add "torch_dtype": "bfloat16"

#3
by wukaixingxp - opened
Meta Llama org
No description provided.
qubvel-hf changed pull request status to closed

Sign up or log in to comment