We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Huggingface-based Llama training with deepspeed
bash accelerate.sh