Accelerate

Run PyTorch training across GPUs with minimal changes.

By Orchestra Research · Official Docs · Date unknown

Accelerate simplifies distributed training to 4 lines of code.

Platforms: linux, macos, windows.

hermes skills install official/mlops/accelerate

• #skills
• #optional
• #mlops
• #distributed training
• #huggingface
• #accelerate
• #deepspeed