File size: 1,093 Bytes
302920f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 |
accelerate launch --config_file "configs/deepspeed_config.yaml" train.py \
--seed 100 \
--model_name_or_path "meta-llama/Llama-2-70b-hf" \
--dataset_name "smangrul/ultrachat-10k-chatml" \
--chat_template_format "chatml" \
--add_special_tokens False \
--append_concat_token False \
--splits "train,test" \
--max_seq_len 2048 \
--num_train_epochs 1 \
--logging_steps 5 \
--log_level "info" \
--logging_strategy "steps" \
--eval_strategy "epoch" \
--save_strategy "epoch" \
--push_to_hub \
--hub_private_repo True \
--hub_strategy "every_save" \
--bf16 True \
--packing True \
--learning_rate 1e-4 \
--lr_scheduler_type "cosine" \
--weight_decay 1e-4 \
--warmup_ratio 0.0 \
--max_grad_norm 1.0 \
--output_dir "mistral-sft-lora-deepspeed" \
--per_device_train_batch_size 8 \
--per_device_eval_batch_size 8 \
--gradient_accumulation_steps 4 \
--gradient_checkpointing True \
--use_reentrant False \
--dataset_text_field "content" \
--use_flash_attn True \
--use_peft_lora True \
--lora_r 8 \
--lora_alpha 16 \
--lora_dropout 0.1 \
--lora_target_modules "all-linear" \
--use_4bit_quantization False |