search

david-svitov / HAHA

HAHA: Highly Articulated Gaussian Human Avatars with Textured Mesh Prior
89 stars 9 forks source link

RuntimeError: Tensors of type TensorImpl do not have sizes #3

Closed zyhbili closed 2 months ago

zyhbili commented 3 months ago

Thanks for your nice work! I install the gaussian rasterization in the ./3rd-party and encounter the following bugs. Could you give me some instructions?

  File "/data5/zhiyh/HAHA/utils/gaussian_rasterizer.py", line 73, in forward
    ret = rasterizer(
  File "/data4/zhiyh/miniconda3/envs/ngt/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
    return forward_call(*input, **kwargs)
  File "/data4/zhiyh/miniconda3/envs/ngt/lib/python3.8/site-packages/diff_gaussian_rasterization/__init__.py", line 218, in forward
    return rasterize_gaussians(
  File "/data4/zhiyh/miniconda3/envs/ngt/lib/python3.8/site-packages/diff_gaussian_rasterization/__init__.py", line 33, in rasterize_gaussians
    return _RasterizeGaussians.apply(
  File "/data4/zhiyh/miniconda3/envs/ngt/lib/python3.8/site-packages/diff_gaussian_rasterization/__init__.py", line 96, in forward
    num_rendered, color, depth, alpha, radii, geomBuffer, binningBuffer, imgBuffer = _C.rasterize_gaussians(*args)
RuntimeError: Tensors of type TensorImpl do not have sizes
david-svitov commented 3 months ago

Hello!

Let's try to understand what caused the error. Have you installed diff_gaussian_rasterization-alphadep as stated in the readme? https://github.com/david-svitov/HAHA/blob/42bd784670eb6a9f1c74c1fef5c8a2532174f8b2/README.md?plain=1#L39

Could you send me the result of this command?

zyhbili commented 3 months ago

Yes, i have installed diff_gaussian_rasterization-alphadep. Since i cannot use docker on my server, i build my env using conda only, and i install nvdiffrast==0.3.1 instead of yours for some bugs. The rest packages, including torch cuda etc., are consistent with yours in the Dockerfile.

zyhbili commented 2 months ago

I set up a new env, and solve it, but i do not figure out the reason. Thanks a lot!