unable to install pytorch3d

[danvisai]

pip install "git+https://github.com/facebookresearch/pytorch3d.git@stable"

note: This error originates from a subprocess, and is likely not a problem with pip.
ERROR: Failed building wheel for pytorch3d
Running setup.py clean for pytorch3d
Failed to build pytorch3d

i cannot tell if its due to nvcc version mismatch
current nvcc version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2023 NVIDIA Corporation
Built on Fri_Jan__6_16:45:21_PST_2023
Cuda compilation tools, release 12.0, V12.0.140
Build cuda_12.0.r12.0/compiler.32267302_0

[NIRVANALAN]

Hi, thanks for your comments.

This issue is pretty wired and is likely due to the mismatch between your pytorch compiled cuda version and your machine's nvcc/cuda version. Any other error message to share?

[danvisai]

Hi, thanks for your comments.

This issue is pretty wired and is likely due to the mismatch between your pytorch compiled cuda version and your machine's nvcc/cuda version. Any other error message to share?

Hi,
I managed to get the versions right and the gradio app runs now. However, the results are not good.

this is what i get in console.-------------
stage-1 glb point cloud saved to ./logs/i23d/stage-1/dino_img//house2-input/sample-0-0.glb
point cloud saved to ./logs/i23d/stage-1/dino_img//house2-input/sample-0-0.glb
loading stage-1 point cloud from: ./logs/i23d/stage-1/dino_img//house2-input/sample-0-0.ply
cfg_scale: 4, seed: 42
size: 1
sfdghggjfghfghfghfg torch.Size([1, 768, 10])
0%| | 0/50 [00:00<?, ?it/s]Numba: Attempted to fork from a non-main thread, the TBB library may be in an invalid state in the child process.
Numba: Attempted to fork from a non-main thread, the TBB library may be in an invalid state in the child process.
100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 50/50 [00:00<00:00, 70.13it/s]
logged video to: ./logs/i23d/stage-1/dino_img//house2-input/sample-0-0.mp4
exporting mesh ...
using aabb
Running tsdf volume integration ...
voxel_size: 0.005358532185916215
sdf_trunc: 0.06430238623099457
depth_truc: 3
TSDF integration progress: 50it [00:00, 73.94it/s]
mesh saved at ./logs/i23d/stage-1/dino_img/house2-input/0-mesh_raw.obj
[Open3D DEBUG] [ClusterConnectedTriangles] Compute triangle adjacency
[Open3D DEBUG] [ClusterConnectedTriangles] Done computing triangle adjacency
[Open3D DEBUG] [ClusterConnectedTriangles] Done clustering, #clusters=32544
mesh post processed saved at ./logs/i23d/stage-1/dino_img/house2-input/0-mesh.obj
Numba: Attempted to fork from a non-main thread, the TBB library may be in an invalid state in the child process.

[NIRVANALAN]

That's weird, seems some libraries are not installed correctly. I have seen this output pattern before when I did not have Apex installed correctly (actually on the huggingface demo, I use the pytorch RMSNorm rather than the apex version to bypass this issue). Did you use apex in your local gradio demo?