Spaces:

Justinrune
/

LLaMA-Factory

Running

Upload folder using huggingface_hub

2852136 verified 7 months ago

222 Bytes

	#!/bin/bash
	# DO NOT use GPTQ/AWQ model in FSDP+QLoRA

	CUDA_VISIBLE_DEVICES=0,1 accelerate launch \
	--config_file examples/accelerate/fsdp_config.yaml \
	src/train.py examples/extras/fsdp_qlora/llama3_lora_sft.yaml