RuntimeError: aten::grid_sampler_2d_backward() - Githubissues

PDillis / stylegan3-fun

Modifications of the official PyTorch implementation of StyleGAN3. Let's easily generate images and videos with StyleGAN2/2-ADA/3!

Other

230 stars 36 forks source link

RuntimeError: aten::grid_sampler_2d_backward() #11

Closed secretsather closed 2 years ago

secretsather commented 2 years ago

Running train.py results in the following runtime error: RuntimeError: aten::grid_sampler_2d_backward() is missing value for argument 'output_mask'. Declaration: aten::grid_sampler_2d_backward(Tensor grad_output, Tensor input, Tensor grid, int interpolation_mode, int padding_mode, bool align_corners, bool[2] output_mask) -> (Tensor, Tensor)

OS: Windows 11
PyTorch version 1.11
CUDA toolkit version 11.6
GPU GTX 3080ti
Docker: did you use Docker? No

Changes from the below commit to conv2d_gradfix.py & grid_sample_gradfix.py will correct issues: https://github.com/NVlabs/stylegan3/commit/407db86e6fe432540a22515310188288687858fa

PDillis commented 2 years ago

Thanks, I have updated the code to work with PyTorch 1.11 in PR #12. Check it in case there's anything wrong.