trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 5268,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2847380410022779,
      "grad_norm": 8.189080238342285,
      "learning_rate": 1.810174639331815e-05,
      "loss": 0.275,
      "step": 500
    },
    {
      "epoch": 0.5694760820045558,
      "grad_norm": 0.24108383059501648,
      "learning_rate": 1.6203492786636296e-05,
      "loss": 0.0972,
      "step": 1000
    },
    {
      "epoch": 0.8542141230068337,
      "grad_norm": 2.321260929107666,
      "learning_rate": 1.4305239179954442e-05,
      "loss": 0.0768,
      "step": 1500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9830492318934977,
      "eval_f1": 0.9187202405814051,
      "eval_loss": 0.05956669896841049,
      "eval_precision": 0.9120915574722176,
      "eval_recall": 0.9254459777852575,
      "eval_runtime": 5.9807,
      "eval_samples_per_second": 543.418,
      "eval_steps_per_second": 68.053,
      "step": 1756
    },
    {
      "epoch": 1.1389521640091116,
      "grad_norm": 4.685558319091797,
      "learning_rate": 1.240698557327259e-05,
      "loss": 0.0589,
      "step": 2000
    },
    {
      "epoch": 1.4236902050113895,
      "grad_norm": 0.9782440066337585,
      "learning_rate": 1.0508731966590738e-05,
      "loss": 0.0434,
      "step": 2500
    },
    {
      "epoch": 1.7084282460136673,
      "grad_norm": 0.514324426651001,
      "learning_rate": 8.610478359908885e-06,
      "loss": 0.0423,
      "step": 3000
    },
    {
      "epoch": 1.9931662870159452,
      "grad_norm": 5.84403657913208,
      "learning_rate": 6.712224753227031e-06,
      "loss": 0.0365,
      "step": 3500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9852415524171129,
      "eval_f1": 0.9326105087572978,
      "eval_loss": 0.05695108696818352,
      "eval_precision": 0.9244378306878307,
      "eval_recall": 0.9409289801413665,
      "eval_runtime": 6.4521,
      "eval_samples_per_second": 503.712,
      "eval_steps_per_second": 63.08,
      "step": 3512
    },
    {
      "epoch": 2.277904328018223,
      "grad_norm": 0.13090287148952484,
      "learning_rate": 4.8139711465451785e-06,
      "loss": 0.0233,
      "step": 4000
    },
    {
      "epoch": 2.562642369020501,
      "grad_norm": 0.17942672967910767,
      "learning_rate": 2.9157175398633257e-06,
      "loss": 0.0219,
      "step": 4500
    },
    {
      "epoch": 2.847380410022779,
      "grad_norm": 2.7603061199188232,
      "learning_rate": 1.0174639331814731e-06,
      "loss": 0.0237,
      "step": 5000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9855275072680191,
      "eval_f1": 0.9348825545431748,
      "eval_loss": 0.059884414076805115,
      "eval_precision": 0.9287493771798705,
      "eval_recall": 0.9410972736452373,
      "eval_runtime": 6.5591,
      "eval_samples_per_second": 495.492,
      "eval_steps_per_second": 62.051,
      "step": 5268
    }
  ],
  "logging_steps": 500,
  "max_steps": 5268,
  "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": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 445994355589020.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}