pytorch / pytorch/vision

Prebuilt .whl and .conda of torchvision for aarch64 + cuda

Open
#8,614 1 comment 1 reaction 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
17.9k
Forks
7.3k
Avg merge
1d 15h
Merged PRs (30d)
13

Description

🚀 The feature

torchvision provides prebuilt binaries for aarch64 + cuda on linux for pip and conda

Motivation, pitch

pytorch now provides aarch64 cuda prebuilt binaries in 2.4.0 (https://github.com/pytorch/pytorch/issues/134117) but currently torchvision doesn't provide corresponding prebuilt binaries. It makes very hard to adopt torch on aarch64 + cuda environment, which is often much cheaper in cloud providers.

Alternatives

Build torchvision from source, which is not ideal for users of this library since it requires lots of work.

Additional context

Some relevant discussion here: https://discuss.pytorch.org/t/cuda-pytorch-package-manager-on-gh200/208582

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No files, tests, or entry points are named. Review the linked PyTorch issue 134117 and the referenced PyTorch forum discussion first, then determine the build and release work needed for aarch64 CUDA wheels and conda packages; done means torchvision binaries are published for the requested platform and package managers.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
build-system, computer-vision, machine-learning, release
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
38/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.