Use your example OBJ file can run the code but use myself have the below error

[Benson1223]

(try_mvlm) D:\python_work\mvlm-main>python inference.py
WARNING: All log messages before absl::InitializeLog() is called are written to STDERR
W0000 00:00:1734603455.067464 26808 face_landmarker_graph.cc:174] Sets FaceBlendshapesGraph acceleration to xnnpack by default.
INFO: Created TensorFlow Lite XNNPACK delegate for CPU.
Render [0] - Prepare 0.000000 s
Render [1] - Setup time: 1.582143 s
Render [2] - Render 1.678741 s
Render [Total]: 3.311789 s
Prediction [Total]: 0.020932 s
Landmarks [0] - From Heatmaps: 0.000000 s
Traceback (most recent call last):
File "D:\python_work\mvlm-main\inference.py", line 16, in
landmarks = dm.predict_one_file(file)
File "D:\python_work\mvlm-main\src\mvlm\pipeline\general_pipeline.py", line 82, in predict_one_file
landmarks, error = self.estimator_3d.estimate_landmarks_from_lines(landmark_stack, lines_s, lines_e)
File "D:\python_work\mvlm-main\src\mvlm\utils\estimator3d.py", line 169, in estimate_landmarks_from_lines
pa, pb = self.filter_lines_based_on_heatmap_value_using_quantiles(landmark_stack, lm_no, pa, pb)
File "D:\python_work\mvlm-main\src\mvlm\utils\estimator3d.py", line 142, in filter_lines_based_on_heatmap_value_using_quantiles
threshold = np.quantile(max_values, self.threshold_quantile)
File "C:\Users\user\anaconda3\envs\try_mvlm\lib\site-packages\numpy\lib_function_base_impl.py", line 4663, in quantile
return _quantile_unchecked(
File "C:\Users\user\anaconda3\envs\try_mvlm\lib\site-packages\numpy\lib_function_base_impl.py", line 4676, in _quantile_unchecked
return _ureduce(a,
File "C:\Users\user\anaconda3\envs\try_mvlm\lib\site-packages\numpy\lib_function_base_impl.py", line 3764, in _ureduce
r = func(a, **kwargs)
File "C:\Users\user\anaconda3\envs\try_mvlm\lib\site-packages\numpy\lib_function_base_impl.py", line 4853, in _quantile_ureduce_func
result = _quantile(arr,
File "C:\Users\user\anaconda3\envs\try_mvlm\lib\site-packages\numpy\lib_function_base_impl.py", line 4975, in _quantile
slices_having_nans = np.isnan(arr[-1, ...])
IndexError: index -1 is out of bounds for axis 0 with size 0

[Benson1223]

My code as below:
from pathlib import Path
import numpy as np
import mvlm
file = Path("./lulu_alignment/texturedMesh.obj")
output_dir = Path("./lulu_alignment/res_point")

dm = mvlm.pipeline.create_pipeline("mediapipe")
landmarks = dm.predict_one_file(file)
np.savetxt((output_dir/f"{file.stem}.txt").as_posix(), landmarks, delimiter=",")

mvlm.utils.VTKViewer(file.as_posix(), landmarks)

[Timozen]

Thank you for pointing out this issue. I will take a look into the possible cause.
The underlying error indicates that the landmark prediction somehow failed.

[Benson1223]

Many thanks for your help! I try to use testmeshA.obj (the example file of Deep-MVLM) to run mediapipe, but can't run. However I use your sample file is available.