script/ft.sh

export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7

model_name_or_path="meta-llama/Llama-2-7b-chat-hf"
super_tokenizer_name_or_path="/home/baaiks/ninglu/code/PluginTransformer/data/outputs/90k_0104+12/super_tokenizer"
output_dir="data/outputs/90k_0104+12-longalpaca_0106"

mkdir -p ${output_dir}

deepspeed --master_port 12345 --module main.train \
    --model_name_or_path ${model_name_or_path} \
    --super_tokenizer_name_or_path ${super_tokenizer_name_or_path} \
    --super_tokenizer_num_hidden_layers 12 \
    --dataset_list "longalpaca_0106" \
    --output_dir ${output_dir} \
    --learning_rate 1e-5 \
    --num_train_epochs 1 \
    --per_device_train_batch_size 1 \
    --max_step 850 \
    --logging_strategy "steps" \
    --logging_steps 50 \
    --save_strategy "steps" \
    --save_steps 850 \
    --gradient_checkpointing \
    --deepspeed "data/ds_config/ds_config_stage1.json" \
    | tee "${output_dir}/train.log"