trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "eval_steps": 20,
  "global_step": 90,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 1.2984,
      "step": 10
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 1.1814,
      "step": 20
    },
    {
      "epoch": 1.11,
      "eval_loss": 1.10374915599823,
      "eval_runtime": 5.2841,
      "eval_samples_per_second": 2.271,
      "eval_steps_per_second": 0.378,
      "step": 20
    },
    {
      "epoch": 1.67,
      "learning_rate": 0.0002,
      "loss": 1.0693,
      "step": 30
    },
    {
      "epoch": 2.22,
      "learning_rate": 0.0002,
      "loss": 1.0172,
      "step": 40
    },
    {
      "epoch": 2.22,
      "eval_loss": 0.9914504885673523,
      "eval_runtime": 5.3058,
      "eval_samples_per_second": 2.262,
      "eval_steps_per_second": 0.377,
      "step": 40
    },
    {
      "epoch": 2.78,
      "learning_rate": 0.0002,
      "loss": 0.9444,
      "step": 50
    },
    {
      "epoch": 3.33,
      "learning_rate": 0.0002,
      "loss": 0.8934,
      "step": 60
    },
    {
      "epoch": 3.33,
      "eval_loss": 0.954533040523529,
      "eval_runtime": 5.3167,
      "eval_samples_per_second": 2.257,
      "eval_steps_per_second": 0.376,
      "step": 60
    },
    {
      "epoch": 3.89,
      "learning_rate": 0.0002,
      "loss": 0.8459,
      "step": 70
    },
    {
      "epoch": 4.44,
      "learning_rate": 0.0002,
      "loss": 0.7643,
      "step": 80
    },
    {
      "epoch": 4.44,
      "eval_loss": 0.9285812973976135,
      "eval_runtime": 5.3161,
      "eval_samples_per_second": 2.257,
      "eval_steps_per_second": 0.376,
      "step": 80
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0002,
      "loss": 0.7527,
      "step": 90
    }
  ],
  "logging_steps": 10,
  "max_steps": 144,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "total_flos": 5.47066223788032e+16,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}