training_parameters.json

{"lora_name": "babel_v8_2e_05_100_16_32", "always_override": false, "save_steps": 0.0, "micro_batch_size": 8, "batch_size": 128, "epochs": 100.0, "learning_rate": "2e-05", "lr_scheduler_type": "linear", "lora_rank": 16, "lora_alpha": 32, "lora_dropout": 0.05, "cutoff_len": 512, "dataset": "babel", "eval_dataset": "None", "format": "alpaca-format-ru", "eval_steps": 100.0, "raw_text_file": "None", "overlap_len": 128, "newline_favor_len": 128, "do_shuffle": true, "higher_rank_limit": false, "warmup_steps": 100.0, "optimizer": "adamw_torch"}