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

7.32k stars 872 forks source link

fix mlu device longTensor bugs #2887

Open huismiling opened 1 week ago

huismiling commented 1 week ago

What does this PR do?

Fixes # (fix mlu LongTensor.to(int num) bugs)

Before submitting

[x] Did you read the contributor guideline, Pull Request section?

HuggingFaceDocBuilderDev commented 15 minutes 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.

huismiling commented 4 minutes ago

Thanks! Can you point to the PR where they fixed this please? Or the version? Just to make sure it's fixed

@muellerzr Hi, MLU fixed this bug in our private torch wheel . It's not a bug of accelerate.