{ | |
"micro_batch_size": 1, | |
"gradient_accumulation_steps": 1, | |
"num_train_epochs": 10, | |
"learning_rate": 0.0003, | |
"cutoff_len": 1024, | |
"val_set_size": 0, | |
"lora_r": 8, | |
"lora_alpha": 16, | |
"lora_dropout": 0.05, | |
"lora_target_modules": [ | |
"q_proj", | |
"v_proj" | |
], | |
"lora_modules_to_save": [], | |
"train_on_inputs": true, | |
"group_by_length": false, | |
"load_in_8bit": true, | |
"fp16": false, | |
"bf16": false, | |
"gradient_checkpointing": false, | |
"save_steps": 500, | |
"save_total_limit": 5, | |
"logging_steps": 10, | |
"additional_training_arguments": null, | |
"additional_lora_config": null | |
} |