We want to create a dedicated Pytorch Training container that includes all important Hugging Face Ecosystem libraries, including transformers, datasets, evaluate, diffusers, trl, peft, sentence-transformers and common ML libraries, like torch, scikit-learn, tensorboard, bitsandbytes, deepspeed etc. to allow Google users to train models easily.
We want to create a dedicated Pytorch Training container that includes all important Hugging Face Ecosystem libraries, including transformers, datasets, evaluate, diffusers, trl, peft, sentence-transformers and common ML libraries, like torch, scikit-learn, tensorboard, bitsandbytes, deepspeed etc. to allow Google users to train models easily.
Pytorch training container with all required Hugging Face Libraries (versions should be pinned) For inspiration see philschmid container, or pytorch DLC, transformers deepspeed, transformers container gpu