trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 186,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 0.7306758515536785,
      "epoch": 0.21621621621621623,
      "grad_norm": 3.109375,
      "learning_rate": 5e-06,
      "loss": 0.9965,
      "mean_token_accuracy": 0.779215482622385,
      "num_tokens": 6826187.0,
      "step": 20
    },
    {
      "entropy": 0.5489349599927664,
      "epoch": 0.43243243243243246,
      "grad_norm": 0.58203125,
      "learning_rate": 9.998873580873848e-06,
      "loss": 0.5535,
      "mean_token_accuracy": 0.8538637027144432,
      "num_tokens": 13653383.0,
      "step": 40
    },
    {
      "entropy": 0.3483663786202669,
      "epoch": 0.6486486486486487,
      "grad_norm": 0.33984375,
      "learning_rate": 9.511402217292927e-06,
      "loss": 0.351,
      "mean_token_accuracy": 0.8950387261807918,
      "num_tokens": 20485861.0,
      "step": 60
    },
    {
      "entropy": 0.32447177954018114,
      "epoch": 0.8648648648648649,
      "grad_norm": 0.3125,
      "learning_rate": 8.222962883121196e-06,
      "loss": 0.3235,
      "mean_token_accuracy": 0.90108412951231,
      "num_tokens": 27314991.0,
      "step": 80
    },
    {
      "entropy": 0.3140074060513423,
      "epoch": 1.0756756756756758,
      "grad_norm": 0.3203125,
      "learning_rate": 6.3623089878341146e-06,
      "loss": 0.3125,
      "mean_token_accuracy": 0.9037734896708758,
      "num_tokens": 33973895.0,
      "step": 100
    },
    {
      "entropy": 0.3088852509856224,
      "epoch": 1.291891891891892,
      "grad_norm": 0.30078125,
      "learning_rate": 4.259786641731344e-06,
      "loss": 0.3071,
      "mean_token_accuracy": 0.9051578566431999,
      "num_tokens": 40804802.0,
      "step": 120
    },
    {
      "entropy": 0.30427912399172785,
      "epoch": 1.5081081081081082,
      "grad_norm": 0.298828125,
      "learning_rate": 2.288684010125325e-06,
      "loss": 0.3034,
      "mean_token_accuracy": 0.9061512276530266,
      "num_tokens": 47638527.0,
      "step": 140
    },
    {
      "entropy": 0.3052296478301287,
      "epoch": 1.7243243243243245,
      "grad_norm": 0.310546875,
      "learning_rate": 7.989566054312286e-07,
      "loss": 0.3036,
      "mean_token_accuracy": 0.9059888675808907,
      "num_tokens": 54467102.0,
      "step": 160
    },
    {
      "entropy": 0.3061117686331272,
      "epoch": 1.9405405405405407,
      "grad_norm": 0.298828125,
      "learning_rate": 5.509512889877333e-08,
      "loss": 0.3049,
      "mean_token_accuracy": 0.9056108377873897,
      "num_tokens": 61294894.0,
      "step": 180
    },
    {
      "entropy": 0.3062901334329085,
      "epoch": 2.0,
      "mean_token_accuracy": 0.9058145718141035,
      "num_tokens": 63165740.0,
      "step": 186,
      "total_flos": 1.38382400484488e+18,
      "train_loss": 0.41364410038917293,
      "train_runtime": 6390.138,
      "train_samples_per_second": 9.726,
      "train_steps_per_second": 0.029
    }
  ],
  "logging_steps": 20,
  "max_steps": 186,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "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": 1.38382400484488e+18,
  "train_batch_size": 84,
  "trial_name": null,
  "trial_params": null
}