.32923368],
[ 0.18559058, -0.0272872 , -0.05311282],
[ 0.18461129, -0.03340921, -0.05030164],
[ 0.16708317, -0.37868547, 0.22585447],
[ 0.17942746, -0.39864632, 0.23951109],
[ 0.30001038, 0.23076457, -0.30585024],
[ 0.3031139 , 0.23258543, -0.30852175],
[ 0.18346733, -0.02970218, -0.03787874],
[ 0.19249558, -0.02399703, -0.04607868],
[ 0.18000631, -0.31728536, 0.21440358],
[ 0.18301687, -0.32329664, 0.21846868]], dtype=float32),
29.0: array([[-0.4005102 , 0.30894297, -0.10763498],
[-0.3531988 , 0.27064303, -0.09417138],
[-0.17050989, -0.0643676 , 0.03592921],
[-0.15674254, -0.08112403, 0.04219102],
[-0.03312076, -0.441771 , 0.18617041],
[-0.03462328, -0.47760978, 0.20119672],
[-0.3592955 , 0.36005738, -0.1500509 ],
[-0.33068544, 0.33114526, -0.13840093],
[-0.14660364, -0.06512623, 0.02802902],
[-0.14527501, -0.09654945, 0.04170308],
[-0.03232286, -0.4757285 , 0.20400454],
[-0.03700366, -0.55843127, 0.23854446],
[ 0.35240394, 0.28602147, -0.26506445],
[ 0.35614777, 0.2830631 , -0.26402545],
[ 0.15112086, -0.03662864, -0.03771642],
[ 0.14961915, -0.04328157, -0.03406437],
[-0.01142525, -0.4027103 , 0.19328536],
[-0.01134351, -0.42396784, 0.20369492],
[ 0.37032467, 0.21888869, -0.2257654 ],
[ 0.37387243, 0.22057217, -0.22772957],
[ 0.14084224, -0.04171887, -0.02512621],
[ 0.15142366, -0.03640699, -0.03113514],
[-0.00624459, -0.3370596 , 0.16577455],
[-0.00659518, -0.34343156, 0.16896673]], dtype=float32),
31.0: array([[-2.5403216e-01, 2.2741334e-01, 2.2207912e-02],
[-2.2462361e-01, 1.9943006e-01, 1.9789523e-02],
[-1.7316157e-01, -2.4185536e-02, 3.2316271e-02],
[-1.6649672e-01, -3.5779927e-02, 3.2248255e-02],
[-1.7608373e-01, -2.6981342e-01, 5.6437414e-02],
[-1.8938933e-01, -2.9179415e-01, 6.0773864e-02],
[-1.9640869e-01, 2.3403136e-01, -9.2913043e-03],
[-1.8077186e-01, 2.1480183e-01, -8.9406027e-03],
[-1.4228705e-01, -3.8025040e-02, 2.7693383e-02],
[-1.5258121e-01, -5.6477647e-02, 3.1169735e-02],
[-1.6414134e-01, -2.9613772e-01, 8.0633931e-02],
[-1.9709951e-01, -3.4648207e-01, 8.9748479e-02],
[ 4.3578359e-01, 1.8969189e-03, -2.1561533e-01],
[ 4.3800837e-01, -2.1760106e-04, -2.1590033e-01],
[ 1.9207144e-01, -1.1127279e-01, -2.1466233e-02],
[ 1.8851870e-01, -1.1441709e-01, -1.9535454e-02],
[ 2.3465948e-03, -2.8141376e-01, 1.8149480e-01],
[ 5.0679259e-03, -2.9688439e-01, 1.9267374e-01],
[ 4.2662501e-01, -2.7377088e-02, -1.9792126e-01],
[ 4.3076158e-01, -2.7943321e-02, -1.9960576e-01],
[ 1.7624626e-01, -1.0117667e-01, -1.1177621e-02],
[ 1.8850681e-01, -1.0264599e-01, -1.6393466e-02],
[ 3.0229732e-02, -2.4176723e-01, 1.7432448e-01],
[ 3.0431338e-02, -2.4620293e-01, 1.7759261e-01]], dtype=float32),
33.0: array([[-0.30939412, 0.25602823, -0.04024093],
[-0.2729689 , 0.22444367, -0.03505757],
[-0.14540924, -0.03632286, 0.03015728],
[-0.13519996, -0.04967344, 0.03301162],
[-0.05827967, -0.32549065, 0.10954671],
[-0.06200996, -0.35197398, 0.11828617],
[-0.25822553, 0.2849875 , -0.07913747],
[-0.2373937 , 0.261985 , -0.07326637],
[-0.12168919, -0.04142176, 0.02295251],
[-0.12313439, -0.06454325, 0.03085142],
[-0.06531238, -0.35134014, 0.13087037],
[-0.07464114, -0.41234308, 0.1504453 ],
[ 0.40212965, 0.14688215, -0.2538089 ],
[ 0.40502876, 0.14444757, -0.25345838],
[ 0.14765052, -0.05676161, -0.03645246],
[ 0.14535622, -0.0613478 , -0.0336886 ],
[-0.07254177, -0.30174014, 0.18461415],
[-0.07621259, -0.3177934 , 0.19535702],
[ 0.4035117 , 0.09991341, -0.22432698],
[ 0.40730777, 0.10055739, -0.22627276],
[ 0.13202561, -0.05636821, -0.02333602],
[ 0.14340957, -0.05422735, -0.02931234],
[-0.06396261, -0.25345033, 0.16908593],
[-0.06537932, -0.25819907, 0.17230119]], dtype=float32),
36.0: array([[-0.50396615, 0.29274705, 0.11235292],
[-0.44456264, 0.25620574, 0.09924819],
[-0.22879232, -0.0870034 , 0.06738545],
[-0.21191655, -0.10371231, 0.06408966],
[-0.07566422, -0.48056102, 0.05607919],
[-0.08024184, -0.51949763, 0.06004917],
[-0.43611053, 0.31565544, 0.04795878],
[-0.4011911 , 0.28948408, 0.04321256],
[-0.19339654, -0.0939235 , 0.05577299],
[-0.19430675, -0.12704915, 0.0584758 ],
[-0.07802735, -0.49644494, 0.10468064],
[-0.09013319, -0.58751464, 0.11188557],
[ 0.5509915 , 0.09985324, -0.45323592],
[ 0.55569 , 0.09697405, -0.45446956],
[ 0.22231162, -0.07597005, -0.06129523],
[ 0.21944131, -0.08167987, -0.0579662 ],
[-0.05070086, -0.30051842, 0.34129518],
[-0.05251679, -0.31465408, 0.3628589 ],
[ 0.5631992 , 0.05266851, -0.4245024 ],
[ 0.5685592 , 0.05288879, -0.42816284],
[ 0.20307161, -0.0650299 , -0.0383005 ],
[ 0.21908702, -0.06422012, -0.04959998],
[-0.03913989, -0.22786357, 0.33567 ],
[-0.04021633, -0.23219548, 0.34196517]], dtype=float32),
43.0: array([[-0.19446045, 0.14111139, -0.07030037],
[-0.17218906, 0.12368871, -0.06144791],
[-0.15839593, -0.02104995, 0.02986392],
[-0.15411961, -0.02843486, 0.03414929],
[-0.19640754, -0.18180351, 0.13685781],
[-0.21151847, -0.19659571, 0.14786866],
[-0.14698116, 0.1345232 , -0.09998616],
[-0.13547467, 0.123155 , -0.09219711],
[-0.1307178 , -0.03346084, 0.02571187],
[-0.14284621, -0.04546097, 0.03546733],
[-0.17770374, -0.19517069, 0.15311405],
[-0.2151542 , -0.22931235, 0.17828181],
[ 0.35657182, -0.07064203, -0.18639596],
[ 0.35819957, -0.0719614 , -0.18573496],
[ 0.17184988, -0.09450629, -0.01604349],
[ 0.16850391, -0.09613059, -0.01340861],
[ 0.0384743 , -0.1609671 , 0.1625384 ],
[ 0.04375331, -0.16953689, 0.17154731],
[ 0.34728843, -0.08188462, -0.15949845],
[ 0.35071227, -0.08284059, -0.16085397],
[ 0.16043657, -0.08185948, -0.00823229],
[ 0.1705197 , -0.08459689, -0.01241025],
[ 0.06988467, -0.13401261, 0.14252807],
[ 0.07083529, -0.13644493, 0.14524056]], dtype=float32),
46.0: array([[-0.07004075, 0.14823915, -0.12062981],
[-0.06191396, 0.12997559, -0.10586517],
[-0.04629326, -0.01873393, 0.00459833],
[-0.04439182, -0.02639409, 0.0105104 ],
[-0.04514502, -0.1830036 , 0.1235109 ],
[-0.04848257, -0.1979015 , 0.1336539 ],
[-0.03757638, 0.17531031, -0.13760066],
[-0.03421605, 0.16141215, -0.12669265],
[-0.03685435, -0.017296 , 0.00126569],
[-0.03946661, -0.0307527 , 0.010909 ],
[-0.0595881 , -0.1985355 , 0.1249932 ],
[-0.06695981, -0.2327856 , 0.14873178],
[ 0.23125271, 0.15430318, -0.09171642],
[ 0.23194242, 0.15288556, -0.09051689],
[ 0.0554679 , -0.00371557, -0.0152194 ],
[ 0.05372222, -0.00674571, -0.01326459],
[-0.11368837, -0.17641218, 0.06042663],
[-0.12047824, -0.1858384 , 0.06292184],
[ 0.21815486, 0.12041678, -0.06729179],
[ 0.22011712, 0.12138867, -0.06788805],
[ 0.04399684, -0.00941188, -0.01043033],
[ 0.04996548, -0.00638132, -0.01222807],
[-0.10761664, -0.14928453, 0.0415298 ],
[-0.10969734, -0.15211557, 0.04237182]], dtype=float32),
49.0: array([[-0.20135659, 0.173211 , 0.09515937],
[-0.17838092, 0.151635 , 0.08385899],
[-0.17266777, -0.04569894, 0.03492574],
[-0.16853195, -0.05543169, 0.03144316],
[-0.22395536, -0.27045548, -0.00536828],
[-0.24130216, -0.2924906 , -0.00617419],
[-0.1683 , 0.15366586, 0.06207301],
[-0.15553586, 0.14017004, 0.05657336],
[-0.144473 , -0.05364803, 0.03064693],
[-0.1583723 , -0.07219444, 0.02919251],
[-0.1854651 , -0.2508234 , 0.02720371],
[-0.22928987, -0.30384007, 0.02383701],
[ 0.26134488, -0.16293277, -0.23967549],
[ 0.26294056, -0.16474764, -0.24071723],
[ 0.17769045, -0.06324313, -0.02047354],
[ 0.17482771, -0.06494926, -0.01906266],
[ 0.15610212, 0.01751273, 0.21085319],
[ 0.16944753, 0.02176814, 0.2246778 ],
[ 0.26199064, -0.1696758 , -0.22916172],
[ 0.26474783, -0.17123303, -0.2310967 ],
[ 0.17359264, -0.04303854, -0.0085063 ],
[ 0.18153597, -0.04781896, -0.01453036],
[ 0.19296354, 0.06014663, 0.21369907],
[ 0.19615784, 0.06120408, 0.21766268]], dtype=float32),
51.0: array([[ 0.05423057, 0.01247072, 0.08354215],
[ 0.04663565, 0.0101702 , 0.0738216 ],
[-0.10592296, -0.08253632, 0.05219272],
[-0.11187241, -0.08576963, 0.04982777],
[-0.3032679 , -0.2083361 , 0.04662673],
[-0.32753304, -0.22505388, 0.05003305],
[ 0.11284761, -0.01568013, 0.04937433],
[ 0.10389626, -0.01590446, 0.04506857],
[-0.08209607, -0.08362238, 0.04298121],
[-0.10391498, -0.0968257 , 0.04545613],
[-0.2942024 , -0.16834351, 0.06427979],
[-0.35227153, -0.21019517, 0.07168412],
[ 0.38482615, -0.19035782, -0.24518085],
[ 0.3840928 , -0.19072337, -0.24601373],
[ 0.14046565, -0.01154285, -0.05303598],
[ 0.13534622, -0.01201138, -0.05120148],
[-0.06735972, 0.17494848, 0.13527331],
[-0.06789238, 0.18893746, 0.1434592 ],
[ 0.34070194, -0.17382467, -0.2325198 ],
[ 0.3439528 , -0.1752367 , -0.23459291],
[ 0.12412901, 0.01127877, -0.04147127],
[ 0.13377658, 0.0067709 , -0.04779555],
[-0.02012438, 0.20933525, 0.12918134],
[-0.02083206, 0.21309596, 0.13165748]], dtype=float32),
68.0: array([[-0.34548748, 0.25186995, 0.22617714],
[-0.305383 , 0.22048324, 0.19919659],
[-0.22268176, -0.06886853, 0.06893239],
[-0.21323173, -0.08307813, 0.06025151],
[-0.20878968, -0.39914298, -0.04644113],
[-0.22452854, -0.43157032, -0.05090456],
[-0.31031954, 0.2612819 , 0.1801766 ],
[-0.28639907, 0.23942372, 0.16506036],
[-0.19090584, -0.07227596, 0.05918448],
[-0.20218311, -0.10023706, 0.05381618],
[-0.16954535, -0.39377213, -0.00624312],
[-0.20996827, -0.47056535, -0.01836516],
[ 0.30257937, 0.01701579, -0.35086057],
[ 0.30554414, 0.01445311, -0.3531867 ],
[ 0.20078875, -0.04359638, -0.05599025],
[ 0.19815093, -0.04783039, -0.05452442],
[ 0.16652963, -0.12863263, 0.24445714],
[ 0.18047899, -0.13240786, 0.26086834],
[ 0.31792435, -0.01711002, -0.34664682],
[ 0.32122737, -0.017359 , -0.34965956],
[ 0.19827957, -0.0317658 , -0.03880182],
[ 0.2078451 , -0.0325252 , -0.04809804],
[ 0.20109268, -0.067109 , 0.25406846],
[ 0.20440647, -0.06848276, 0.25880724]], dtype=float32),
93.0: array([[-0.14368315, 0.2932776 , -0.0936086 ],
[-0.12734419, 0.25726846, -0.08201108],
[-0.12981503, -0.02027437, 0.01871139],
[-0.12705499, -0.03500349, 0.02377617],
[-0.17561604, -0.32166493, 0.13196208],
[-0.18919094, -0.3482897 , 0.14270103],
[-0.09987956, 0.23017864, -0.11107448],
[-0.09201773, 0.2096558 , -0.10220388],
[-0.10721786, -0.03038035, 0.01478375],
[-0.11824836, -0.05389083, 0.02462836],
[-0.16524684, -0.24797955, 0.13318977],
[-0.19847237, -0.3144232 , 0.15843274],
[ 0.32211155, -0.47713566, -0.10014185],
[ 0.32331243, -0.48034674, -0.09919468],
[ 0.13845916, -0.06773003, -0.01835995],
[ 0.13533346, -0.06789365, -0.01634242],
[-0.00680299, 0.34951344, 0.0623918 ],
[-0.00482513, 0.37611473, 0.06489561],
[ 0.30963567, -0.4695673 , -0.07841181],
[ 0.3126368 , -0.47357744, -0.07911362],
[ 0.12751502, -0.03034689, -0.0144863 ],
[ 0.13639992, -0.04285953, -0.01659761],
[ 0.02087965, 0.40337712, 0.04218986],
[ 0.02098497, 0.4107398 , 0.04305407]], dtype=float32),
162.0: array([[-1.9727017e-01, -4.4334605e-02, 1.4260431e-01],
[-1.7515738e-01, -3.8771003e-02, 1.2592117e-01],
[-2.1244104e-01, 1.4815328e-02, 7.8050502e-02],
[-2.0974469e-01, 1.7525828e-02, 7.3688120e-02],
[-3.2267377e-01, 7.6507017e-02, 5.3131707e-02],
[-3.4783560e-01, 8.3056070e-02, 5.6940284e-02],
[-1.3740374e-01, -1.4423163e-02, 1.3233699e-01],
[-1.2694649e-01, -1.2368227e-02, 1.2206805e-01],
[-1.7771706e-01, -7.4426387e-03, 7.3126167e-02],
[-1.9833694e-01, 5.8927335e-04, 7.4860401e-02],
[-2.9461169e-01, -2.6569694e-02, 5.5812616e-02],
[-3.5647836e-01, -5.7319026e-03, 6.5119587e-02],
[ 4.4086254e-01, 2.0722334e-01, -9.5875286e-02],
[ 4.4241121e-01, 2.0818883e-01, -9.7216718e-02],
[ 2.1230680e-01, -1.6446747e-01, -2.6930336e-02],
[ 2.0738715e-01, -1.6565020e-01, -2.6183102e-02],
[ 4.6934202e-02, -6.3952762e-01, 4.0254384e-02],
[ 5.4369818e-02, -6.8266791e-01, 4.2140432e-02],
[ 4.2378816e-01, 1.9387028e-01, -1.0517057e-01],
[ 4.2798233e-01, 1.9490911e-01, -1.0613864e-01],
[ 2.0066445e-01, -1.7155507e-01, -2.8574029e-02],
[ 2.1298161e-01, -1.6766459e-01, -3.1486735e-02],
[ 9.8323628e-02, -6.5786541e-01, 3.2294940e-02],
[ 9.9679306e-02, -6.6972762e-01, 3.2960348e-02]], dtype=float32)},
'1001B': {-18.0: array([[ 1.18133366e-01, -1.88309908e-01, -3.48824933e-02],
[ 1.74951598e-01, -2.79627442e-01, -5.14742956e-02],
[-2.78264703e-03, 6.94553275e-03, 2.22006142e-02],
[-4.83069709e-03, 1.12034138e-02, 3.46923731e-02],
[-3.36618125e-02, 5.56390733e-02, 2.77447663e-02],
[-5.52991852e-02, 9.13749784e-02, 4.55625392e-02],
[ 8.42830688e-02, -7.50173628e-02, -2.25346535e-02],
[ 1.26282990e-01, -1.54178485e-01, -3.59191559e-02],
[ 1.01538850e-02, 1.34188905e-02, 7.22976029e-03],
[ 1.31846424e-02, 2.56407983e-03, 1.23187602e-02],
[-1.24517800e-02, 2.24618837e-02, 1.06134443e-02],
[-2.29785517e-02, 3.95101309e-02, 1.93641372e-02],
[ 4.21158187e-02, 3.71672571e-01, 1.04648471e-02],
[ 4.08236310e-02, 3.08882892e-01, 7.57816853e-03],
[ 1.82827413e-02, 1.77068159e-01, 4.55470337e-03],
[ 1.58165749e-02, 1.48515105e-01, 4.98660048e-03],
[ 2.89575837e-04, 1.06156133e-02, 2.87179049e-04],
[-3.21495609e-04, 1.09609328e-02, 1.10675488e-03],
[ 3.77708003e-02, 3.75395179e-01, 5.11166919e-03],
[ 3.09147388e-02, 3.26939762e-01, 9.02234297e-03],
[ 3.00402120e-02, 1.38398051e-01, -3.41300704e-02],
[ 1.73018761e-02, 1.42095968e-01, -4.77384729e-03],
[ 9.30709019e-03, -1.82230351e-03, -2.12248545e-02],
[ 3.00935120e-03, 6.00902364e-03, -5.30830165e-03]], dtype=float32),
-13.0: array([[-0.0112452 , 0.14295611, 0.10049051],
[-0.01686104, 0.21182632, 0.14850718],
[-0.03546895, -0.06755801, -0.03952306],
[-0.05508869, -0.10537651, -0.06192107],
[-0.02654133, -0.09409943, -0.0595783 ],
[-0.04357116, -0.15451375, -0.09784607],
[-0.01985774, 0.04342896, 0.06710176],
[-0.02059358, 0.10775694, 0.10464401],
[-0.0218706 , -0.04322306, -0.00940886],
[-0.03125443, -0.05163033, -0.01715003],
[-0.01065568, -0.03748884, -0.02268458],
[-0.01902706, -0.06680565, -0.0414253 ],
[-0.10092748, -0.39926842, -0.00871534],
[-0.08649848, -0.33391133, -0.00280779],
[-0.04714739, -0.19204253, -0.00826428],
[-0.04108718, -0.16178189, -0.00603913],
[-0.00276733, -0.01302059, -0.00278414],
[-0.00325748, -0.01331649, -0.00288727],
[-0.0949869 , -0.40667695, -0.02190183],
[-0.08499439, -0.3501074 , -0.01436343],
[-0.01744868, -0.18215896, -0.04467044],
[-0.03268857, -0.15996172, -0.01512256],
[ 0.01080004, -0.01700355, -0.02143635],
[ 0.00110853, -0.01135189, -0.00585271]], dtype=float32),
8.0: array([[-0.07475785, 0.29503447, 0.00611002],
[-0.1107536 , 0.4374371 , 0.00860781],
[-0.01946294, -0.06711604, -0.02355612],
[-0.02989203, -0.10514888, -0.03680401],
[ 0.00364675, -0.13403197, -0.0212734 ],
[ 0.00601186, -0.2201054 , -0.03493803],
[-0.06677445, 0.12526144, 0.0200495 ],
[-0.09010354, 0.24902599, 0.0195592 ],
[-0.02191887, -0.04126787, -0.00191023],
[-0.02851316, -0.04155755, -0.01037002],
[ 0.00065557, -0.05304032, -0.00769964],
[ 0.00178967, -0.09466433, -0.01453274],
[-0.13136095, -0.54102904, 0.11125209],
[-0.11519007, -0.4482959 , 0.0953928 ],
[-0.05993262, -0.26113248, 0.04997608],
[-0.05202586, -0.21893172, 0.04272253],
[-0.0025961 , -0.01782043, 0.0013432 ],
[-0.00273245, -0.0183415 , 0.00150852],
[-0.12127904, -0.5558006 , 0.10212881],
[-0.10782851, -0.4790861 , 0.09309475],
[-0.02796945, -0.24385269, 0.00533589],
[-0.04279139, -0.21765903, 0.03262716],
[ 0.01043928, -0.02023044, -0.01951924],
[ 0.00057348, -0.01475031, -0.00321994]], dtype=float32),
13.0: array([[-1.50526792e-01, 1.46242410e-01, 1.16960280e-01],
[-2.22728580e-01, 2.16727033e-01, 1.72913417e-01],
[ 1.39365597e-02, -7.42558986e-02, -2.36520730e-02],
[ 2.23793797e-02, -1.15764685e-01, -3.73613946e-02],
[ 5.15612774e-02, -1.00536592e-01, -5.07483408e-02],
[ 8.46999511e-02, -1.65080488e-01, -8.33569467e-02],
[-1.14425756e-01, 3.80315855e-02, 9.03270394e-02],
[-1.66747645e-01, 1.05269030e-01, 1.30992979e-01],
[-1.18839527e-02, -4.94832993e-02, 4.44862200e-03],
[-1.19188186e-02, -5.88586777e-02, 5.73510479e-04],
[ 1.90001745e-02, -4.01986726e-02, -1.88462473e-02],
[ 3.52013744e-02, -7.15074241e-02, -3.48312333e-02],
[-1.02355830e-01, -4.57062095e-01, 8.62820968e-02],
[-9.37937200e-02, -3.83010656e-01, 7.88099021e-02],
[-4.42633778e-02, -2.19350085e-01, 3.63511518e-02],
[-3.86833064e-02, -1.84928447e-01, 3.20412964e-02],
[-4.77059686e-04, -1.46224741e-02, -2.33387516e-04],
[-1.13505623e-04, -1.49272261e-02, -3.28588649e-04],
[-8.83236453e-02, -4.63942707e-01, 7.08840489e-02],
[-7.91515037e-02, -3.99770170e-01, 6.61816299e-02],
[-1.58829130e-02, -2.05032870e-01, -8.13293178e-03],
[-3.03206444e-02, -1.81968302e-01, 2.04377957e-02],
[ 1.02425264e-02, -1.77637208e-02, -2.05123164e-02],
[ 1.10441656e-03, -1.25322584e-02, -4.01904294e-03]], dtype=float32),
17.0: array([[ 1.05999485e-01, 6.96798787e-02, -1.01033643e-01],
[ 1.57700494e-01, 1.02432035e-01, -1.49876952e-01],
[-7.93236215e-03, -1.07838206e-01, 2.65160333e-02],
[-1.23913139e-02, -1.68009445e-01, 4.14474383e-02],
[-3.46218832e-02, -1.08276501e-01, 4.88251932e-02],
[-5.68460785e-02, -1.77789241e-01, 8.01754072e-02],
[ 1.61692277e-02, 2.57226117e-02, -3.41605805e-02],
[ 6.61601052e-02, 5.79037070e-02, -7.84138814e-02],
[-2.17959434e-02, -5.07823415e-02, 1.98256988e-02],
[-1.38959214e-02, -7.72777349e-02, 1.99718047e-02],
[-1.47277778e-02, -4.22130972e-02, 1.95557214e-02],
[-2.52160095e-02, -7.63386488e-02, 3.46926451e-02],
[-4.02615130e-01, -1.81526199e-01, 2.50307024e-01],
[-3.38747233e-01, -1.51283994e-01, 2.09195167e-01],
[-1.90156579e-01, -8.96021500e-02, 1.18234165e-01],
[-1.60192698e-01, -7.67306015e-02, 1.00490339e-01],
[-1.06764352e-02, -7.92888645e-03, 6.85321912e-03],
[-1.07887499e-02, -8.69159587e-03, 7.61635881e-03],
[-4.00986612e-01, -1.86940730e-01, 2.45923862e-01],
[-3.49304706e-01, -1.60080716e-01, 2.17678264e-01],
[-1.47623882e-01, -9.04135182e-02, 6.32960424e-02],
[-1.51744977e-01, -7.47964755e-02, 8.79914761e-02],
[ 2.06660177e-03, -1.17530441e-02, -1.73030347e-02],
[-6.38063345e-03, -6.23983797e-03, -1.61702977e-04]], dtype=float32),
22.0: array([[-9.11770388e-02, -1.31094322e-01, 2.97883272e-01],
[-1.35447577e-01, -1.94972143e-01, 4.41338480e-01],
[-5.88189922e-02, -6.13654740e-02, -2.41629463e-02],
[-9.10781845e-02, -9.50692296e-02, -3.86870056e-02],
[-2.47784369e-02, -1.63619202e-02, -9.91136879e-02],
[-4.06556614e-02, -2.68393904e-02, -1.62798315e-01],
[-8.03407282e-02, -8.01510140e-02, 1.85120404e-01],
[-1.08194910e-01, -1.27985865e-01, 2.97044903e-01],
[-4.11466137e-02, -3.22834998e-02, 5.10238810e-03],
[-5.95600829e-02, -5.59967756e-02, 1.08728055e-02],
[-1.04441987e-02, -6.49153395e-03, -3.77581008e-02],
[-1.83432028e-02, -1.19411359e-02, -6.86419681e-02],
[-1.61743000e-01, 3.44373807e-02, -1.01534091e-01],
[-1.41626939e-01, 2.41151210e-02, -7.36141205e-02],
[-7.49408528e-02, 1.62507873e-02, -5.47885559e-02],
[-6.55113608e-02, 1.19890636e-02, -4.35573868e-02],
[-4.16387571e-03, 2.04114811e-04, -6.40431885e-03],
[-4.56530415e-03, 1.57867529e-04, -6.77879760e-03],
[-1.50832579e-01, 3.84181365e-02, -1.25137389e-01],
[-1.33271307e-01, 3.50215882e-02, -1.03597917e-01],
[-4.13053520e-02, 1.45732774e-03, -8.74355808e-02],
[-5.44436425e-02, 1.21557973e-02, -5.50708920e-02],
[ 9.14254412e-03, -7.67410686e-03, -2.36991588e-02],
[-2.73344951e-04, -1.29791780e-03, -8.21637362e-03]], dtype=float32),
26.0: array([[-0.1368274 , -0.11533888, 0.18143742],
[-0.20252408, -0.17194901, 0.26889113],
[-0.01198337, -0.07009092, 0.00532349],
[-0.01793809, -0.10883739, 0.00756914],
[ 0.02635919, -0.02784307, -0.04369224],
[ 0.04332242, -0.04570227, -0.07178073],
[-0.11809674, -0.05156573, 0.12243323],
[-0.16219324, -0.09715518, 0.18815291],
[-0.02807509, -0.02614654, 0.01629342],
[-0.03370347, -0.05387031, 0.02456341],
[ 0.00891865, -0.01031007, -0.01626516],
[ 0.0172211 , -0.01948772, -0.02983904],
[-0.20377831, 0.16581263, 0.01494089],
[-0.17954312, 0.13678491, 0.0206255 ],
[-0.09285554, 0.07743537, 0.00251964],
[-0.07996275, 0.06349955, 0.00452129],
[-0.00378551, 0.00313708, -0.00186021],
[-0.00352625, 0.00278937, -0.0017744 ],
[-0.19023848, 0.16763484, -0.00324981],
[-0.1669738 , 0.14677817, 0.00201439],
[-0.06096201, 0.05556365, -0.03848111],
[-0.07030387, 0.06227095, -0.0081778 ],
[ 0.00631598, -0.00449238, -0.02192551],
[-0.00165565, 0.0017312 , -0.00565192]], dtype=float32),
30.0: array([[-2.00018778e-01, -1.15553133e-01, 1.87804133e-01],
[-2.95991033e-01, -1.71273038e-01, 2.77993649e-01],
[ 1.00621553e-02, -7.97187630e-03, -2.98362672e-02],
[ 1.66010354e-02, -1.19405407e-02, -4.71787006e-02],
[ 6.14772812e-02, 2.40133628e-02, -7.46669099e-02],
[ 1.00996301e-01, 3.94552462e-02, -1.22639738e-01],
[-1.56228051e-01, -7.97872990e-02, 1.24851100e-01],
[-2.24698022e-01, -1.21162534e-01, 1.94091484e-01],
[-2.14052405e-02, -1.32097369e-02, 9.08556045e-04],
[-2.34439895e-02, -1.96001232e-02, -6.42296567e-04],
[ 2.24003009e-02, 8.82335287e-03, -2.82538682e-02],
[ 4.17214036e-02, 1.62750315e-02, -5.16378433e-02],
[-1.69260755e-01, -2.45188251e-02, -8.04692227e-03],
[-1.52892932e-01, -2.56773662e-02, 1.43011182e-03],
[-7.49621987e-02, -1.03285955e-02, -9.23031569e-03],
[-6.49583191e-02, -9.30300076e-03, -6.02099299e-03],
[-1.83337741e-03, -1.23536389e-04, -3.30191175e-03],
[-1.26854237e-03, 3.54641874e-04, -3.53847677e-03],
[-1.51948512e-01, -2.05066130e-02, -2.64336392e-02],
[-1.34089008e-01, -1.60035528e-02, -1.81214716e-02],
[-4.34023738e-02, -2.26475149e-02, -4.74739708e-02],
[-5.51630445e-02, -1.05781769e-02, -1.70457605e-02],
[ 8.15709960e-03, -8.78625363e-03, -2.20778622e-02],
[-5.17860579e-04, -2.58265692e-03, -6.09858287e-03]], dtype=float32),
35.0: array([[-2.27499679e-01, -2.42018163e-01, 3.66758138e-01],
[-3.37255239e-01, -3.59272927e-01, 5.43432415e-01],
[-2.46942863e-02, -8.23185593e-02, -2.23686732e-02],
[-3.74993905e-02, -1.27108961e-01, -3.61541621e-02],
[ 3.97855900e-02, -4.35548229e-03, -1.15884945e-01],
[ 6.53794631e-02, -7.09141046e-03, -1.90350354e-01],
[-1.59854680e-01, -1.75580293e-01, 2.29799435e-01],
[-2.40536645e-01, -2.58962631e-01, 3.67044568e-01],
[-3.11594624e-02, -5.96841238e-02, 1.03301387e-02],
[-4.60647717e-02, -9.12753791e-02, 1.93753764e-02],
[ 1.44043900e-02, -2.86675547e-03, -4.40571979e-02],
[ 2.67062392e-02, -4.55724029e-03, -8.01439285e-02],
[-7.07146674e-02, -1.31669119e-01, -1.09063499e-01],
[-6.99403509e-02, -1.22092299e-01, -7.71167427e-02],
[-2.90846340e-02, -6.02537841e-02, -5.94733059e-02],
[-2.71634590e-02, -5.37455864e-02, -4.69634272e-02],
[ 4.56977177e-05, -3.42764333e-03, -7.14857318e-03],
[ 2.21787443e-04, -3.18200979e-03, -7.67358905e-03],
[-5.27907759e-02, -1.19319119e-01, -1.36697754e-01],
[-4.84120175e-02, -1.01781793e-01, -1.13485537e-01],
[-1.39163539e-03, -6.19780906e-02, -9.28540230e-02],
[-1.66221950e-02, -4.85427603e-02, -5.96612468e-02],
[ 1.08844656e-02, -1.00248484e-02, -2.43217088e-02],
[ 1.87564362e-03, -4.60565602e-03, -8.57876893e-03]], dtype=float32),
43.0: array([[-0.03422698, 0.0219276 , 0.26586667],
[-0.05132531, 0.03292428, 0.3939278 ],
[-0.05826478, -0.0135453 , -0.02735849],
[-0.09058046, -0.02087476, -0.04350309],
[-0.03946101, -0.01701864, -0.09327448],
[-0.06478548, -0.02792886, -0.15320142],
[-0.02141931, -0.02907741, 0.15883563],
[-0.03281641, -0.0118154 , 0.26017493],
[-0.02664969, -0.02517635, -0.00098631],
[-0.04512969, -0.02332332, 0.00320028],
[-0.0152822 , -0.00786365, -0.03575705],
[-0.02787559, -0.01296202, -0.06480987],
[-0.00593319, -0.32611504, -0.13941337],
[-0.00642949, -0.27687377, -0.10723824],
[-0.00269838, -0.15442222, -0.07213566],
[-0.00424764, -0.12983485, -0.05844666],
[-0.00072422, -0.00896054, -0.0071668 ],
[-0.0015052 , -0.00848231, -0.00742251],
[ 0.00053042, -0.32753608, -0.16097912],
[-0.0025122 , -0.28133264, -0.13451897],
[ 0.02341778, -0.15215296, -0.10316559],
[ 0.00451439, -0.12983732, -0.06910806],
[ 0.01366246, -0.01689863, -0.02500806],
[ 0.00347639, -0.00994114, -0.00916432]], dtype=float32),
48.0: array([[ 0.11825339, 0.01286125, 0.15518437],
[ 0.1752402 , 0.02020713, 0.2304122 ],
[-0.02440295, 0.0214928 , 0.03128562],
[-0.03831647, 0.03389681, 0.0481727 ],
[-0.05165716, 0.01462595, -0.01507856],
[-0.08483513, 0.02404398, -0.02478757],
[ 0.0590561 , -0.06302807, 0.09257875],
[ 0.10678815, -0.04418084, 0.15064746],
[-0.01119609, -0.02432105, 0.01948496],
[-0.01161104, -0.00941616, 0.03574726],
[-0.02016772, 0.00346453, -0.00568719],
[-0.03626468, 0.00855273, -0.01014906],
[-0.14991498, -0.5220059 , -0.06947006],
[-0.12320855, -0.44479254, -0.05260213],
[-0.06930507, -0.24227755, -0.03548918],
[-0.06041632, -0.20526552, -0.02754301],
[-0.00337124, -0.01131709, -0.00285257],
[-0.00502101, -0.01131174, -0.00252493],
[-0.13741295, -0.50586635, -0.0826842 ],
[-0.13009161, -0.44497445, -0.06828923],
[ 0.0313036 , -0.15293588, -0.05923861],
[-0.03673863, -0.18521914, -0.03664908],
[ 0.0489165 , 0.02220794, -0.01652002],
[ 0.01007678, -0.00304122, -0.00563322]], dtype=float32),
56.0: array([[-1.39564812e-01, 1.48426682e-01, 1.10185228e-01],
[-2.06970513e-01, 2.20243320e-01, 1.63048610e-01],
[-8.09035543e-03, -8.75890329e-02, -7.01760314e-03],
[-1.20916991e-02, -1.36312827e-01, -1.14356754e-02],
[ 3.02670095e-02, -1.12163804e-01, -3.50933038e-02],
[ 4.97246869e-02, -1.84157908e-01, -5.76506667e-02],
[-8.69229287e-02, 7.64358556e-03, 8.53211358e-02],
[-1.38867468e-01, 8.25122818e-02, 1.23226777e-01],
[-1.07478080e-02, -7.03140274e-02, 1.07999137e-02],
[-1.91006400e-02, -7.98767805e-02, 1.15828430e-02],
[ 1.14464341e-02, -4.56688143e-02, -1.28420079e-02],
[ 2.07842086e-02, -8.04724395e-02, -2.38659717e-02],
[ 4.09063436e-02, -6.95723534e-01, 8.70965943e-02],
[ 2.88544670e-02, -5.86317599e-01, 7.91614801e-02],
[ 2.22455040e-02, -3.31475437e-01, 3.72198150e-02],
[ 1.70306042e-02, -2.79840142e-01, 3.30695696e-02],
[ 2.47649965e-03, -2.07547396e-02, 2.13530933e-04],
[ 2.41362699e-03, -2.09644586e-02, 3.05745605e-04],
[ 5.28594181e-02, -6.99248195e-01, 7.19475076e-02],
[ 4.29646038e-02, -6.04120195e-01, 6.72362223e-02],
[ 4.31329347e-02, -2.96833515e-01, -8.77996627e-03],
[ 2.44133063e-02, -2.71981925e-01, 2.06509605e-02],
[ 1.36583345e-02, -1.95983667e-02, -2.11282000e-02],
[ 4.39605815e-03, -1.70498081e-02, -4.15294897e-03]], dtype=float32),
65.0: array([[ 1.00193754e-01, -1.09461159e-01, 1.35065779e-01],
[ 1.48223922e-01, -1.62042603e-01, 1.99826688e-01],
[-2.66734101e-02, -3.98326442e-02, -1.40413204e-02],
[-4.19021510e-02, -6.12889901e-02, -2.24771462e-02],
[-4.87915725e-02, -4.06869641e-03, -4.75463383e-02],
[-8.01339820e-02, -6.63701445e-03, -7.81039223e-02],
[ 6.48655593e-02, -1.16098583e-01, 1.03387475e-01],
[ 1.02413535e-01, -1.46236002e-01, 1.50225535e-01],
[-4.92355460e-03, -4.57473509e-02, 1.03479307e-02],
[-8.70311167e-03, -5.66163659e-02, 9.85320006e-03],
[-1.85753889e-02, -3.25081102e-03, -1.75244194e-02],
[-3.38895880e-02, -4.45767585e-03, -3.24695297e-02],
[-1.97089948e-02, -3.31211746e-01, 9.65207741e-02],
[-1.25784911e-02, -2.86693931e-01, 8.82637054e-02],
[-1.07662855e-02, -1.55180678e-01, 4.11147326e-02],
[-9.41807218e-03, -1.31949276e-01, 3.63514833e-02],
[-1.52185000e-03, -8.55916273e-03, 1.07492306e-04],
[-2.31644535e-03, -7.98362400e-03, 2.99270469e-05],
[-2.12671701e-02, -3.24824035e-01, 8.01502913e-02],
[-2.09669601e-02, -2.79233545e-01, 7.41291270e-02],
[ 1.17288632e-02, -1.49619520e-01, -3.66265094e-03],
[-4.40787384e-03, -1.28533006e-01, 2.41445098e-02],
[ 1.16240820e-02, -1.60149038e-02, -1.99354403e-02],
[ 2.57951068e-03, -9.66545008e-03, -3.71187087e-03]], dtype=float32),
93.0: array([[ 0.03387733, -0.03269347, 0.20284744],
[ 0.05009072, -0.04739994, 0.30098483],
[-0.03208399, 0.01771432, 0.0275381 ],
[-0.04996682, 0.02815826, 0.04214362],
[-0.03568201, 0.02355189, -0.03084492],
[-0.05858772, 0.03870765, -0.05068607],
[ 0.00605299, -0.08838265, 0.12567939],
[ 0.02253845, -0.08708252, 0.20094107],
[-0.01909234, -0.02685613, 0.02202544],
[-0.02532668, -0.01535899, 0.03884961],
[-0.0141524 , 0.00693186, -0.01158481],
[-0.02534322, 0.0147431 , -0.02097993],
[-0.12901762, -0.4864478 , -0.05999541],
[-0.10918865, -0.41659138, -0.04245479],
[-0.05657459, -0.2236111 , -0.03221293],
[-0.05152459, -0.19077094, -0.02433904],
[-0.0014283 , -0.00944967, -0.00325171],
[-0.00354453, -0.00986327, -0.00300564],
[-0.10362079, -0.46166003, -0.07738842],
[-0.10605905, -0.41073835, -0.06286491],
[ 0.08585218, -0.10326439, -0.0634507 ],
[-0.01610331, -0.16226546, -0.03579432],
[ 0.07352694, 0.04163834, -0.02017326],
[ 0.01692932, 0.00266931, -0.00647708]], dtype=float32)},
'1004B': {-2.5: array([[-4.9269088e-02, 7.0253588e-02, 1.5979638e-02],
[-2.2227876e-02, 7.7585109e-02, 5.9659332e-03],
[ 2.0059975e-02, -6.4141909e-03, -5.7245540e-03],
[ 1.9896487e-02, -5.7877542e-04, -6.3721365e-03],
[ 1.5116483e-01, -3.0431725e-02, -4.8283156e-02],
[ 9.4835952e-02, -1.8845901e-02, -3.0321309e-02],
[-1.6076533e-01, 7.1748425e-03, 5.8123033e-02],
[-2.0262763e-02, 3.4287576e-02, 6.4535746e-03],
[-2.3573207e-02, -2.3675224e-02, 1.1320283e-02],
[ 3.1886149e-02, -6.8608490e-03, -9.9371914e-03],
[ 2.2914739e-01, -4.7698654e-02, -7.2994366e-02],
[ 1.6145971e-01, -3.2227099e-02, -5.1608320e-02],
[-1.0788523e-01, -1.6615471e-02, 3.9557178e-02],
[-8.4514720e-03, 7.2295417e-04, 3.0529252e-03],
[-4.0272888e-02, -1.1276133e-02, 1.5345521e-02],
[ 4.1297274e-03, -2.2328070e-03, -1.1297760e-03],
[ 5.4203216e-02, -1.2067673e-02, -1.7166398e-02],
[ 3.4290820e-02, -6.9121635e-03, -1.0952126e-02],
[-8.7242862e-03, -1.4849324e-03, 3.2017203e-03],
[-7.4087409e-04, -1.1080279e-04, 2.7173405e-04],
[-3.4837236e-03, -7.0597418e-04, 1.2922820e-03],
[-1.2910664e-04, -1.0474942e-04, 5.7992504e-05],
[ 1.1447904e-03, -3.4757238e-04, -3.5074141e-04],
[ 1.0151347e-03, -2.1127274e-04, -3.2337700e-04]], dtype=float32),
2.0: array([[-4.75066118e-02, 7.18677565e-02, 1.52170612e-02],
[-2.65195742e-02, 7.61082619e-02, 8.53245705e-03],
[ 5.01976758e-02, 1.24573931e-02, -2.60474011e-02],
[ 3.61963548e-02, 9.85481869e-03, -1.75942592e-02],
[ 2.64017045e-01, 4.15130332e-02, -1.26017049e-01],
[ 1.65046066e-01, 2.59303935e-02, -7.87013397e-02],
[-1.31244317e-01, 2.25897823e-02, 4.21097055e-02],
[-1.94653720e-02, 3.51067223e-02, 6.03849301e-03],
[ 3.47587056e-02, 1.14631364e-02, -2.65104156e-02],
[ 6.14138767e-02, 1.18438387e-02, -3.01160309e-02],
[ 4.04114068e-01, 6.37374818e-02, -1.93391785e-01],
[ 2.81258851e-01, 4.41663414e-02, -1.34150177e-01],
[-8.63789916e-02, -5.66156069e-03, 2.79017612e-02],
[-6.77520363e-03, 1.61726680e-03, 2.12541176e-03],
[-2.10426953e-02, -3.65976739e-04, 3.63380439e-03],
[ 1.10138580e-02, 2.05377955e-03, -5.75426454e-03],
[ 9.75721404e-02, 1.55008165e-02, -4.69489023e-02],
[ 5.99006489e-02, 9.41405632e-03, -2.85921060e-02],
[-6.99245324e-03, -6.06534421e-04, 2.26541818e-03],
[-5.89341798e-04, -3.31359188e-05, 1.89117854e-04],
[-2.52112770e-03, -1.89970102e-04, 7.40200223e-04],
[ 1.07333100e-04, 3.68764777e-05, -9.45125212e-05],
[ 2.29527173e-03, 3.77760531e-04, -1.13399071e-03],
[ 1.79006776e-03, 2.82285531e-04, -8.56626080e-04]], dtype=float32),
5.0: array([[-1.7156062e-01, -3.0334314e-02, 1.0692491e-01],
[-1.5186127e-01, -2.5210133e-02, 1.0100596e-01],
[ 3.7026159e-02, -1.9346474e-02, -1.0590827e-02],
[ 2.6104443e-02, -1.0485310e-02, -6.7428360e-03],
[ 2.3924948e-01, -6.2245902e-02, -8.4258355e-02],
[ 1.4964461e-01, -3.8671780e-02, -5.2703563e-02],
[-1.9956408e-01, -4.2549726e-02, 9.3369596e-02],
[-7.9246514e-02, -1.4356094e-02, 5.0327171e-02],
[ 1.5753334e-02, -4.0227223e-02, -3.3104648e-03],
[ 5.1724706e-02, -1.7468456e-02, -1.7015718e-02],
[ 3.6557236e-01, -9.6882634e-02, -1.2869804e-01],
[ 2.5502604e-01, -6.5999456e-02, -8.9835882e-02],
[-9.9214040e-02, -2.2638733e-02, 3.7864558e-02],
[-1.1036652e-02, -2.4512447e-03, 5.3423196e-03],
[-2.7787842e-02, -1.5519940e-02, 1.0673237e-02],
[ 9.1581056e-03, -4.2074234e-03, -3.1003356e-03],
[ 8.7928019e-02, -2.4205614e-02, -3.0922597e-02],
[ 5.4287616e-02, -1.4124919e-02, -1.9121738e-02],
[-7.7959974e-03, -1.7772205e-03, 2.8943757e-03],
[-6.8627921e-04, -1.5959526e-04, 2.6475795e-04],
[-2.9049809e-03, -8.8400947e-04, 1.0834557e-03],
[ 3.8187558e-05, -1.6362818e-04, -7.3817982e-06],
[ 2.0288876e-03, -6.6379114e-04, -7.0968125e-04],
[ 1.6194314e-03, -4.2906689e-04, -5.7013793e-04]], dtype=float32),
7.0: array([[-4.66910601e-02, 6.33070469e-02, 1.73988529e-02],
[-2.37727500e-02, 7.48846084e-02, 7.53558148e-03],
[ 4.48971651e-02, -3.56182940e-02, -6.62765186e-03],
[ 3.36174853e-02, -1.69562064e-02, -6.84961025e-03],
[ 2.45622307e-01, -1.41680077e-01, -5.21154702e-02],
[ 1.53621867e-01, -8.80928636e-02, -3.27075198e-02],
[-1.39939412e-01, -1.67711750e-02, 5.82763776e-02],
[-1.89734641e-02, 3.08327153e-02, 7.13091157e-03],
[ 2.27869041e-02, -7.70312026e-02, 9.65003110e-03],
[ 5.64687289e-02, -3.58227231e-02, -1.08864475e-02],
[ 3.75464886e-01, -2.19955161e-01, -7.89220035e-02],
[ 2.61755735e-01, -1.50364026e-01, -5.56795523e-02],
[-9.32000577e-02, -3.22292782e-02, 3.92418876e-02],
[-7.24425865e-03, -6.90057641e-04, 3.06420657e-03],
[-2.58060824e-02, -2.74495501e-02, 1.48776779e-02],
[ 9.76908859e-03, -8.81359074e-03, -1.34762551e-03],
[ 9.04066861e-02, -5.46549149e-02, -1.86284930e-02],
[ 5.57257943e-02, -3.21569666e-02, -1.18218530e-02],
[-7.54715595e-03, -2.72667664e-03, 3.17389611e-03],
[-6.36785815e-04, -2.22069299e-04, 2.69537821e-04],
[-2.79596797e-03, -1.45308417e-03, 1.27287128e-03],
[ 5.78002728e-05, -3.19048733e-04, 5.12159131e-05],
[ 2.09794473e-03, -1.46492186e-03, -3.88712739e-04],
[ 1.66318426e-03, -9.74208699e-04, -3.49631970e-04]], dtype=float32),
8.5: array([[-4.92597111e-02, 7.42106661e-02, 1.43896183e-02],
[-2.76343767e-02, 7.91683048e-02, 6.94074295e-03],
[ 4.50091958e-02, 1.16754863e-02, -2.28559189e-02],
[ 3.32429558e-02, 9.64136329e-03, -1.59210823e-02],
[ 2.44154751e-01, 3.90866175e-02, -1.14047796e-01],
[ 1.52680367e-01, 2.44323239e-02, -7.12568760e-02],
[-1.35251611e-01, 2.05621570e-02, 4.52183038e-02],
[-2.03400794e-02, 3.62897180e-02, 5.62427472e-03],
[ 2.55047288e-02, 8.86086561e-03, -2.00991463e-02],
[ 5.62487245e-02, 1.11770760e-02, -2.69930214e-02],
[ 3.73374283e-01, 5.99018931e-02, -1.74821392e-01],
[ 2.60163993e-01, 4.16037068e-02, -1.21446311e-01],
[-8.87996405e-02, -7.94139132e-03, 3.04865502e-02],
[-7.01636309e-03, 1.52187748e-03, 2.29646591e-03],
[-2.37454586e-02, -1.69286935e-03, 5.83320996e-03],
[ 9.85236466e-03, 1.83915382e-03, -5.01320697e-03],
[ 8.99796113e-02, 1.45139731e-02, -4.23395112e-02],
[ 5.53935133e-02, 8.86303559e-03, -2.58757826e-02],
[-7.18324538e-03, -7.96514621e-04, 2.47581885e-03],
[-6.06714282e-04, -4.87486868e-05, 2.07215737e-04],
[-2.64110183e-03, -2.76946055e-04, 8.53940146e-04],
[ 7.03188925e-05, 2.56310814e-05, -6.83582912e-05],
[ 2.09689583e-03, 3.47510446e-04, -1.01099804e-03],
[ 1.65391876e-03, 2.65299343e-04, -7.74377258e-04]], dtype=float32),
10.5: array([[-4.8757210e-02, 6.9497101e-02, 1.5730826e-02],
[-2.4018072e-02, 7.8169808e-02, 7.4425046e-03],
[ 3.1567704e-02, -1.3658900e-02, -1.5672835e-02],
[ 2.6121164e-02, -4.5375745e-03, -1.1833688e-02],
[ 1.9431807e-01, -5.7702754e-02, -8.6174287e-02],
[ 1.2168377e-01, -3.5814829e-02, -5.3901821e-02],
[-1.4963685e-01, 4.2855527e-04, 4.9897294e-02],
[-2.0034360e-02, 3.3924285e-02, 6.3202917e-03],
[-1.3207954e-03, -3.7473816e-02, -7.3875422e-03],
[ 4.3170106e-02, -1.3984833e-02, -1.9787570e-02],
[ 2.9604992e-01, -8.9963824e-02, -1.3169578e-01],
[ 2.0726989e-01, -6.1179761e-02, -9.1840461e-02],
[-9.9732459e-02, -2.1411840e-02, 3.3515956e-02],
[-7.8196945e-03, 3.3621307e-04, 2.5796464e-03],
[-3.2952208e-02, -1.5724033e-02, 9.4598690e-03],
[ 6.7594410e-03, -3.8816207e-03, -3.3972787e-03],
[ 7.0785038e-02, -2.2536065e-02, -3.1694245e-02],
[ 4.4083696e-02, -1.3100771e-02, -1.9550726e-02],
[-8.0675259e-03, -1.8701344e-03, 2.7157525e-03],
[-6.8342849e-04, -1.4468492e-04, 2.2899490e-04],
[-3.1179921e-03, -9.2415448e-04, 1.0103692e-03],
[-3.8916140e-05, -1.6054211e-04, -1.7538892e-05],
[ 1.5845250e-03, -6.2438904e-04, -7.3359744e-04],
[ 1.3114482e-03, -3.9846703e-04, -5.8336934e-04]], dtype=float32),
12.5: array([[-4.9026076e-02, 7.5596273e-02, 1.3509233e-02],
[-2.7129848e-02, 7.7647403e-02, 6.5284329e-03],
[ 4.2978253e-02, 2.7397834e-02, -2.5928102e-02],
[ 3.2106433e-02, 1.8230369e-02, -1.7639322e-02],
[ 2.3618002e-01, 9.8345540e-02, -1.2566811e-01],
[ 1.4771535e-01, 6.1305471e-02, -7.8489378e-02],
[-1.3636662e-01, 3.5028521e-02, 4.2505633e-02],
[-2.0239921e-02, 3.6955986e-02, 5.1915227e-03],
[ 2.1850543e-02, 3.8795412e-02, -2.5756272e-02],
[ 5.4200146e-02, 2.6648207e-02, -3.0027075e-02],
[ 3.6103404e-01, 1.5174034e-01, -1.9281840e-01],
[ 2.5169381e-01, 1.0451723e-01, -1.3378537e-01],
[-8.9690827e-02, 2.4151439e-03, 2.8743092e-02],
[-7.0823664e-03, 2.3507976e-03, 2.1370309e-03],
[-2.4802126e-02, 7.9431664e-03, 4.0907701e-03],
[ 9.3893502e-03, 5.4194313e-03, -5.7048267e-03],
[ 8.6932711e-02, 3.7259728e-02, -4.6791013e-02],
[ 5.3583842e-02, 2.2310656e-02, -2.8512705e-02],
[-7.2550736e-03, 3.5648263e-05, 2.3372378e-03],
[-6.1307452e-04, 2.4406820e-05, 1.9480936e-04],
[-2.6872866e-03, 1.9506851e-04, 7.7207515e-04],
[ 5.5651617e-05, 1.4666974e-04, -9.1101145e-05],
[ 2.0174093e-03, 9.4885059e-04, -1.1280269e-03],
[ 1.5992624e-03, 6.7205599e-04, -8.5408695e-04]], dtype=float32),
14.5: array([[-5.62312156e-02, 6.66079819e-02, 1.78768020e-02],
[-2.88238302e-02, 7.87258521e-02, 1.03787519e-02],
[ 1.06634526e-02, -3.46612073e-02, -1.50409574e-02],
[ 1.42465457e-02, -1.61629319e-02, -1.11742923e-02],
[ 1.14596710e-01, -1.37472108e-01, -8.25117081e-02],
[ 7.20540658e-02, -8.54599476e-02, -5.16028441e-02],
[-1.66496143e-01, -1.74111500e-02, 4.73619476e-02],
[-2.37447545e-02, 3.24901976e-02, 7.44536752e-03],
[-3.88014987e-02, -7.66016692e-02, -7.96739012e-03],
[ 2.24072114e-02, -3.47595327e-02, -1.89395919e-02],
[ 1.72652587e-01, -2.13526696e-01, -1.26149103e-01],
[ 1.22604765e-01, -1.45879999e-01, -8.79283473e-02],
[-1.09902002e-01, -3.37808765e-02, 3.08841653e-02],
[-8.83033779e-03, -7.00296834e-04, 2.46389676e-03],
[-4.40164134e-02, -2.79171262e-02, 8.43676087e-03],
[ 2.08055787e-03, -8.64310376e-03, -3.29670077e-03],
[ 4.02982272e-02, -5.31083420e-02, -3.03843711e-02],
[ 2.59933881e-02, -3.12025528e-02, -1.87201258e-02],
[-8.86945799e-03, -2.86019314e-03, 2.49648746e-03],
[-7.56306516e-04, -2.32408333e-04, 2.11133258e-04],
[-3.61513323e-03, -1.50316081e-03, 9.22265346e-04],
[-1.88939666e-04, -3.18223378e-04, -2.09808259e-05],
[ 7.87009834e-04, -1.42917549e-03, -7.06179359e-04],
[ 7.64912576e-04, -9.45734559e-04, -5.58798725e-04]], dtype=float32),
16.5: array([[-5.21655194e-02, 6.40237182e-02, 1.93627309e-02],
[-3.16902809e-02, 7.26106390e-02, 1.19688921e-02],
[ 4.96370718e-02, -1.94649082e-02, -1.71494689e-02],
[ 3.56320590e-02, -8.14305898e-03, -1.23923291e-02],
[ 2.61876225e-01, -8.06191713e-02, -9.14763287e-02],
[ 1.63702965e-01, -5.00979945e-02, -5.71891367e-02],
[-1.31571516e-01, -2.31336569e-03, 4.78092469e-02],
[-2.17617769e-02, 3.11784036e-02, 8.12539179e-03],
[ 3.48662585e-02, -4.62748185e-02, -1.13952067e-02],
[ 6.08260110e-02, -1.99029520e-02, -2.11712308e-02],
[ 4.00864840e-01, -1.25327602e-01, -1.39987513e-01],
[ 2.78974354e-01, -8.55365321e-02, -9.74565297e-02],
[-8.51455703e-02, -2.16754340e-02, 3.07945814e-02],
[-6.82028104e-03, 1.36852919e-04, 2.48482870e-03],
[-2.05539037e-02, -1.75696015e-02, 7.67942611e-03],
[ 1.09425485e-02, -5.12905791e-03, -3.78289842e-03],
[ 9.68006775e-02, -3.12202554e-02, -3.37842852e-02],
[ 5.94153404e-02, -1.83003880e-02, -2.07544006e-02],
[-6.88259117e-03, -1.87742861e-03, 2.48775375e-03],
[-5.81215310e-04, -1.47543673e-04, 2.10338039e-04],
[-2.47725518e-03, -9.70211113e-04, 9.01394291e-04],
[ 1.08601773e-04, -1.94662134e-04, -3.45606895e-05],
[ 2.27862410e-03, -8.45583505e-04, -7.92958133e-04],
[ 1.77567638e-03, -5.55098406e-04, -6.20092615e-04]], dtype=float32),
19.0: array([[-4.59347032e-02, 6.55248985e-02, 1.53686032e-02],
[-2.22758073e-02, 7.61762410e-02, 7.08056707e-03],
[ 4.11518738e-02, -2.83000842e-02, -1.68732218e-02],
[ 3.16018686e-02, -1.28022414e-02, -1.25626447e-02],
[ 2.31316939e-01, -1.13650046e-01, -9.10786539e-02],
[ 1.44720778e-01, -7.06425831e-02, -5.69585972e-02],
[-1.42962530e-01, -1.11476332e-02, 4.97077629e-02],
[-1.86069719e-02, 3.19416933e-02, 6.12278795e-03],
[ 1.55719062e-02, -6.39298782e-02, -9.23464354e-03],
[ 5.27594611e-02, -2.85367202e-02, -2.10333858e-02],
[ 3.53293657e-01, -1.76572070e-01, -1.39263317e-01],
[ 2.46568039e-01, -1.20595016e-01, -9.70534384e-02],
[-9.56869349e-02, -2.87563242e-02, 3.35070267e-02],
[-7.41398940e-03, -3.52778676e-04, 2.56207446e-03],
[-2.81384345e-02, -2.36046724e-02, 9.07868054e-03],
[ 8.90567061e-03, -7.17244949e-03, -3.66219599e-03],
[ 8.49151909e-02, -4.39383462e-02, -3.35520916e-02],
[ 5.24794050e-02, -2.57963967e-02, -2.06635520e-02],
[-7.74906436e-03, -2.45229690e-03, 2.71650450e-03],
[-6.54285599e-04, -1.97177767e-04, 2.28761128e-04],
[-2.91044964e-03, -1.28153292e-03, 1.00144395e-03],
[ 2.85412298e-05, -2.66606890e-04, -2.47465268e-05],
[ 1.95273012e-03, -1.18478341e-03, -7.80861068e-04],
[ 1.56498631e-03, -7.82061426e-04, -6.16887526e-04]], dtype=float32),
21.0: array([[-5.34694009e-02, 6.47808835e-02, 1.88612193e-02],
[-3.00062969e-02, 7.37764686e-02, 1.05850603e-02],
[ 3.36500965e-02, -2.10885629e-02, -1.21319611e-02],
[ 2.69101467e-02, -8.98332428e-03, -9.65036079e-03],
[ 2.01657698e-01, -8.67570490e-02, -7.22247586e-02],
[ 1.26233414e-01, -5.39153367e-02, -4.52082418e-02],
[-1.46374479e-01, -3.78774083e-03, 5.16380034e-02],
[-2.23854743e-02, 3.15522030e-02, 7.89120048e-03],
[ 4.36447095e-03, -4.95723784e-02, -1.94246427e-03],
[ 4.51014526e-02, -2.14974973e-02, -1.61841102e-02],
[ 3.07534784e-01, -1.34851545e-01, -1.10164531e-01],
[ 2.15042859e-01, -9.20510143e-02, -7.70150423e-02],
[-9.57809612e-02, -2.30147745e-02, 3.37948576e-02],
[-7.66737340e-03, 5.71863129e-05, 2.70362059e-03],
[-3.04013602e-02, -1.87080652e-02, 1.06433565e-02],
[ 7.30020553e-03, -5.50947059e-03, -2.63362238e-03],
[ 7.36833066e-02, -3.35846543e-02, -2.64044777e-02],
[ 4.57499400e-02, -1.96933504e-02, -1.63856242e-02],
[-7.73749081e-03, -1.98709266e-03, 2.73023639e-03],
[-6.56313612e-04, -1.56898532e-04, 2.31537328e-04],
[-2.96092848e-03, -1.02933147e-03, 1.04276650e-03],
[-1.47705596e-05, -2.08145095e-04, 3.63551680e-06],
[ 1.66721246e-03, -9.08721529e-04, -5.98590472e-04],
[ 1.36231352e-03, -5.97280334e-04, -4.88004356e-04]], dtype=float32),
23.0: array([[-4.74029854e-02, 7.53735080e-02, 1.37492567e-02],
[-2.56043747e-02, 8.14283863e-02, 5.38585661e-03],
[ 4.71761115e-02, 7.36717833e-03, -1.90913789e-02],
[ 3.46303321e-02, 7.40389153e-03, -1.39468461e-02],
[ 2.53041953e-01, 2.30490603e-02, -1.00044124e-01],
[ 1.58222675e-01, 1.44590922e-02, -6.25477061e-02],
[-1.35022387e-01, 1.60772149e-02, 4.91505004e-02],
[-1.93890817e-02, 3.68778966e-02, 5.29493392e-03],
[ 2.85062119e-02, 7.61373376e-05, -1.25071043e-02],
[ 5.85115030e-02, 6.98140077e-03, -2.33259350e-02],
[ 3.87067616e-01, 3.50093320e-02, -1.53090373e-01],
[ 2.69613773e-01, 2.45837457e-02, -1.06584072e-01],
[-8.92505273e-02, -1.16661489e-02, 3.36325280e-02],
[-6.98387204e-03, 1.27635687e-03, 2.51374068e-03],
[-2.32962575e-02, -4.76788869e-03, 8.46205000e-03],
[ 1.03155095e-02, 8.36531573e-04, -4.14155424e-03],
[ 9.33317393e-02, 8.33029673e-03, -3.69434394e-02],
[ 5.74099235e-02, 5.22341300e-03, -2.26977803e-02],
[-7.22499518e-03, -1.09984435e-03, 2.73128413e-03],
[-6.09428389e-04, -7.47932936e-05, 2.29248981e-04],
[-2.64250208e-03, -4.38614137e-04, 9.91008594e-04],
[ 8.17513646e-05, -1.02771046e-05, -3.73827461e-05],
[ 2.18107807e-03, 1.81943236e-04, -8.66769929e-04],
[ 1.71456963e-03, 1.55049827e-04, -6.78130251e-04]], dtype=float32),
24.5: array([[-5.21849580e-02, 6.94204792e-02, 1.71629451e-02],
[-3.21539268e-02, 7.88262188e-02, 9.97947436e-03],
[ 5.13067357e-02, -1.69938095e-02, -2.10437737e-02],
[ 3.64992805e-02, -6.31976733e-03, -1.47058554e-02],
[ 2.67939478e-01, -7.00942576e-02, -1.06694013e-01],
[ 1.67472646e-01, -4.35226820e-02, -6.66688383e-02],
[-1.29481524e-01, -3.07125528e-03, 4.55650426e-02],
[-2.17853710e-02, 3.38995270e-02, 7.01678824e-03],
[ 3.83132249e-02, -4.40529101e-02, -1.78156458e-02],
[ 6.24298416e-02, -1.72355156e-02, -2.51053050e-02],
[ 4.10281867e-01, -1.09184913e-01, -1.63502961e-01],
[ 2.85407811e-01, -7.43323565e-02, -1.13625064e-01],
[-8.35690200e-02, -2.40076222e-02, 2.98420731e-02],
[-6.70464896e-03, 1.39366050e-04, 2.33830395e-03],
[-1.93141419e-02, -1.79511402e-02, 6.05830597e-03],
[ 1.13280043e-02, -4.64617414e-03, -4.63960925e-03],
[ 9.91430134e-02, -2.73049455e-02, -3.95746604e-02],
[ 6.07913546e-02, -1.59128401e-02, -2.42074095e-02],
[-6.75499719e-03, -2.07962329e-03, 2.41676695e-03],
[-5.70183678e-04, -1.62928642e-04, 2.03180927e-04],
[-2.41058297e-03, -1.03822164e-03, 8.42074689e-04],
[ 1.22699304e-04, -1.87283847e-04, -5.98858605e-05],
[ 2.34167930e-03, -7.51410960e-04, -9.42234590e-04],
[ 1.81738404e-03, -4.83597483e-04, -7.24245561e-04]], dtype=float32),
27.0: array([[-2.57173806e-01, -1.87496886e-01, 1.99893862e-01],
[-2.37589449e-01, -1.88666657e-01, 1.96282163e-01],
[ 1.76368896e-02, -1.64031535e-02, 7.41399417e-04],
[ 1.31711755e-02, -1.28956996e-02, 2.12298054e-03],
[ 1.80815890e-01, -2.12673601e-02, -5.68992868e-02],
[ 1.13264076e-01, -1.31841954e-02, -3.56484838e-02],
[-2.50186354e-01, -1.07096143e-01, 1.38054252e-01],
[-1.20627031e-01, -9.01997611e-02, 9.53026861e-02],
[-1.47120915e-02, -2.48913001e-02, 1.05036059e-02],
[ 3.44923511e-02, -1.07087567e-02, -7.77327176e-03],
[ 2.75086939e-01, -3.34226526e-02, -8.64328891e-02],
[ 1.92997232e-01, -2.24427860e-02, -6.07887171e-02],
[-1.10545851e-01, -2.24002711e-02, 4.24674936e-02],
[-1.42064542e-02, -6.62138034e-03, 8.20749160e-03],
[-3.68605964e-02, -1.04015870e-02, 1.39667550e-02],
[ 5.55636175e-03, -2.04444537e-03, -1.42135913e-03],
[ 6.55401051e-02, -8.53826851e-03, -2.05066241e-02],
[ 4.10329551e-02, -4.81561199e-03, -1.29195396e-02],
[-8.54563341e-03, -1.46173069e-03, 3.08625586e-03],
[-7.71460822e-04, -1.66671278e-04, 3.03158478e-04],
[-3.33251990e-03, -6.56519784e-04, 1.21027674e-03],
[-7.84134245e-05, -9.06288915e-05, 4.08598135e-05],
[ 1.43937476e-03, -2.55599472e-04, -4.39950789e-04],
[ 1.21870544e-03, -1.47920626e-04, -3.82983300e-04]], dtype=float32),
30.0: array([[-5.28024249e-02, 7.60688484e-02, 1.64442603e-02],
[-3.38067524e-02, 7.70589113e-02, 9.41110589e-03],
[ 5.30809574e-02, 3.37698162e-02, -2.27727070e-02],
[ 3.72355320e-02, 2.17467919e-02, -1.57114826e-02],
[ 2.73402005e-01, 1.22621648e-01, -1.13403164e-01],
[ 1.70855373e-01, 7.64138550e-02, -7.08471164e-02],
[-1.24901772e-01, 4.02735248e-02, 4.44496274e-02],
[-2.21568402e-02, 3.71960141e-02, 6.65420666e-03],
[ 4.30909172e-02, 5.06936684e-02, -2.07765941e-02],
[ 6.39678538e-02, 3.29606943e-02, -2.68423595e-02],
[ 4.18853760e-01, 1.89344183e-01, -1.73877582e-01],
[ 2.91187584e-01, 1.30294040e-01, -1.20752156e-01],
[-7.98985586e-02, 6.16739737e-03, 2.92464122e-02],
[-6.46356540e-03, 2.65443302e-03, 2.26853252e-03],
[-1.70934908e-02, 1.16527323e-02, 5.25502907e-03],
[ 1.17585678e-02, 6.86789863e-03, -5.02374768e-03],
[ 1.01318538e-01, 4.65637632e-02, -4.21327874e-02],
[ 6.20313548e-02, 2.78195776e-02, -2.57298071e-02],
[-6.45545637e-03, 3.36756289e-04, 2.37066159e-03],
[-5.44800190e-04, 5.09952224e-05, 1.98832306e-04],
[-2.27061426e-03, 3.71227128e-04, 8.09494290e-04],
[ 1.42849429e-04, 1.94648281e-04, -7.16485447e-05],
[ 2.40512588e-03, 1.19376765e-03, -1.00858230e-03],
[ 1.85534544e-03, 8.38605920e-04, -7.70195853e-04]], dtype=float32),
43.0: array([[-8.31093043e-02, 7.33238608e-02, 2.66411006e-02],
[-5.87848537e-02, 8.07350576e-02, 1.65618379e-02],
[ 7.32064098e-02, 2.20983401e-02, -3.86417955e-02],
[ 5.06204143e-02, 1.66853424e-02, -2.60952786e-02],
[ 3.75989616e-01, 8.76939967e-02, -1.88348755e-01],
[ 2.34931186e-01, 5.47837503e-02, -1.17633246e-01],
[-1.69423997e-01, 7.91978277e-03, 6.57091364e-02],
[-3.56833003e-02, 3.63395847e-02, 1.08869895e-02],
[ 6.30458519e-02, 1.99661292e-02, -3.89608741e-02],
[ 8.79732370e-02, 2.29629278e-02, -4.48813774e-02],
[ 5.76231122e-01, 1.34535596e-01, -2.89020211e-01],
[ 4.00411844e-01, 9.33396742e-02, -2.00511187e-01],
[-1.04767330e-01, -1.75298117e-02, 4.25864235e-02],
[-8.82458221e-03, 9.16570076e-04, 3.34213162e-03],
[-2.08945833e-02, -2.57997052e-03, 5.81681542e-03],
[ 1.63566824e-02, 4.10402101e-03, -8.56631808e-03],
[ 1.39492139e-01, 3.26525830e-02, -7.01491609e-02],
[ 8.53084996e-02, 1.98913999e-02, -4.27347533e-02],
[-8.44082795e-03, -1.58423057e-03, 3.45049263e-03],
[-7.14718481e-04, -1.13302092e-04, 2.88988434e-04],
[-2.93271802e-03, -5.25636307e-04, 1.13390712e-03],
[ 2.14069485e-04, 6.55682597e-05, -1.38651434e-04],
[ 3.32340202e-03, 7.87954370e-04, -1.69261871e-03],
[ 2.55245715e-03, 5.95884223e-04, -1.28021394e-03]], dtype=float32),
50.0: array([[-1.50551707e-01, -2.43838802e-02, 1.08773768e-01],
[-1.28343940e-01, -2.02106275e-02, 1.02698557e-01],
[ 2.40156706e-02, -1.00590494e-02, -7.73886545e-03],
[ 1.93407573e-02, -5.09899762e-03, -5.04588755e-03],
[ 1.84579939e-01, -2.70465538e-02, -7.31809214e-02],
[ 1.15611807e-01, -1.67552438e-02, -4.58031893e-02],
[-1.98662132e-01, -3.46042514e-02, 9.51672420e-02],
[-6.91822022e-02, -1.14235841e-02, 5.12526631e-02],
[-8.98227189e-03, -2.40133740e-02, 1.39843882e-03],
[ 3.81425358e-02, -8.26211832e-03, -1.41441021e-02],
[ 2.80963778e-01, -4.24203761e-02, -1.11578934e-01],
[ 1.96957052e-01, -2.86140051e-02, -7.80666545e-02],
[-1.04935400e-01, -1.86039899e-02, 3.86048928e-02],
[-1.09383631e-02, -1.96439447e-03, 5.45886531e-03],
[-3.50670815e-02, -1.08874179e-02, 1.18733337e-02],
[ 6.00524805e-03, -2.15752306e-03, -2.47487845e-03],
[ 6.70347661e-02, -1.07601946e-02, -2.67064683e-02],
[ 4.18803990e-02, -6.13775104e-03, -1.66081954e-02],
[-8.29413999e-03, -1.46537297e-03, 2.94940220e-03],
[-7.25980557e-04, -1.30383080e-04, 2.70306773e-04],
[-3.22573446e-03, -6.83001708e-04, 1.12806552e-03],
[-6.25675675e-05, -9.90892368e-05, 1.11888103e-05],
[ 1.48356601e-03, -3.13210301e-04, -6.00900443e-04],
[ 1.24468061e-03, -1.87850033e-04, -4.94315405e-04]], dtype=float32),
52.0: array([[-4.91133891e-02, 8.06416720e-02, 1.11895762e-02],
[-2.32583582e-02, 8.30661356e-02, 3.09890253e-03],
[ 2.30978876e-02, 3.18676680e-02, -2.17700694e-02],
[ 2.13583875e-02, 2.11142786e-02, -1.55127123e-02],
[ 1.61383122e-01, 1.16389431e-01, -1.09977141e-01],
[ 1.01179324e-01, 7.25579783e-02, -6.87318444e-02],
[-1.55023068e-01, 3.61581333e-02, 4.63572629e-02],
[-2.02549938e-02, 3.95011269e-02, 4.04526154e-03],
[-1.65471025e-02, 4.49683331e-02, -1.71896312e-02],
[ 3.46710049e-02, 3.12764384e-02, -2.59507596e-02],
[ 2.45080695e-01, 1.79544538e-01, -1.68463528e-01],
[ 1.72290042e-01, 1.23702593e-01, -1.17132947e-01],
[-1.03591882e-01, 1.60352362e-03, 3.23515460e-02],
[-8.14015046e-03, 2.45829555e-03, 2.34287395e-03],
[-3.73880826e-02, 8.89103487e-03, 7.09555717e-03],
[ 4.84012673e-03, 6.36407034e-03, -4.72561875e-03],
[ 5.81976064e-02, 4.40674983e-02, -4.07409742e-02],
[ 3.66099738e-02, 2.64045410e-02, -2.49516610e-02],
[-8.37621372e-03, -4.31097687e-05, 2.63336883e-03],
[-7.11010653e-04, 1.96915262e-05, 2.19969777e-04],
[-3.31382640e-03, 1.94715147e-04, 9.30072332e-04],
[-9.99785334e-05, 1.70125524e-04, -5.59964938e-05],
[ 1.25583389e-03, 1.11997011e-03, -9.66048159e-04],
[ 1.08570023e-03, 7.95206113e-04, -7.46218371e-04]], dtype=float32),
54.0: array([[-5.5085484e-02, 7.2300032e-02, 1.5592797e-02],
[-3.3175237e-02, 7.7256247e-02, 9.8006269e-03],
[ 3.7426513e-02, 9.2738131e-03, -2.9302945e-02],
[ 2.8677396e-02, 8.1656072e-03, -1.9272178e-02],
[ 2.1467881e-01, 2.9582797e-02, -1.3787527e-01],
[ 1.3431680e-01, 1.8509412e-02, -8.6073555e-02],
[-1.3995038e-01, 1.9476853e-02, 3.8111515e-02],
[-2.3250423e-02, 3.5326723e-02, 6.2490362e-03],
[ 1.3234792e-02, 5.1563489e-03, -3.3280060e-02],
[ 4.8603624e-02, 8.7274192e-03, -3.3246834e-02],
[ 3.2783690e-01, 4.5232575e-02, -2.1181096e-01],
[ 2.2884470e-01, 3.1503078e-02, -1.4673212e-01],
[-9.0418272e-02, -8.1247929e-03, 2.4779795e-02],
[-7.3270537e-03, 1.4438650e-03, 1.9057610e-03],
[-2.6757060e-02, -2.4985259e-03, 1.1864156e-03],
[ 8.2008038e-03, 1.3191983e-03, -6.5094307e-03],
[ 7.8771815e-02, 1.0910049e-02, -5.1531296e-02],
[ 4.8705336e-02, 6.7066327e-03, -3.1283259e-02],
[-7.2992793e-03, -8.0627087e-04, 2.0120780e-03],
[-6.1911909e-04, -5.0421251e-05, 1.6729841e-04],
[-2.7465820e-03, -2.9898912e-04, 6.0813391e-04],
[ 2.2129523e-05, 1.1209732e-05, -1.2223655e-04],
[ 1.8085424e-03, 2.5553169e-04, -1.2574503e-03],
[ 1.4522284e-03, 2.0032671e-04, -9.3820412e-04]], dtype=float32),
57.0: array([[-4.74663042e-02, 6.55223355e-02, 1.67945623e-02],
[-2.67220959e-02, 7.58264884e-02, 9.49087087e-03],
[ 5.36576807e-02, -2.60515325e-02, -2.12841835e-02],
[ 3.82074006e-02, -1.15551874e-02, -1.48937749e-02],
[ 2.77929246e-01, -1.04983807e-01, -1.07841186e-01],
[ 1.73712447e-01, -6.52483255e-02, -6.73862994e-02],
[-1.30135700e-01, -9.53046046e-03, 4.59290966e-02],
[-1.94026232e-02, 3.19482908e-02, 6.82266522e-03],
[ 4.05535959e-02, -5.97872846e-02, -1.79493669e-02],
[ 6.49608597e-02, -2.62938403e-02, -2.53841225e-02],
[ 4.25611168e-01, -1.63153216e-01, -1.65257126e-01],
[ 2.96041012e-01, -1.11392029e-01, -1.14847325e-01],
[-8.56051072e-02, -2.75576115e-02, 3.02417297e-02],
[-6.70109084e-03, -2.58593063e-04, 2.35386821e-03],
[-1.95850823e-02, -2.23470535e-02, 6.16824906e-03],
[ 1.17927799e-02, -6.66065235e-03, -4.68676537e-03],
[ 1.02864698e-01, -4.06208336e-02, -3.99974138e-02],
[ 6.30575567e-02, -2.38297917e-02, -2.44676396e-02],
[-6.93166209e-03, -2.35595321e-03, 2.45023984e-03],
[-5.83590474e-04, -1.88665697e-04, 2.05872653e-04],
[-2.46850424e-03, -1.22345448e-03, 8.54415179e-04],
[ 1.30317349e-04, -2.49929377e-04, -6.02282453e-05],
[ 2.43161572e-03, -1.09775004e-03, -9.52090544e-04],
[ 1.88528211e-03, -7.22628203e-04, -7.32015527e-04]], dtype=float32),
65.5: array([[-5.45864515e-02, 7.19663873e-02, 1.68590024e-02],
[-2.76965071e-02, 7.94319510e-02, 6.84002042e-03],
[ 1.89041682e-02, -6.13191212e-03, -5.49137872e-03],
[ 1.90771762e-02, -3.42367013e-04, -6.22080313e-03],
[ 1.47446305e-01, -2.94366311e-02, -4.75542843e-02],
[ 9.25176144e-02, -1.82236247e-02, -2.98675653e-02],
[-1.63215026e-01, 7.55440770e-03, 5.86733110e-02],
[-2.28474382e-02, 3.51275243e-02, 6.87859813e-03],
[-2.51041297e-02, -2.34901085e-02, 1.16982860e-02],
[ 3.08185369e-02, -6.58289297e-03, -9.72666033e-03],
[ 2.23410234e-01, -4.61754799e-02, -7.18655288e-02],
[ 1.57508448e-01, -3.11675314e-02, -5.08345887e-02],
[-1.08058140e-01, -1.68717541e-02, 3.96961719e-02],
[-8.61234870e-03, 7.53248343e-04, 3.08720977e-03],
[-4.05845456e-02, -1.13435620e-02, 1.54597256e-02],
[ 3.92091693e-03, -2.18805484e-03, -1.08500372e-03],
[ 5.27941845e-02, -1.16995703e-02, -1.68870464e-02],
[ 3.34473997e-02, -6.68651704e-03, -1.07867820e-02],
[-8.72755144e-03, -1.50928146e-03, 3.21124215e-03],
[-7.42514268e-04, -1.12427348e-04, 2.72768433e-04],
[-3.49171809e-03, -7.14770751e-04, 1.29770511e-03],
[-1.34732982e-04, -1.04261424e-04, 5.94535886e-05],
[ 1.10887107e-03, -3.38869577e-04, -3.43379943e-04],
[ 9.89726163e-04, -2.04526586e-04, -3.18377715e-04]], dtype=float32),
93.0: array([[-5.47065213e-02, 8.76432136e-02, 1.10226702e-02],
[-2.82389317e-02, 8.52111652e-02, 3.18982336e-03],
[ 1.20213795e-02, 6.32044673e-02, -2.21188050e-02],
[ 1.47918509e-02, 3.85870300e-02, -1.56444255e-02],
[ 1.17953427e-01, 2.35274523e-01, -1.10865153e-01],
[ 7.41251633e-02, 1.46551743e-01, -6.92796782e-02],
[-1.61074236e-01, 6.30297214e-02, 4.50209081e-02],
[-2.30889767e-02, 4.29594107e-02, 3.98615375e-03],
[-3.48297171e-02, 1.02813758e-01, -1.82399936e-02],
[ 2.34617442e-02, 6.22695982e-02, -2.62243338e-02],
[ 1.77971140e-01, 3.63669395e-01, -1.69870868e-01],
[ 1.26146346e-01, 2.49941081e-01, -1.18069865e-01],
[-1.06256828e-01, 1.94018632e-02, 3.13872769e-02],
[-8.53617862e-03, 4.03262582e-03, 2.27144687e-03],
[-4.19964790e-02, 2.67219469e-02, 6.55847788e-03],
[ 2.39738869e-03, 1.34377452e-02, -4.80935024e-03],
[ 4.16734256e-02, 8.96105692e-02, -4.11049873e-02],
[ 2.67554075e-02, 5.33824638e-02, -2.51532663e-02],
[-8.57576728e-03, 1.37539778e-03, 2.55522272e-03],
[-7.31025357e-04, 1.46225808e-04, 2.13297055e-04],
[-3.48186400e-03, 1.03273895e-03, 8.94600875e-04],
[-1.72130458e-04, 4.02828999e-04, -6.04612251e-05],
[ 8.29916156e-04, 2.31731404e-03, -9.77417687e-04],
[ 7.88463338e-04, 1.61070633e-03, -7.52450200e-04]], dtype=float32)}}
[27]:
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split
from sklearn.metrics import r2_score
X = []
y = []
subjects_ids = ["1004B", "1001B"]
for subject_id in subjects_ids:
subject_data = vel_week[subject_id]
for gest_week, velocity_matrix in subject_data.items():
# print(velocity_matrix)
# print(gest_week)
X.append(gest_week)
y.append(velocity_matrix.flatten())
X = np.array(X).reshape(-1, 1)
y = np.array(y)
print(f"shape of feature matrix X (weeks): {X.shape}")
print(f"shape of target matrix y (flattened velocities): {y.shape}")
shape of feature matrix X (weeks): (36, 1)
shape of target matrix y (flattened velocities): (36, 72)
[28]:
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.3, random_state=42
)
# try regular linear regression with velocities -- still bad r2 values
model = LinearRegression()
model.fit(X_train, y_train)
print(f"shape of coefficients (slopes): {model.coef_.shape}")
print(f"shape of model intercept: {model.intercept_.shape}")
# Evaluate the model on the test set
test_score = model.score(X_test, y_test)
print(f"\nR2 score on the test set: {test_score:.3f}")
shape of coefficients (slopes): (72, 1)
shape of model intercept: (72,)
R2 score on the test set: -0.035
[29]:
X_train_list = []
y_train_list = []
subjects_to_train = ["1004B",]# "1001B"]
# wait i don't remember why i tried training on just one subject
for subject_id in subjects_to_train:
subject_data = vel_week[subject_id]
for gest_week, velocity_matrix in subject_data.items():
X_train_list.append(gest_week)
y_train_list.append(velocity_matrix.flatten())
# test data from subject 01
X_test_list = []
y_test_list = []
subject_to_test = "01"
subject_data_test = vel_week[subject_to_test]
for gest_week, velocity_matrix in subject_data_test.items():
X_test_list.append(gest_week)
y_test_list.append(velocity_matrix.flatten())
X_train = np.array(X_train_list).reshape(-1, 1)
y_train = np.array(y_train_list)
X_test = np.array(X_test_list).reshape(-1, 1)
y_test = np.array(y_test_list)
print(f"shape of training feature matrix X (weeks): {X_train.shape}")
print(f"shape of training target matrix y (vlocities): {y_train.shape}")
print(f"shape of test feature matrix X (week): {X_test.shape}")
print(f"shape of test target matrix y (velocities): {y_test.shape}")
# now trying nonlinear regression (and ridge for overfitting)
from sklearn.linear_model import Ridge
from sklearn.preprocessing import SplineTransformer
from sklearn.pipeline import make_pipeline
n_knots = 4
degree = 3
model = make_pipeline(
SplineTransformer(n_knots=n_knots, degree=degree),
LinearRegression()
)
# model = Ridge(alpha=1.0)
# model = LinearRegression()
model.fit(X_train, y_train)
# still terrible R2
test_score = model.score(X_test, y_test)
print(f"\nR2 score on the test set (Subject '01'): {test_score:.3f}")
shape of training feature matrix X (weeks): (22, 1)
shape of training target matrix y (vlocities): (22, 72)
shape of test feature matrix X (week): (26, 1)
shape of test target matrix y (velocities): (26, 72)
R2 score on the test set (Subject '01'): -1.761
[30]:
import matplotlib.pyplot as plt
component_to_plot = 10 # picked 10 for no reason in particular but wanted to plot something simple rather than 72 plots for each velocity and each dimension
X_plot = np.linspace(0, 40, 100).reshape(-1, 1)
y_plot_pred = model.predict(X_plot)
plt.figure(figsize=(10, 6))
plt.scatter(X_train, y_train[:, component_to_plot], label='training set (Subjects 1004B, 1001B)', c='blue')
plt.scatter(X_test, y_test[:, component_to_plot], label="testing set (Subject '01')", c='red', marker='x', s=100)
plt.plot(X_plot, y_plot_pred[:, component_to_plot], color='green', linewidth=2, label='spline prediction')
plt.title(f'spline fit for velocity comp #{component_to_plot}')
plt.xlabel('Gestational week')
plt.ylabel('Velocity component value')
plt.legend()
plt.grid(True)
plt.show()
Loaded backend module://matplotlib_inline.backend_inline version unknown.
Loaded backend module://matplotlib_inline.backend_inline version unknown.
findfont: Matching sans\-serif:style=normal:variant=normal:weight=normal:stretch=normal:size=10.0.
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizTwoSymReg.ttf', name='STIXSizeTwoSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralBol.ttf', name='STIXGeneral', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-Oblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 1.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniBol.ttf', name='STIXNonUnicode', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizOneSymBol.ttf', name='STIXSizeOneSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansDisplay.ttf', name='DejaVu Sans Display', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralItalic.ttf', name='STIXGeneral', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizThreeSymBol.ttf', name='STIXSizeThreeSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-Bold.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmss10.ttf', name='cmss10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-BoldItalic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-Italic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 0.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-Bold.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 0.33499999999999996
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-Bold.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmsy10.ttf', name='cmsy10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmex10.ttf', name='cmex10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmb10.ttf', name='cmb10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerifDisplay.ttf', name='DejaVu Serif Display', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-Oblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmr10.ttf', name='cmr10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFiveSymReg.ttf', name='STIXSizeFiveSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmmi10.ttf', name='cmmi10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-BoldOblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 1.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralBolIta.ttf', name='STIXGeneral', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniIta.ttf', name='STIXNonUnicode', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-BoldOblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFourSymReg.ttf', name='STIXSizeFourSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUni.ttf', name='STIXNonUnicode', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniBolIta.ttf', name='STIXNonUnicode', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizThreeSymReg.ttf', name='STIXSizeThreeSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizTwoSymBol.ttf', name='STIXSizeTwoSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFourSymBol.ttf', name='STIXSizeFourSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizOneSymReg.ttf', name='STIXSizeOneSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneral.ttf', name='STIXGeneral', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmtt10.ttf', name='cmtt10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed-Oblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=400, stretch='condensed', size='scalable')) = 1.25
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-HeavyItalic.ttf', name='Lato', style='italic', variant='normal', weight=800, stretch='normal', size='scalable')) = 11.43
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed-Bold.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=700, stretch='condensed', size='scalable')) = 0.5349999999999999
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-HairlineItalic.ttf', name='Lato', style='italic', variant='normal', weight=100, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono-Oblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-ThinItalic.ttf', name='Lato', style='italic', variant='normal', weight=200, stretch='normal', size='scalable')) = 11.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Heavy.ttf', name='Lato', style='normal', variant='normal', weight=800, stretch='normal', size='scalable')) = 10.43
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif-Bold.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Bold.ttf', name='Lato', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-BoldItalic.ttf', name='Lato', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed-BoldItalic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=700, stretch='condensed', size='scalable')) = 11.535
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-SemiboldItalic.ttf', name='Lato', style='italic', variant='normal', weight=600, stretch='normal', size='scalable')) = 11.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Black.ttf', name='Lato', style='normal', variant='normal', weight=900, stretch='normal', size='scalable')) = 10.525
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed-Bold.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=700, stretch='condensed', size='scalable')) = 10.535
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 0.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-BoldOblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 1.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Italic.ttf', name='Lato', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed-BoldOblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=700, stretch='condensed', size='scalable')) = 1.535
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed-Italic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=400, stretch='condensed', size='scalable')) = 11.25
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Thin.ttf', name='Lato', style='normal', variant='normal', weight=200, stretch='normal', size='scalable')) = 10.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Regular.ttf', name='Lato', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Light.ttf', name='Lato', style='normal', variant='normal', weight=300, stretch='normal', size='scalable')) = 10.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono-Bold.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif-Italic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuMathTeXGyre.ttf', name='DejaVu Math TeX Gyre', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-BlackItalic.ttf', name='Lato', style='italic', variant='normal', weight=900, stretch='normal', size='scalable')) = 11.525
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-MediumItalic.ttf', name='Lato', style='italic', variant='normal', weight=500, stretch='normal', size='scalable')) = 11.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Medium.ttf', name='Lato', style='normal', variant='normal', weight=500, stretch='normal', size='scalable')) = 10.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-LightItalic.ttf', name='Lato', style='italic', variant='normal', weight=300, stretch='normal', size='scalable')) = 11.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-Oblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 1.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono-BoldOblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-Bold.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 0.33499999999999996
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif-BoldItalic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-ExtraLight.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=200, stretch='normal', size='scalable')) = 0.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Semibold.ttf', name='Lato', style='normal', variant='normal', weight=600, stretch='normal', size='scalable')) = 10.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Hairline.ttf', name='Lato', style='normal', variant='normal', weight=100, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=400, stretch='condensed', size='scalable')) = 10.25
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=400, stretch='condensed', size='scalable')) = 0.25
findfont: Matching sans\-serif:style=normal:variant=normal:weight=normal:stretch=normal:size=10.0 to DejaVu Sans ('/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans.ttf') with score of 0.050000.
findfont: Matching sans\-serif:style=normal:variant=normal:weight=normal:stretch=normal:size=12.0.
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizTwoSymReg.ttf', name='STIXSizeTwoSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralBol.ttf', name='STIXGeneral', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-Oblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 1.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniBol.ttf', name='STIXNonUnicode', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizOneSymBol.ttf', name='STIXSizeOneSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansDisplay.ttf', name='DejaVu Sans Display', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralItalic.ttf', name='STIXGeneral', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizThreeSymBol.ttf', name='STIXSizeThreeSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-Bold.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmss10.ttf', name='cmss10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-BoldItalic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-Italic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 0.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-Bold.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 0.33499999999999996
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-Bold.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmsy10.ttf', name='cmsy10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmex10.ttf', name='cmex10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmb10.ttf', name='cmb10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerifDisplay.ttf', name='DejaVu Serif Display', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-Oblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmr10.ttf', name='cmr10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFiveSymReg.ttf', name='STIXSizeFiveSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmmi10.ttf', name='cmmi10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-BoldOblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 1.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralBolIta.ttf', name='STIXGeneral', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniIta.ttf', name='STIXNonUnicode', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-BoldOblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFourSymReg.ttf', name='STIXSizeFourSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUni.ttf', name='STIXNonUnicode', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniBolIta.ttf', name='STIXNonUnicode', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizThreeSymReg.ttf', name='STIXSizeThreeSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizTwoSymBol.ttf', name='STIXSizeTwoSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFourSymBol.ttf', name='STIXSizeFourSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizOneSymReg.ttf', name='STIXSizeOneSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneral.ttf', name='STIXGeneral', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmtt10.ttf', name='cmtt10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed-Oblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=400, stretch='condensed', size='scalable')) = 1.25
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-HeavyItalic.ttf', name='Lato', style='italic', variant='normal', weight=800, stretch='normal', size='scalable')) = 11.43
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed-Bold.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=700, stretch='condensed', size='scalable')) = 0.5349999999999999
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-HairlineItalic.ttf', name='Lato', style='italic', variant='normal', weight=100, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono-Oblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-ThinItalic.ttf', name='Lato', style='italic', variant='normal', weight=200, stretch='normal', size='scalable')) = 11.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Heavy.ttf', name='Lato', style='normal', variant='normal', weight=800, stretch='normal', size='scalable')) = 10.43
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif-Bold.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Bold.ttf', name='Lato', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-BoldItalic.ttf', name='Lato', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed-BoldItalic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=700, stretch='condensed', size='scalable')) = 11.535
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-SemiboldItalic.ttf', name='Lato', style='italic', variant='normal', weight=600, stretch='normal', size='scalable')) = 11.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Black.ttf', name='Lato', style='normal', variant='normal', weight=900, stretch='normal', size='scalable')) = 10.525
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed-Bold.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=700, stretch='condensed', size='scalable')) = 10.535
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 0.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-BoldOblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 1.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Italic.ttf', name='Lato', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed-BoldOblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=700, stretch='condensed', size='scalable')) = 1.535
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed-Italic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=400, stretch='condensed', size='scalable')) = 11.25
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Thin.ttf', name='Lato', style='normal', variant='normal', weight=200, stretch='normal', size='scalable')) = 10.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Regular.ttf', name='Lato', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Light.ttf', name='Lato', style='normal', variant='normal', weight=300, stretch='normal', size='scalable')) = 10.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono-Bold.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif-Italic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuMathTeXGyre.ttf', name='DejaVu Math TeX Gyre', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-BlackItalic.ttf', name='Lato', style='italic', variant='normal', weight=900, stretch='normal', size='scalable')) = 11.525
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-MediumItalic.ttf', name='Lato', style='italic', variant='normal', weight=500, stretch='normal', size='scalable')) = 11.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Medium.ttf', name='Lato', style='normal', variant='normal', weight=500, stretch='normal', size='scalable')) = 10.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-LightItalic.ttf', name='Lato', style='italic', variant='normal', weight=300, stretch='normal', size='scalable')) = 11.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-Oblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 1.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono-BoldOblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-Bold.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 0.33499999999999996
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif-BoldItalic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-ExtraLight.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=200, stretch='normal', size='scalable')) = 0.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Semibold.ttf', name='Lato', style='normal', variant='normal', weight=600, stretch='normal', size='scalable')) = 10.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Hairline.ttf', name='Lato', style='normal', variant='normal', weight=100, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=400, stretch='condensed', size='scalable')) = 10.25
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=400, stretch='condensed', size='scalable')) = 0.25
findfont: Matching sans\-serif:style=normal:variant=normal:weight=normal:stretch=normal:size=12.0 to DejaVu Sans ('/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans.ttf') with score of 0.050000.
[31]:
# now i am trying mixed effects models
# need to do more in depth analysis of regular linreg and spline reg than just looking at R2s, but for now i moved onto LME
all_data = []
for subject_id, subject_data in vel_week.items():
for week, matrix in subject_data.items():
# each row in the matrix is a control point's velocity vector
for cp_index, velocity_vec in enumerate(matrix):
all_data.append({
"subject": subject_id,
"week": week,
"cp_index": cp_index,
"value_x": velocity_vec[0],
"value_y": velocity_vec[1],
"value_z": velocity_vec[2]
})
[32]:
import statsmodels.api as sm
import statsmodels.formula.api as smf
df = pd.DataFrame(all_data)
print("--- Reshaped Data (first 5 rows) ---")
print(df.head())
--- Reshaped Data (first 5 rows) ---
subject week cp_index value_x value_y value_z
0 01 -3.0 0 0.002202 0.051423 -0.045844
1 01 -3.0 1 0.002034 0.045535 -0.040385
2 01 -3.0 2 0.011171 0.042089 -0.014809
3 01 -3.0 3 0.011415 0.040953 -0.013083
4 01 -3.0 4 0.024612 0.052422 0.007442
[33]:
df
[33]:
| subject | week | cp_index | value_x | value_y | value_z | |
|---|---|---|---|---|---|---|
| 0 | 01 | -3.0 | 0 | 0.002202 | 0.051423 | -0.045844 |
| 1 | 01 | -3.0 | 1 | 0.002034 | 0.045535 | -0.040385 |
| 2 | 01 | -3.0 | 2 | 0.011171 | 0.042089 | -0.014809 |
| 3 | 01 | -3.0 | 3 | 0.011415 | 0.040953 | -0.013083 |
| 4 | 01 | -3.0 | 4 | 0.024612 | 0.052422 | 0.007442 |
| ... | ... | ... | ... | ... | ... | ... |
| 1483 | 1004B | 93.0 | 19 | -0.000731 | 0.000146 | 0.000213 |
| 1484 | 1004B | 93.0 | 20 | -0.003482 | 0.001033 | 0.000895 |
| 1485 | 1004B | 93.0 | 21 | -0.000172 | 0.000403 | -0.000060 |
| 1486 | 1004B | 93.0 | 22 | 0.000830 | 0.002317 | -0.000977 |
| 1487 | 1004B | 93.0 | 23 | 0.000788 | 0.001611 | -0.000752 |
1488 rows × 6 columns
[ ]:
import statsmodels.formula.api as smf
# i get convergence warnings here
# filter for one specific control point to model
df_cp0 = df[df['cp_index'] == 0]
# Model for X component
lme_model_x = smf.mixedlm("value_x ~ week", data=df_cp0, groups="subject")
results_x = lme_model_x.fit()
# Model for Y component
lme_model_y = smf.mixedlm("value_y ~ week", data=df_cp0, groups="subject")
results_y = lme_model_y.fit()
# Model for Z component
lme_model_z = smf.mixedlm("value_z ~ week", data=df_cp0, groups="subject")
results_z = lme_model_z.fit()
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
[35]:
X_pred = pd.DataFrame({'week': [25]})
pred_x = results_x.predict(X_pred)[0]
pred_y = results_y.predict(X_pred)[0]
pred_z = results_z.predict(X_pred)[0]
predicted_3d_vector = [pred_x, pred_y, pred_z]
print(f"predicted velocity at week 25: {predicted_3d_vector}")
predicted velocity at week 25: [np.float64(-0.1100075156282499), np.float64(0.06282665386116554), np.float64(0.07686207842449523)]
[36]:
print(results_x.summary())
Mixed Linear Model Regression Results
======================================================
Model: MixedLM Dependent Variable: value_x
No. Observations: 62 Method: REML
No. Groups: 3 Scale: 0.0227
Min. group size: 14 Log-Likelihood: 18.9563
Max. group size: 26 Converged: Yes
Mean group size: 20.7
------------------------------------------------------
Coef. Std.Err. z P>|z| [0.025 0.975]
------------------------------------------------------
Intercept -0.107 0.056 -1.910 0.056 -0.217 0.003
week -0.000 0.001 -0.168 0.867 -0.001 0.001
subject Var 0.007 0.056
======================================================
[37]:
print(results_y.summary())
# higher order mixed effects? again, need to do more analysis but just tried it for fun
lme_model_y_quad = smf.mixedlm("value_y ~ week + I(week**2)", data=df_cp0, groups="subject")
results_y_quad = lme_model_y_quad.fit()
print(results_y_quad.summary())
Mixed Linear Model Regression Results
=====================================================
Model: MixedLM Dependent Variable: value_y
No. Observations: 62 Method: REML
No. Groups: 3 Scale: 0.0202
Min. group size: 14 Log-Likelihood: 22.7625
Max. group size: 26 Converged: Yes
Mean group size: 20.7
-----------------------------------------------------
Coef. Std.Err. z P>|z| [0.025 0.975]
-----------------------------------------------------
Intercept 0.057 0.046 1.232 0.218 -0.034 0.148
week 0.000 0.001 0.381 0.703 -0.001 0.001
subject Var 0.004 0.039
=====================================================
Mixed Linear Model Regression Results
=======================================================
Model: MixedLM Dependent Variable: value_y
No. Observations: 62 Method: REML
No. Groups: 3 Scale: 0.0191
Min. group size: 14 Log-Likelihood: 14.2319
Max. group size: 26 Converged: Yes
Mean group size: 20.7
-------------------------------------------------------
Coef. Std.Err. z P>|z| [0.025 0.975]
-------------------------------------------------------
Intercept 0.024 0.051 0.473 0.637 -0.076 0.124
week 0.003 0.001 1.994 0.046 0.000 0.005
I(week ** 2) -0.000 0.000 -2.044 0.041 -0.000 -0.000
subject Var 0.005 0.045
=======================================================
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
[ ]:
results_summary = []
control_point_indices = df['cp_index'].unique()
# here i also get convergence warnings
# now i do this for each control point
# i need to figure out how to do multivariate
for cp_index in control_point_indices:
df_cp = df[df['cp_index'] == cp_index]
try:
p_x = smf.mixedlm("value_x ~ week", df_cp, groups="subject").fit().pvalues['week']
p_y = smf.mixedlm("value_y ~ week", df_cp, groups="subject").fit().pvalues['week']
p_z = smf.mixedlm("value_z ~ week", df_cp, groups="subject").fit().pvalues['week']
results_summary.append({
'cp_index': cp_index,
'p_value_x': p_x,
'p_value_y': p_y,
'p_value_z': p_z,
'min_p_value': min(p_x, p_y, p_z)
})
except Exception as e:
print(f"can't fit model for cp {cp_index}: {e}")
df_results = pd.DataFrame(results_summary)
best_result = df_results.loc[df_results['min_p_value'].idxmin()]
best_cp_index = int(best_result['cp_index'])
print("\n--- most significant ---")
print(best_result)
print(f"\nbest CP is #{best_cp_index}")
Analyzing 24 control points...
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/base/model.py:607: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals
warnings.warn("Maximum Likelihood optimization failed to "
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2200: ConvergenceWarning: Retrying MixedLM optimization with lbfgs
warnings.warn(
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2261: ConvergenceWarning: The Hessian matrix at the estimated parameter values is not positive definite.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/base/model.py:607: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals
warnings.warn("Maximum Likelihood optimization failed to "
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2200: ConvergenceWarning: Retrying MixedLM optimization with lbfgs
warnings.warn(
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/base/model.py:607: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals
warnings.warn("Maximum Likelihood optimization failed to "
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2200: ConvergenceWarning: Retrying MixedLM optimization with cg
warnings.warn(
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/base/model.py:607: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals
warnings.warn("Maximum Likelihood optimization failed to "
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2206: ConvergenceWarning: MixedLM optimization failed, trying a different optimizer may help.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2218: ConvergenceWarning: Gradient optimization failed, |grad| = 0.438618
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:1634: UserWarning: Random effects covariance is singular
warnings.warn(msg)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
--- most significant ---
cp_index 20.000000
p_value_x 0.000008
p_value_y 0.004790
p_value_z 0.230180
min_p_value 0.000008
Name: 20, dtype: float64
best CP is #20
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2261: ConvergenceWarning: The Hessian matrix at the estimated parameter values is not positive definite.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/base/model.py:607: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals
warnings.warn("Maximum Likelihood optimization failed to "
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2200: ConvergenceWarning: Retrying MixedLM optimization with lbfgs
warnings.warn(
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
[39]:
import seaborn as sns
# oh yeah, i installed this package
cp_to_plot = best_cp_index
df_plot = df[df['cp_index'] == cp_to_plot]
# determine which dimension had the minimum p-value
best_dim_info = best_result[['p_value_x', 'p_value_y', 'p_value_z']].idxmin()
main_dim_col = best_dim_info.replace('p_value_', 'value_') # e.g., 'value_y'
main_dim_label = main_dim_col[-1].upper() # e.g., 'Y'
model_x = smf.mixedlm(f"value_x ~ week", df_plot, groups="subject").fit()
model_y = smf.mixedlm(f"value_y ~ week", df_plot, groups="subject").fit()
model_z = smf.mixedlm(f"value_z ~ week", df_plot, groups="subject").fit()
models = {'X': model_x, 'Y': model_y, 'Z': model_z}
fig = plt.figure(figsize=(14, 7))
gs = fig.add_gridspec(2, 2, width_ratios=[2, 1], height_ratios=[1, 1])
ax_main = fig.add_subplot(gs[:, 0])
ax_sub1 = fig.add_subplot(gs[0, 1])
ax_sub2 = fig.add_subplot(gs[1, 1])
def plot_model_fit(ax, model, data, dim_label, is_main_plot=False):
value_col = f'value_{dim_label.lower()}'
p_value = model.pvalues['week']
sns.scatterplot(data=data, x='week', y=value_col, hue='subject', ax=ax, legend=is_main_plot, s=(50 if is_main_plot else 25))
x_range = pd.DataFrame({'week': np.linspace(data['week'].min(), data['week'].max(), 100)})
# plot average trend
y_avg = model.predict(x_range)
ax.plot(x_range['week'], y_avg, color='red', linestyle='--', linewidth=(3 if is_main_plot else 2), label='Average Trend')
# plot subject specific trends
for subject_id in data['subject'].unique():
x_subj = x_range.copy()
x_subj['subject'] = subject_id
y_subj = model.predict(x_subj)
subj_color = sns.color_palette(n_colors=data['subject'].nunique())[list(data['subject'].unique()).index(subject_id)]
ax.plot(x_subj['week'], y_subj, color=subj_color, linewidth=1, alpha=0.7)
title_str = f'trend in {dim_label} dimension'
if is_main_plot:
title_str = f'most significant trend in {dim_label} dimension'
ax.set_title(f'{title_str} (p = {p_value:.5f})', fontsize=(14 if is_main_plot else 10))
ax.set_ylabel(f'velocity component {dim_label}')
ax.set_xlabel('gestational week')
ax.grid(True, linestyle='--', alpha=0.6)
if is_main_plot:
handles, labels = ax.get_legend_handles_labels()
unique_labels = {}
for handle, label in zip(handles, labels):
if label not in unique_labels:
unique_labels[label] = handle
ax.legend(unique_labels.values(), unique_labels.keys(), title='Subject')
plot_model_fit(ax_main, models[main_dim_label], df_plot, main_dim_label, is_main_plot=True)
sub_dims = [dim for dim in ['X', 'Y', 'Z'] if dim != main_dim_label]
plot_model_fit(ax_sub1, models[sub_dims[0]], df_plot, sub_dims[0])
plot_model_fit(ax_sub2, models[sub_dims[1]], df_plot, sub_dims[1])
fig.suptitle(f'mixed effects model results for control point {cp_to_plot}', fontsize=18, y=1.02)
plt.tight_layout(rect=[0, 0, 1, 0.98])
plt.show()
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/statsmodels/regression/mixed_linear_model.py:2237: ConvergenceWarning: The MLE may be on the boundary of the parameter space.
warnings.warn(msg, ConvergenceWarning)
findfont: Matching sans\-serif:style=normal:variant=normal:weight=normal:stretch=normal:size=14.0.
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizTwoSymReg.ttf', name='STIXSizeTwoSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralBol.ttf', name='STIXGeneral', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-Oblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 1.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniBol.ttf', name='STIXNonUnicode', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizOneSymBol.ttf', name='STIXSizeOneSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansDisplay.ttf', name='DejaVu Sans Display', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralItalic.ttf', name='STIXGeneral', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizThreeSymBol.ttf', name='STIXSizeThreeSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-Bold.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmss10.ttf', name='cmss10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-BoldItalic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-Italic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 0.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-Bold.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 0.33499999999999996
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-Bold.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmsy10.ttf', name='cmsy10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmex10.ttf', name='cmex10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmb10.ttf', name='cmb10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerifDisplay.ttf', name='DejaVu Serif Display', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-Oblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmr10.ttf', name='cmr10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFiveSymReg.ttf', name='STIXSizeFiveSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmmi10.ttf', name='cmmi10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-BoldOblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 1.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralBolIta.ttf', name='STIXGeneral', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniIta.ttf', name='STIXNonUnicode', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-BoldOblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFourSymReg.ttf', name='STIXSizeFourSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUni.ttf', name='STIXNonUnicode', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniBolIta.ttf', name='STIXNonUnicode', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizThreeSymReg.ttf', name='STIXSizeThreeSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizTwoSymBol.ttf', name='STIXSizeTwoSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFourSymBol.ttf', name='STIXSizeFourSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizOneSymReg.ttf', name='STIXSizeOneSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneral.ttf', name='STIXGeneral', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmtt10.ttf', name='cmtt10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed-Oblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=400, stretch='condensed', size='scalable')) = 1.25
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-HeavyItalic.ttf', name='Lato', style='italic', variant='normal', weight=800, stretch='normal', size='scalable')) = 11.43
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed-Bold.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=700, stretch='condensed', size='scalable')) = 0.5349999999999999
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-HairlineItalic.ttf', name='Lato', style='italic', variant='normal', weight=100, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono-Oblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-ThinItalic.ttf', name='Lato', style='italic', variant='normal', weight=200, stretch='normal', size='scalable')) = 11.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Heavy.ttf', name='Lato', style='normal', variant='normal', weight=800, stretch='normal', size='scalable')) = 10.43
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif-Bold.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Bold.ttf', name='Lato', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-BoldItalic.ttf', name='Lato', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed-BoldItalic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=700, stretch='condensed', size='scalable')) = 11.535
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-SemiboldItalic.ttf', name='Lato', style='italic', variant='normal', weight=600, stretch='normal', size='scalable')) = 11.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Black.ttf', name='Lato', style='normal', variant='normal', weight=900, stretch='normal', size='scalable')) = 10.525
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed-Bold.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=700, stretch='condensed', size='scalable')) = 10.535
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 0.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-BoldOblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 1.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Italic.ttf', name='Lato', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed-BoldOblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=700, stretch='condensed', size='scalable')) = 1.535
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed-Italic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=400, stretch='condensed', size='scalable')) = 11.25
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Thin.ttf', name='Lato', style='normal', variant='normal', weight=200, stretch='normal', size='scalable')) = 10.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Regular.ttf', name='Lato', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Light.ttf', name='Lato', style='normal', variant='normal', weight=300, stretch='normal', size='scalable')) = 10.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono-Bold.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif-Italic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuMathTeXGyre.ttf', name='DejaVu Math TeX Gyre', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-BlackItalic.ttf', name='Lato', style='italic', variant='normal', weight=900, stretch='normal', size='scalable')) = 11.525
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-MediumItalic.ttf', name='Lato', style='italic', variant='normal', weight=500, stretch='normal', size='scalable')) = 11.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Medium.ttf', name='Lato', style='normal', variant='normal', weight=500, stretch='normal', size='scalable')) = 10.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-LightItalic.ttf', name='Lato', style='italic', variant='normal', weight=300, stretch='normal', size='scalable')) = 11.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-Oblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 1.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono-BoldOblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-Bold.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 0.33499999999999996
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif-BoldItalic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-ExtraLight.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=200, stretch='normal', size='scalable')) = 0.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Semibold.ttf', name='Lato', style='normal', variant='normal', weight=600, stretch='normal', size='scalable')) = 10.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Hairline.ttf', name='Lato', style='normal', variant='normal', weight=100, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=400, stretch='condensed', size='scalable')) = 10.25
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=400, stretch='condensed', size='scalable')) = 0.25
findfont: Matching sans\-serif:style=normal:variant=normal:weight=normal:stretch=normal:size=14.0 to DejaVu Sans ('/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans.ttf') with score of 0.050000.
findfont: Matching sans\-serif:style=normal:variant=normal:weight=normal:stretch=normal:size=18.0.
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizTwoSymReg.ttf', name='STIXSizeTwoSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralBol.ttf', name='STIXGeneral', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-Oblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 1.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniBol.ttf', name='STIXNonUnicode', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizOneSymBol.ttf', name='STIXSizeOneSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansDisplay.ttf', name='DejaVu Sans Display', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralItalic.ttf', name='STIXGeneral', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizThreeSymBol.ttf', name='STIXSizeThreeSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-Bold.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmss10.ttf', name='cmss10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-BoldItalic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-Italic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 0.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-Bold.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 0.33499999999999996
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-Bold.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmsy10.ttf', name='cmsy10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmex10.ttf', name='cmex10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmb10.ttf', name='cmb10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerifDisplay.ttf', name='DejaVu Serif Display', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-Oblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmr10.ttf', name='cmr10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFiveSymReg.ttf', name='STIXSizeFiveSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmmi10.ttf', name='cmmi10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-BoldOblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 1.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralBolIta.ttf', name='STIXGeneral', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniIta.ttf', name='STIXNonUnicode', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-BoldOblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFourSymReg.ttf', name='STIXSizeFourSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUni.ttf', name='STIXNonUnicode', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniBolIta.ttf', name='STIXNonUnicode', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizThreeSymReg.ttf', name='STIXSizeThreeSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizTwoSymBol.ttf', name='STIXSizeTwoSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFourSymBol.ttf', name='STIXSizeFourSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXSizOneSymReg.ttf', name='STIXSizeOneSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneral.ttf', name='STIXGeneral', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/cmtt10.ttf', name='cmtt10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed-Oblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=400, stretch='condensed', size='scalable')) = 1.25
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-HeavyItalic.ttf', name='Lato', style='italic', variant='normal', weight=800, stretch='normal', size='scalable')) = 11.43
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed-Bold.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=700, stretch='condensed', size='scalable')) = 0.5349999999999999
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-HairlineItalic.ttf', name='Lato', style='italic', variant='normal', weight=100, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono-Oblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-ThinItalic.ttf', name='Lato', style='italic', variant='normal', weight=200, stretch='normal', size='scalable')) = 11.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Heavy.ttf', name='Lato', style='normal', variant='normal', weight=800, stretch='normal', size='scalable')) = 10.43
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif-Bold.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Bold.ttf', name='Lato', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-BoldItalic.ttf', name='Lato', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed-BoldItalic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=700, stretch='condensed', size='scalable')) = 11.535
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-SemiboldItalic.ttf', name='Lato', style='italic', variant='normal', weight=600, stretch='normal', size='scalable')) = 11.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Black.ttf', name='Lato', style='normal', variant='normal', weight=900, stretch='normal', size='scalable')) = 10.525
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed-Bold.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=700, stretch='condensed', size='scalable')) = 10.535
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 0.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-BoldOblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 1.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Italic.ttf', name='Lato', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed-BoldOblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=700, stretch='condensed', size='scalable')) = 1.535
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed-Italic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=400, stretch='condensed', size='scalable')) = 11.25
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Thin.ttf', name='Lato', style='normal', variant='normal', weight=200, stretch='normal', size='scalable')) = 10.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Regular.ttf', name='Lato', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Light.ttf', name='Lato', style='normal', variant='normal', weight=300, stretch='normal', size='scalable')) = 10.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono-Bold.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif-Italic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuMathTeXGyre.ttf', name='DejaVu Math TeX Gyre', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-BlackItalic.ttf', name='Lato', style='italic', variant='normal', weight=900, stretch='normal', size='scalable')) = 11.525
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-MediumItalic.ttf', name='Lato', style='italic', variant='normal', weight=500, stretch='normal', size='scalable')) = 11.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Medium.ttf', name='Lato', style='normal', variant='normal', weight=500, stretch='normal', size='scalable')) = 10.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-LightItalic.ttf', name='Lato', style='italic', variant='normal', weight=300, stretch='normal', size='scalable')) = 11.145
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-Oblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 1.05
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansMono-BoldOblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-Bold.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 0.33499999999999996
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerif-BoldItalic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSans-ExtraLight.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=200, stretch='normal', size='scalable')) = 0.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Semibold.ttf', name='Lato', style='normal', variant='normal', weight=600, stretch='normal', size='scalable')) = 10.24
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/lato/Lato-Hairline.ttf', name='Lato', style='normal', variant='normal', weight=100, stretch='normal', size='scalable')) = 10.335
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSerifCondensed.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=400, stretch='condensed', size='scalable')) = 10.25
findfont: score(FontEntry(fname='/usr/share/fonts/truetype/dejavu/DejaVuSansCondensed.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=400, stretch='condensed', size='scalable')) = 0.25
findfont: Matching sans\-serif:style=normal:variant=normal:weight=normal:stretch=normal:size=18.0 to DejaVu Sans ('/home/sak/anaconda3/envs/deformetrica/lib/python3.12/site-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans.ttf') with score of 0.050000.
[40]:
df_plot = df[df['cp_index'] == 20]
g = sns.relplot(
data=df_plot,
x="week",
y="value_x",
hue="subject",
col="subject",
kind="scatter",
height=4,
aspect=1,
)
g.map(sns.regplot, "week", "value_x", scatter=False, ci=None, line_kws={'color':'black', 'linestyle':'--'})
g.fig.suptitle("Subject trends for control point 20 (x dimension)", y=1.03)
plt.show()
[41]:
# was googling about how to evaluate mixed effects models and apparently regular r2 doesn't make sense so chatgpt made this function for me
def calculate_r2_mixed(model):
"""
Calculates Marginal and Conditional R-squared for a statsmodels MixedLM object.
This corrected version robustly calculates the fixed-effects variance.
Args:
model: A fitted MixedLM results object.
Returns:
A dictionary containing the marginal and conditional R-squared values.
"""
# Variance of the fixed effects component
# This is the variance of the predictions based ONLY on the fixed effects
# It's the dot product of the design matrix (X) and the fixed-effect coefficients (beta)
var_fixed = np.var(np.dot(model.model.exog, model.fe_params))
# Variance of the random effects component (same as before)
var_random = float(model.cov_re.iloc[0, 0])
# Variance of the residuals (same as before)
var_residual = model.scale
# Total variance is the sum of all components
var_total = var_fixed + var_random + var_residual
# Marginal R-squared (variance explained by fixed effects)
r2_marginal = var_fixed / var_total
# Conditional R-squared (variance explained by fixed and random effects)
r2_conditional = (var_fixed + var_random) / var_total
return {
"marginal_r2": r2_marginal,
"conditional_r2": r2_conditional
}
[42]:
r2_values = calculate_r2_mixed(results_x)
print(f"R-squared for the model:")
print(f" - Marginal R² (Fixed effects only): {r2_values['marginal_r2']:.5f}")
print(f" - Conditional R² (Fixed + Random effects): {r2_values['conditional_r2']:.5f}")
R-squared for the model:
- Marginal R² (Fixed effects only): 0.00034
- Conditional R² (Fixed + Random effects): 0.24078