Added doc for nvdec #335
Added doc for nvdec #335
Conversation
facebook-github-bot
added
the
CLA Signed
There was a problem hiding this comment.
Great tutorial! It's great to see that the UX for installing gpu-compatible wheels is fairly straightforward. I left a few comments below.
Doc job is failing with a minor build issue: https://github.com/pytorch/torchcodec/actions/runs/11734305974/job/32690104327?pr=335
As a result I haven't been able to look at the rendered docs, please let me do another round before landing :)
There was a problem hiding this comment.
Great tutorial! It's great to see that the UX for installing gpu-compatible wheels is fairly straightforward. I left a few comments below.
Doc job is failing with a minor build issue: https://github.com/pytorch/torchcodec/actions/runs/11734305974/job/32690104327?pr=335
As a result I haven't been able to look at the rendered docs, please let me do another round before landing :)
| run: | | ||
| cd docs | ||
| python -m pip install -r requirements.txt | ||
| ${CONDA_RUN} python -m pip install -r requirements.txt |
There was a problem hiding this comment.
Do we understand why we need CONDA_RUN here?
There was a problem hiding this comment.
Everything is installed in a conda env. Without CONDA_RUN the pip that's used is the outside pip
| first_cpu_frame = cpu_frames[0].data.to("cpu") | ||
| first_cuda_frame = cuda_frames[0].data.to("cpu") | ||
| frames_equal = torch.equal(first_cpu_frame, first_cuda_frame) | ||
| print(f"{frames_equal=}") |
There was a problem hiding this comment.
Nit: Instead of this binary indicator, we may want to print max abs diff and mean abs diff? We could even do it across all frames.
4 participants
Result screenshot: