huggingface / accelerate

🚀 A simple way to launch, train, and use PyTorch models on almost any device and distributed configuration, automatic mixed precision (including fp8), and easy-to-configure FSDP and DeepSpeed support
https://huggingface.co/docs/accelerate
Apache License 2.0
7.97k stars 970 forks source link

eliminate dead code #3198

Closed statelesshz closed 3 weeks ago

statelesshz commented 4 weeks ago

What does this PR do?

I noticed that the minimum version of safetensors that accelerate depends on is 0.4.3 https://github.com/huggingface/accelerate/blob/ba7ab93f5e688466ea56908ea3b056fae2f9a023/setup.py#L80

Before submitting

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR.

cc @muellerzr

HuggingFaceDocBuilderDev commented 4 weeks ago

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

BenjaminBossan commented 4 weeks ago

@statelesshz The linter is complaining because the compare_versions is now obsolete. Could you please remove it?

statelesshz commented 3 weeks ago

I think this PR is ready to be merged :) @muellerzr