trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.9647058823529413,
  "eval_steps": 500,
  "global_step": 126,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.23529411764705882,
      "grad_norm": 6.322516853155574,
      "learning_rate": 5e-06,
      "loss": 1.0515,
      "step": 10
    },
    {
      "epoch": 0.47058823529411764,
      "grad_norm": 14.560379257369785,
      "learning_rate": 5e-06,
      "loss": 0.9513,
      "step": 20
    },
    {
      "epoch": 0.7058823529411765,
      "grad_norm": 4.479057532531375,
      "learning_rate": 5e-06,
      "loss": 0.9027,
      "step": 30
    },
    {
      "epoch": 0.9411764705882353,
      "grad_norm": 3.759321153679887,
      "learning_rate": 5e-06,
      "loss": 0.8767,
      "step": 40
    },
    {
      "epoch": 0.9882352941176471,
      "eval_loss": 0.8602203726768494,
      "eval_runtime": 29.141,
      "eval_samples_per_second": 39.223,
      "eval_steps_per_second": 0.618,
      "step": 42
    },
    {
      "epoch": 1.1764705882352942,
      "grad_norm": 3.2738441668490785,
      "learning_rate": 5e-06,
      "loss": 0.8788,
      "step": 50
    },
    {
      "epoch": 1.4117647058823528,
      "grad_norm": 1.1791187368515386,
      "learning_rate": 5e-06,
      "loss": 0.8148,
      "step": 60
    },
    {
      "epoch": 1.6470588235294117,
      "grad_norm": 1.5795343558619792,
      "learning_rate": 5e-06,
      "loss": 0.8062,
      "step": 70
    },
    {
      "epoch": 1.8823529411764706,
      "grad_norm": 1.257105712897486,
      "learning_rate": 5e-06,
      "loss": 0.7964,
      "step": 80
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.8092531561851501,
      "eval_runtime": 28.3785,
      "eval_samples_per_second": 40.277,
      "eval_steps_per_second": 0.634,
      "step": 85
    },
    {
      "epoch": 2.1176470588235294,
      "grad_norm": 1.2441883026945793,
      "learning_rate": 5e-06,
      "loss": 0.8096,
      "step": 90
    },
    {
      "epoch": 2.3529411764705883,
      "grad_norm": 0.9557577996128518,
      "learning_rate": 5e-06,
      "loss": 0.7382,
      "step": 100
    },
    {
      "epoch": 2.588235294117647,
      "grad_norm": 1.3694309859333422,
      "learning_rate": 5e-06,
      "loss": 0.7448,
      "step": 110
    },
    {
      "epoch": 2.8235294117647056,
      "grad_norm": 1.4014362052372222,
      "learning_rate": 5e-06,
      "loss": 0.7383,
      "step": 120
    },
    {
      "epoch": 2.9647058823529413,
      "eval_loss": 0.7979101538658142,
      "eval_runtime": 27.3252,
      "eval_samples_per_second": 41.83,
      "eval_steps_per_second": 0.659,
      "step": 126
    },
    {
      "epoch": 2.9647058823529413,
      "step": 126,
      "total_flos": 210845313269760.0,
      "train_loss": 0.8375598930177235,
      "train_runtime": 4383.6848,
      "train_samples_per_second": 14.857,
      "train_steps_per_second": 0.029
    }
  ],
  "logging_steps": 10,
  "max_steps": 126,
  "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": 210845313269760.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}