| { | |
| "lora_name": "Nemo", | |
| "always_override": true, | |
| "q_proj_en": true, | |
| "v_proj_en": true, | |
| "k_proj_en": true, | |
| "o_proj_en": true, | |
| "gate_proj_en": true, | |
| "down_proj_en": true, | |
| "up_proj_en": true, | |
| "save_steps": 1000, | |
| "micro_batch_size": 1, | |
| "batch_size": 2, | |
| "epochs": 42, | |
| "learning_rate": "1e-4", | |
| "lr_scheduler_type": "polynomial", | |
| "lora_rank": 2048, | |
| "lora_alpha": 4096, | |
| "lora_dropout": 0, | |
| "cutoff_len": 4096, | |
| "dataset": "BLACKSHEEP-Toxic_Assistant", | |
| "eval_dataset": "None", | |
| "format": "unwoke_alpaca-chatbot-format", | |
| "eval_steps": 100, | |
| "raw_text_file": "None", | |
| "overlap_len": 512, | |
| "newline_favor_len": 512, | |
| "higher_rank_limit": true, | |
| "warmup_steps": 0, | |
| "optimizer": "adamw_bnb_8bit", | |
| "hard_cut_string": "\\n\\n\\n", | |
| "train_only_after": "", | |
| "stop_at_loss": 0, | |
| "add_eos_token": false, | |
| "min_chars": 0, | |
| "report_to": "tensorboard" | |
| } |