trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 7500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2,
      "grad_norm": 5.575812339782715,
      "learning_rate": 1.9781476007338058e-05,
      "loss": 0.9324,
      "step": 500
    },
    {
      "epoch": 0.4,
      "grad_norm": 4.583652973175049,
      "learning_rate": 1.913545457642601e-05,
      "loss": 0.9619,
      "step": 1000
    },
    {
      "epoch": 0.6,
      "grad_norm": 3.9456448554992676,
      "learning_rate": 1.8090169943749477e-05,
      "loss": 0.979,
      "step": 1500
    },
    {
      "epoch": 0.8,
      "grad_norm": 5.700883865356445,
      "learning_rate": 1.6691306063588583e-05,
      "loss": 0.9591,
      "step": 2000
    },
    {
      "epoch": 1.0,
      "grad_norm": 3.7766757011413574,
      "learning_rate": 1.5000000000000002e-05,
      "loss": 0.9381,
      "step": 2500
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.937586784362793,
      "eval_runtime": 43.9207,
      "eval_samples_per_second": 45.537,
      "eval_steps_per_second": 11.384,
      "step": 2500
    },
    {
      "epoch": 1.2,
      "grad_norm": 4.477228164672852,
      "learning_rate": 1.3090169943749475e-05,
      "loss": 0.4559,
      "step": 3000
    },
    {
      "epoch": 1.4,
      "grad_norm": 4.878340721130371,
      "learning_rate": 1.1045284632676535e-05,
      "loss": 0.4521,
      "step": 3500
    },
    {
      "epoch": 1.6,
      "grad_norm": 3.9909627437591553,
      "learning_rate": 8.954715367323468e-06,
      "loss": 0.4468,
      "step": 4000
    },
    {
      "epoch": 1.8,
      "grad_norm": 2.667670726776123,
      "learning_rate": 6.909830056250527e-06,
      "loss": 0.4321,
      "step": 4500
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.369731903076172,
      "learning_rate": 5.000000000000003e-06,
      "loss": 0.4124,
      "step": 5000
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.0478131771087646,
      "eval_runtime": 41.0932,
      "eval_samples_per_second": 48.67,
      "eval_steps_per_second": 12.167,
      "step": 5000
    },
    {
      "epoch": 2.2,
      "grad_norm": 2.2593352794647217,
      "learning_rate": 3.308693936411421e-06,
      "loss": 0.1657,
      "step": 5500
    },
    {
      "epoch": 2.4,
      "grad_norm": 2.3517982959747314,
      "learning_rate": 1.9098300562505266e-06,
      "loss": 0.1616,
      "step": 6000
    },
    {
      "epoch": 2.6,
      "grad_norm": 2.2167694568634033,
      "learning_rate": 8.645454235739903e-07,
      "loss": 0.153,
      "step": 6500
    },
    {
      "epoch": 2.8,
      "grad_norm": 2.2378008365631104,
      "learning_rate": 2.1852399266194312e-07,
      "loss": 0.1521,
      "step": 7000
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.863773226737976,
      "learning_rate": 0.0,
      "loss": 0.1515,
      "step": 7500
    },
    {
      "epoch": 3.0,
      "eval_loss": 1.4480849504470825,
      "eval_runtime": 41.4492,
      "eval_samples_per_second": 48.252,
      "eval_steps_per_second": 12.063,
      "step": 7500
    },
    {
      "epoch": 3.0,
      "step": 7500,
      "total_flos": 2.3924465601190298e+17,
      "train_loss": 0.5169153635660807,
      "train_runtime": 5042.7608,
      "train_samples_per_second": 5.949,
      "train_steps_per_second": 1.487
    }
  ],
  "logging_steps": 500,
  "max_steps": 7500,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.3924465601190298e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}