trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 300.0,
  "global_step": 2826,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5307855626326964,
      "grad_norm": 10.279387474060059,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 1.9545,
      "step": 500
    },
    {
      "epoch": 1.0615711252653928,
      "grad_norm": 7.024175643920898,
      "learning_rate": 1.6406109613656783e-05,
      "loss": 0.9557,
      "step": 1000
    },
    {
      "epoch": 1.5923566878980893,
      "grad_norm": 7.754688262939453,
      "learning_rate": 1.1913746630727763e-05,
      "loss": 0.7767,
      "step": 1500
    },
    {
      "epoch": 2.1231422505307855,
      "grad_norm": 9.531620979309082,
      "learning_rate": 7.421383647798742e-06,
      "loss": 0.6899,
      "step": 2000
    },
    {
      "epoch": 2.653927813163482,
      "grad_norm": 6.811396598815918,
      "learning_rate": 2.929020664869722e-06,
      "loss": 0.5649,
      "step": 2500
    },
    {
      "epoch": 3.0,
      "step": 2826,
      "total_flos": 2.377533515518771e+16,
      "train_loss": 0.9355160604532273,
      "train_runtime": 47353.9083,
      "train_samples_per_second": 1.908,
      "train_steps_per_second": 0.06
    }
  ],
  "logging_steps": 500,
  "max_steps": 2826,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.377533515518771e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}