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": 7.551333904266357,
      "learning_rate": 1.810174639331815e-05,
      "loss": 0.2775,
      "step": 500
    },
    {
      "epoch": 0.5694760820045558,
      "grad_norm": 0.47765791416168213,
      "learning_rate": 1.6203492786636296e-05,
      "loss": 0.096,
      "step": 1000
    },
    {
      "epoch": 0.8542141230068337,
      "grad_norm": 2.6445720195770264,
      "learning_rate": 1.4305239179954442e-05,
      "loss": 0.0765,
      "step": 1500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9827791634231973,
      "eval_f1": 0.9163957656080687,
      "eval_loss": 0.06094542145729065,
      "eval_precision": 0.9078447563996697,
      "eval_recall": 0.925109390777516,
      "eval_runtime": 6.5104,
      "eval_samples_per_second": 499.199,
      "eval_steps_per_second": 62.515,
      "step": 1756
    },
    {
      "epoch": 1.1389521640091116,
      "grad_norm": 6.109570026397705,
      "learning_rate": 1.240698557327259e-05,
      "loss": 0.0578,
      "step": 2000
    },
    {
      "epoch": 1.4236902050113895,
      "grad_norm": 1.6886531114578247,
      "learning_rate": 1.0508731966590738e-05,
      "loss": 0.0428,
      "step": 2500
    },
    {
      "epoch": 1.7084282460136673,
      "grad_norm": 2.353813648223877,
      "learning_rate": 8.610478359908885e-06,
      "loss": 0.0427,
      "step": 3000
    },
    {
      "epoch": 1.9931662870159452,
      "grad_norm": 7.191911220550537,
      "learning_rate": 6.712224753227031e-06,
      "loss": 0.0347,
      "step": 3500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9854957345068073,
      "eval_f1": 0.9329772139220431,
      "eval_loss": 0.057699523866176605,
      "eval_precision": 0.9254843517138599,
      "eval_recall": 0.9405923931336251,
      "eval_runtime": 6.8321,
      "eval_samples_per_second": 475.697,
      "eval_steps_per_second": 59.572,
      "step": 3512
    },
    {
      "epoch": 2.277904328018223,
      "grad_norm": 0.13887672126293182,
      "learning_rate": 4.8139711465451785e-06,
      "loss": 0.023,
      "step": 4000
    },
    {
      "epoch": 2.562642369020501,
      "grad_norm": 0.30556178092956543,
      "learning_rate": 2.9157175398633257e-06,
      "loss": 0.0213,
      "step": 4500
    },
    {
      "epoch": 2.847380410022779,
      "grad_norm": 1.6076356172561646,
      "learning_rate": 1.0174639331814731e-06,
      "loss": 0.0225,
      "step": 5000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9857658029771077,
      "eval_f1": 0.9344248600785231,
      "eval_loss": 0.059937961399555206,
      "eval_precision": 0.9276828661469564,
      "eval_recall": 0.941265567149108,
      "eval_runtime": 6.7177,
      "eval_samples_per_second": 483.793,
      "eval_steps_per_second": 60.586,
      "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
}