{
  "train_file": "data/historytrans/train.jsonl",
  "extra_train_file": [],
  "extra_repeat": 1,
  "validation_file": "data/historytrans/validation.jsonl",
  "model": "Qwen/Qwen3-1.7B",
  "output_dir": "outputs/qwen3-1.7b-wenyan-historytrans-full",
  "epochs": 1.0,
  "max_steps": -1,
  "max_length": 512,
  "batch_size": 1,
  "eval_batch_size": 1,
  "gradient_accumulation": 16,
  "learning_rate": 0.0002,
  "warmup_ratio": 0.03,
  "weight_decay": 0.01,
  "lora_rank": 16,
  "lora_alpha": 32,
  "lora_dropout": 0.05,
  "logging_steps": 10,
  "save_steps": 500,
  "save_total_limit": 2,
  "seed": 42,
  "train_limit": null,
  "validation_limit": null,
  "resume": false,
  "device": "cuda",
  "dtype": "torch.bfloat16",
  "train_examples": 747630,
  "base_train_examples": 747630,
  "extra_train_examples": 0,
  "validation_examples": 7644,
  "parameters": {
    "total": 1738007552,
    "trainable": 17432576,
    "trainable_percent": 1.0030207279559622
  },
  "first_example_supervised_tokens": 21,
  "estimated_update_steps": 46727,
  "warmup_steps": 1401
}