MiniCPM/finetune/lora_finetune_ocnli.sh
2024-03-16 01:58:08 +08:00

18 lines
811 B
Bash

formatted_time=$(date +"%Y%m%d%H%M%S")
echo $formatted_time
deepspeed --include localhost:1 --master_port 19888 finetune.py \
--model_name_or_path MiniCPM-2B-sft-bf16 \
--output_dir output/OCNLILoRA/$formatted_time/ \
--train_data_path data/ocnli_public_chatml/train.json \
--eval_data_path data/ocnli_public_chatml/dev.json \
--learning_rate 5e-5 --per_device_train_batch_size 80 \
--per_device_eval_batch_size 128 --model_max_length 128 --bf16 --use_lora \
--gradient_accumulation_steps 1 --warmup_steps 100 \
--max_steps 1000 --weight_decay 0.01 \
--evaluation_strategy steps --eval_steps 500 \
--save_strategy steps --save_steps 500 --seed 42 \
--log_level info --logging_strategy steps --logging_steps 10 \
--deepspeed configs/ds_config_zero3_offload.json