training_info.json

{
  "base_model": "WeiboAI/VibeThinker-1.5B",
  "training_samples": 80000,
  "lora_config": {
    "r": 16,
    "alpha": 32,
    "dropout": 0.05,
    "target_modules": [
      "q_proj",
      "k_proj",
      "v_proj",
      "o_proj"
    ]
  },
  "training_args": {
    "epochs": 1,
    "batch_size": 4,
    "gradient_accumulation": 4,
    "learning_rate": 3e-05,
    "warmup_ratio": 0.03
  },
  "total_steps": 5000
}