huangyangyi / TeCH

[3DV 2024] Official repo of "TeCH: Text-guided Reconstruction of Lifelike Clothed Humans"

Home Page:https://huangyangyi.github.io/TeCH/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Dependency issues

michaltemp opened this issue · comments

Hi, your paper looks great. However, I am having some troubles with setting up environment according to provided dependencies.

Firstly, pyrender 0.1.45 depends on PyOpenGL==3.1.0, whereas requirements.txt asks for PyOpenGL==3.1.5.
Further, pytorch-lightning 1.9.1 depends on torchmetrics>=0.7.0 (not torchmetrics==0.6.0).
Then I get error from sklearn: "The 'sklearn' PyPI package is deprecated, use 'scikit-learn', ..." (installed as a dependency to one of the other packages)
Also, I get: "Could not find a version that satisfies the requirement taming-transformers (unavailable) (from versions: 0.0.1).

All tried in docker: pytorch/pytorch:1.13.0-cuda11.6-cudnn8-devel

Do you happen to have a Dockerfile that would work? Otherwise, resolving those dependency issues seems like a very tedious task.

The original package dependency is quite complex, you could check install.md, which is an updated installation introduction that works well on my local workstation. I have removed the version numbers as much as possible and added a few necessary packages that are missing in the original requirements.txt.

The original package dependency is quite complex, you could check install.md, which is an updated installation introduction that works well on my local workstation. I have removed the version numbers as much as possible and added a few necessary packages that are missing in the original requirements.txt.

Hello, could you share your driver version? I've encountered some issues with my A100, and it seems to be related to pyrender and EGL. Through my research, it appears that changing to an appropriate driver version might solve this problem.

  File "/media/shared1/linjie/TeCH-XIU/cores/main.py", line 84, in <module>
    model = Renderer(cfg)
  File "/media/shared1/linjie/TeCH-XIU/cores/lib/renderer.py", line 143, in __init__
    self.dmtet_network.init_mesh(
  File "/media/shared1/linjie/TeCH-XIU/cores/lib/dmtet_network.py", line 201, in init_mesh
    mesh_to_sdf.mesh_to_sdf(mesh,
  File "/media/shared1/linjie/.conda/envs/TeCH/lib/python3.10/site-packages/mesh_to_sdf/__init__.py", line 34, in mesh_to_sdf
    point_cloud = get_surface_point_cloud(mesh, surface_point_method, bounding_radius, scan_count, scan_resolution, sample_point_count, calculate_normals=sign_method=='normal')
  File "/media/shared1/linjie/.conda/envs/TeCH/lib/python3.10/site-packages/mesh_to_sdf/__init__.py", line 17, in get_surface_point_cloud
    return surface_point_cloud.create_from_scans(mesh, bounding_radius=bounding_radius, scan_count=scan_count, scan_resolution=scan_resolution, calculate_normals=calculate_normals)
  File "/media/shared1/linjie/.conda/envs/TeCH/lib/python3.10/site-packages/mesh_to_sdf/surface_point_cloud.py", line 162, in create_from_scans
    scans.append(Scan(mesh,
  File "/media/shared1/linjie/.conda/envs/TeCH/lib/python3.10/site-packages/mesh_to_sdf/scan.py", line 58, in __init__
    color, depth = render_normal_and_depth_buffers(mesh, camera, self.camera_transform, resolution)
  File "/media/shared1/linjie/.conda/envs/TeCH/lib/python3.10/site-packages/mesh_to_sdf/pyrender_wrapper.py", line 58, in render_normal_and_depth_buffers
    renderer = pyrender.OffscreenRenderer(resolution, resolution)
  File "/media/shared1/linjie/.conda/envs/TeCH/lib/python3.10/site-packages/pyrender/offscreen.py", line 31, in __init__
    self._create()
  File "/media/shared1/linjie/.conda/envs/TeCH/lib/python3.10/site-packages/pyrender/offscreen.py", line 137, in _create
    egl_device = egl.get_device_by_index(device_id)
  File "/media/shared1/linjie/.conda/envs/TeCH/lib/python3.10/site-packages/pyrender/platforms/egl.py", line 83, in get_device_by_index
    raise ValueError('Invalid device ID ({})'.format(device_id, len(devices)))
ValueError: Invalid device ID (0)