Downloaded Pytorch, conversion doesn't run anymore (returns immediately)

[aleph65]

I downloaded Pytorch to run conversion on the GPU

conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia

I am running on a conda env for 3.9 or 3.10 Python.

(nougat-env) C:\Users\user6\OneDrive\Desktop>nougat ./MyPDF.pdf -o ./TestOutput C:\Users\user6\miniconda3\envs\nougat-env\lib\site-packages\torch\functional.py:504: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at C:\cb\pytorch_1000000000000\work\aten\src\ATen\native\TensorShape.cpp:3527.) return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined]

The following just returns immediately; I however don't see the "No GPU found. Conversion on CPU is very slow." error anymore. I've restarted computer, recreated conda envs several times with slightly modified versionings.

Any suggestions please?

[noobmaster29]

I'm also getting the following error: /usr/local/lib/python3.10/dist-packages/torch/functional.py:504: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at ../aten/src/ATen/native/TensorShape.cpp:3526.) return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined]

It seems to be running though but I'm not sure if the result will be valid.

[aleph65]

@noobmaster29 thanks, did it output anything for you?

[noobmaster29]

Yep took a long time. It did run and output for me.

[aleph65]

@noobmaster29 thanks so much, i will try again.. if you find any way to display the progress let me know.. thanks :)