last-checkpoint/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0434782608695652,
  "eval_steps": 3,
  "global_step": 12,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.08695652173913043,
      "eval_loss": 8.65756607055664,
      "eval_runtime": 0.7568,
      "eval_samples_per_second": 6.607,
      "eval_steps_per_second": 3.964,
      "step": 1
    },
    {
      "epoch": 0.2608695652173913,
      "eval_loss": 8.638962745666504,
      "eval_runtime": 0.7572,
      "eval_samples_per_second": 6.603,
      "eval_steps_per_second": 3.962,
      "step": 3
    },
    {
      "epoch": 0.43478260869565216,
      "grad_norm": 17.402477264404297,
      "learning_rate": 5e-05,
      "loss": 37.8497,
      "step": 5
    },
    {
      "epoch": 0.5217391304347826,
      "eval_loss": 8.449831008911133,
      "eval_runtime": 0.757,
      "eval_samples_per_second": 6.605,
      "eval_steps_per_second": 3.963,
      "step": 6
    },
    {
      "epoch": 0.782608695652174,
      "eval_loss": 7.207528591156006,
      "eval_runtime": 0.7574,
      "eval_samples_per_second": 6.602,
      "eval_steps_per_second": 3.961,
      "step": 9
    },
    {
      "epoch": 0.8695652173913043,
      "grad_norm": 29.091482162475586,
      "learning_rate": 0.0001,
      "loss": 36.0964,
      "step": 10
    },
    {
      "epoch": 1.0434782608695652,
      "eval_loss": 4.865574836730957,
      "eval_runtime": 0.758,
      "eval_samples_per_second": 6.596,
      "eval_steps_per_second": 3.958,
      "step": 12
    }
  ],
  "logging_steps": 5,
  "max_steps": 12,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 3,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4709758654218240.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}