# export MODEL=Meta-Llama-3-8B-Instruct-hf/
export MODEL=Llama-2-7b-chat-hf/
mkdir -p minigpt4/ft/$MODEL
# /home/ubuntu/miniconda3/envs/prot/bin/torchrun --master_port 28888 protein_gpt.py --cfg-path configs/train_modality_alignment.yaml
/home/ubuntu/miniconda3/envs/prot/bin/torchrun --master_port 28888 protein_gpt.py --cfg-path configs/train_instruction_tuning.yaml