training_parameters.json

{
  "lora_name": "mbtk.py",
  "always_override": true,
  "save_steps": 0.0,
  "micro_batch_size": 2,
  "batch_size": 96,
  "epochs": 48.0,
  "learning_rate": "3e-4",
  "lr_scheduler_type": "inverse_sqrt",
  "lora_rank": 16,
  "lora_alpha": 24,
  "lora_dropout": 0.05,
  "cutoff_len": 448,
  "dataset": "None",
  "eval_dataset": "None",
  "format": "None",
  "eval_steps": 100.0,
  "raw_text_file": "einmbtkreaktorkannnichteinfach",
  "overlap_len": 96,
  "newline_favor_len": 0,
  "higher_rank_limit": false,
  "warmup_steps": 100.0,
  "optimizer": "adamw_torch",
  "hard_cut_string": "/0",
  "train_only_after": "",
  "stop_at_loss": 0,
  "add_eos_token": false,
  "min_chars": 0.0,
  "report_to": "None"
}