trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.9047619047619047,
  "eval_steps": 500,
  "global_step": 20,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.19047619047619047,
      "grad_norm": 2.075279951095581,
      "learning_rate": 0.00018,
      "loss": 9.5584,
      "step": 2
    },
    {
      "epoch": 0.38095238095238093,
      "grad_norm": 1.2216403484344482,
      "learning_rate": 0.00016,
      "loss": 9.2621,
      "step": 4
    },
    {
      "epoch": 0.5714285714285714,
      "grad_norm": 1.9454573392868042,
      "learning_rate": 0.00014,
      "loss": 8.8941,
      "step": 6
    },
    {
      "epoch": 0.7619047619047619,
      "grad_norm": 2.4250237941741943,
      "learning_rate": 0.00012,
      "loss": 8.4539,
      "step": 8
    },
    {
      "epoch": 0.9523809523809523,
      "grad_norm": 1.720076560974121,
      "learning_rate": 0.0001,
      "loss": 8.1194,
      "step": 10
    },
    {
      "epoch": 1.1428571428571428,
      "grad_norm": 1.4687391519546509,
      "learning_rate": 8e-05,
      "loss": 7.9072,
      "step": 12
    },
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 1.1075403690338135,
      "learning_rate": 6e-05,
      "loss": 7.7305,
      "step": 14
    },
    {
      "epoch": 1.5238095238095237,
      "grad_norm": 1.532745361328125,
      "learning_rate": 4e-05,
      "loss": 7.6572,
      "step": 16
    },
    {
      "epoch": 1.7142857142857144,
      "grad_norm": 1.2055246829986572,
      "learning_rate": 2e-05,
      "loss": 7.6057,
      "step": 18
    },
    {
      "epoch": 1.9047619047619047,
      "grad_norm": 0.9867807030677795,
      "learning_rate": 0.0,
      "loss": 7.5695,
      "step": 20
    },
    {
      "epoch": 1.9047619047619047,
      "step": 20,
      "total_flos": 99814369025136.0,
      "train_loss": 8.27579402923584,
      "train_runtime": 84.8689,
      "train_samples_per_second": 3.959,
      "train_steps_per_second": 0.236
    }
  ],
  "logging_steps": 2,
  "max_steps": 20,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 99814369025136.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}