checkpoint-75/trainer_state.json

{
  "best_global_step": 50,
  "best_metric": 0.37461215257644653,
  "best_model_checkpoint": "./deepseek-pentest-model-v2/checkpoint-50",
  "epoch": 3.0,
  "eval_steps": 50,
  "global_step": 75,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.40404040404040403,
      "grad_norm": 0.6140235662460327,
      "learning_rate": 9e-05,
      "loss": 2.1515,
      "step": 10
    },
    {
      "epoch": 0.8080808080808081,
      "grad_norm": 0.7330130934715271,
      "learning_rate": 0.00019,
      "loss": 1.7919,
      "step": 20
    },
    {
      "epoch": 1.202020202020202,
      "grad_norm": 0.5969696640968323,
      "learning_rate": 0.00016727272727272728,
      "loss": 1.0693,
      "step": 30
    },
    {
      "epoch": 1.606060606060606,
      "grad_norm": 0.5808431506156921,
      "learning_rate": 0.00013090909090909093,
      "loss": 0.7384,
      "step": 40
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.6547220945358276,
      "learning_rate": 9.454545454545455e-05,
      "loss": 0.4897,
      "step": 50
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.37461215257644653,
      "eval_runtime": 5.244,
      "eval_samples_per_second": 8.391,
      "eval_steps_per_second": 4.195,
      "step": 50
    },
    {
      "epoch": 2.404040404040404,
      "grad_norm": 0.44966480135917664,
      "learning_rate": 5.818181818181818e-05,
      "loss": 0.2932,
      "step": 60
    },
    {
      "epoch": 2.808080808080808,
      "grad_norm": 0.5388813614845276,
      "learning_rate": 2.1818181818181818e-05,
      "loss": 0.2162,
      "step": 70
    }
  ],
  "logging_steps": 10,
  "max_steps": 75,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4684388851777536.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}