trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 7325,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.3412969283276451,
      "grad_norm": 1.6108611822128296,
      "learning_rate": 9.343344709897611e-06,
      "loss": 0.8698,
      "mean_token_accuracy": 0.7602084985375405,
      "step": 500
    },
    {
      "epoch": 0.6825938566552902,
      "grad_norm": 1.4324573278427124,
      "learning_rate": 8.662116040955632e-06,
      "loss": 0.7561,
      "mean_token_accuracy": 0.7782786710858345,
      "step": 1000
    },
    {
      "epoch": 1.023890784982935,
      "grad_norm": 1.472659707069397,
      "learning_rate": 7.979522184300343e-06,
      "loss": 0.7244,
      "mean_token_accuracy": 0.7850172371268272,
      "step": 1500
    },
    {
      "epoch": 1.36518771331058,
      "grad_norm": 1.4533411264419556,
      "learning_rate": 7.296928327645052e-06,
      "loss": 0.6679,
      "mean_token_accuracy": 0.7980254861712456,
      "step": 2000
    },
    {
      "epoch": 1.7064846416382253,
      "grad_norm": 1.3612010478973389,
      "learning_rate": 6.614334470989762e-06,
      "loss": 0.6608,
      "mean_token_accuracy": 0.7996546522378921,
      "step": 2500
    },
    {
      "epoch": 2.04778156996587,
      "grad_norm": 1.4121689796447754,
      "learning_rate": 5.931740614334472e-06,
      "loss": 0.6454,
      "mean_token_accuracy": 0.8033261907696724,
      "step": 3000
    },
    {
      "epoch": 2.3890784982935154,
      "grad_norm": 1.5300949811935425,
      "learning_rate": 5.249146757679181e-06,
      "loss": 0.6032,
      "mean_token_accuracy": 0.8137686540484428,
      "step": 3500
    },
    {
      "epoch": 2.73037542662116,
      "grad_norm": 1.4691298007965088,
      "learning_rate": 4.567918088737202e-06,
      "loss": 0.6026,
      "mean_token_accuracy": 0.8138553571701049,
      "step": 4000
    },
    {
      "epoch": 3.0716723549488054,
      "grad_norm": 1.4363794326782227,
      "learning_rate": 3.885324232081912e-06,
      "loss": 0.5918,
      "mean_token_accuracy": 0.8168109318614006,
      "step": 4500
    },
    {
      "epoch": 3.4129692832764507,
      "grad_norm": 1.4927068948745728,
      "learning_rate": 3.2040955631399316e-06,
      "loss": 0.5586,
      "mean_token_accuracy": 0.8255660828351975,
      "step": 5000
    },
    {
      "epoch": 3.7542662116040955,
      "grad_norm": 1.522266149520874,
      "learning_rate": 2.521501706484642e-06,
      "loss": 0.5575,
      "mean_token_accuracy": 0.8258042855262756,
      "step": 5500
    },
    {
      "epoch": 4.09556313993174,
      "grad_norm": 1.545706868171692,
      "learning_rate": 1.8402730375426621e-06,
      "loss": 0.5472,
      "mean_token_accuracy": 0.8288247205018997,
      "step": 6000
    },
    {
      "epoch": 4.436860068259386,
      "grad_norm": 1.6156072616577148,
      "learning_rate": 1.157679180887372e-06,
      "loss": 0.5262,
      "mean_token_accuracy": 0.8347329877018929,
      "step": 6500
    },
    {
      "epoch": 4.778156996587031,
      "grad_norm": 1.533884048461914,
      "learning_rate": 4.750853242320819e-07,
      "loss": 0.5253,
      "mean_token_accuracy": 0.835058969438076,
      "step": 7000
    },
    {
      "epoch": 5.0,
      "mean_token_accuracy": 0.8352870517510634,
      "step": 7325,
      "total_flos": 2.0101550322378342e+18,
      "train_loss": 0.6264336887398677,
      "train_runtime": 22003.5983,
      "train_samples_per_second": 21.299,
      "train_steps_per_second": 0.333
    }
  ],
  "logging_steps": 500,
  "max_steps": 7325,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "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": 2.0101550322378342e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}