last-checkpoint/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.003126914707063024,
  "eval_steps": 5,
  "global_step": 37,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 8.451120829900065e-05,
      "eval_loss": 4.7536301612854,
      "eval_runtime": 211.0833,
      "eval_samples_per_second": 23.607,
      "eval_steps_per_second": 11.806,
      "step": 1
    },
    {
      "epoch": 0.00025353362489700194,
      "grad_norm": 9.546398162841797,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 4.5522,
      "step": 3
    },
    {
      "epoch": 0.00042255604149500327,
      "eval_loss": 4.59783935546875,
      "eval_runtime": 210.419,
      "eval_samples_per_second": 23.681,
      "eval_steps_per_second": 11.843,
      "step": 5
    },
    {
      "epoch": 0.0005070672497940039,
      "grad_norm": 8.098456382751465,
      "learning_rate": 6.666666666666667e-05,
      "loss": 4.6234,
      "step": 6
    },
    {
      "epoch": 0.0007606008746910059,
      "grad_norm": 2.886258840560913,
      "learning_rate": 0.0001,
      "loss": 4.2417,
      "step": 9
    },
    {
      "epoch": 0.0008451120829900065,
      "eval_loss": 3.9523026943206787,
      "eval_runtime": 210.1781,
      "eval_samples_per_second": 23.708,
      "eval_steps_per_second": 11.857,
      "step": 10
    },
    {
      "epoch": 0.0010141344995880078,
      "grad_norm": 0.9398317337036133,
      "learning_rate": 0.00013333333333333334,
      "loss": 3.7207,
      "step": 12
    },
    {
      "epoch": 0.0012676681244850098,
      "grad_norm": 0.7323232889175415,
      "learning_rate": 0.0001666666666666667,
      "loss": 3.6179,
      "step": 15
    },
    {
      "epoch": 0.0012676681244850098,
      "eval_loss": 3.8297524452209473,
      "eval_runtime": 210.5731,
      "eval_samples_per_second": 23.664,
      "eval_steps_per_second": 11.834,
      "step": 15
    },
    {
      "epoch": 0.0015212017493820118,
      "grad_norm": 0.7559709548950195,
      "learning_rate": 0.0002,
      "loss": 3.6705,
      "step": 18
    },
    {
      "epoch": 0.001690224165980013,
      "eval_loss": 3.7751104831695557,
      "eval_runtime": 210.5833,
      "eval_samples_per_second": 23.663,
      "eval_steps_per_second": 11.834,
      "step": 20
    },
    {
      "epoch": 0.0017747353742790138,
      "grad_norm": 0.6443007588386536,
      "learning_rate": 0.0001879473751206489,
      "loss": 3.5284,
      "step": 21
    },
    {
      "epoch": 0.0020282689991760155,
      "grad_norm": 0.5663931965827942,
      "learning_rate": 0.00015469481581224272,
      "loss": 3.4157,
      "step": 24
    },
    {
      "epoch": 0.002112780207475016,
      "eval_loss": 3.723708152770996,
      "eval_runtime": 210.5845,
      "eval_samples_per_second": 23.663,
      "eval_steps_per_second": 11.834,
      "step": 25
    },
    {
      "epoch": 0.0022818026240730175,
      "grad_norm": 0.6620070338249207,
      "learning_rate": 0.00010825793454723325,
      "loss": 3.5988,
      "step": 27
    },
    {
      "epoch": 0.0025353362489700195,
      "grad_norm": 0.5293451547622681,
      "learning_rate": 5.983045753470308e-05,
      "loss": 3.4641,
      "step": 30
    },
    {
      "epoch": 0.0025353362489700195,
      "eval_loss": 3.6815202236175537,
      "eval_runtime": 210.5611,
      "eval_samples_per_second": 23.665,
      "eval_steps_per_second": 11.835,
      "step": 30
    },
    {
      "epoch": 0.0027888698738670215,
      "grad_norm": 0.5607340335845947,
      "learning_rate": 2.1085949060360654e-05,
      "loss": 3.4125,
      "step": 33
    },
    {
      "epoch": 0.002957892290465023,
      "eval_loss": 3.6705963611602783,
      "eval_runtime": 210.593,
      "eval_samples_per_second": 23.662,
      "eval_steps_per_second": 11.833,
      "step": 35
    },
    {
      "epoch": 0.0030424034987640235,
      "grad_norm": 0.6579822897911072,
      "learning_rate": 1.3638696597277679e-06,
      "loss": 3.4316,
      "step": 36
    }
  ],
  "logging_steps": 3,
  "max_steps": 37,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 18,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 598445660307456.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}