trainer_state.json

{
  "best_global_step": 120,
  "best_metric": 0.7853881278538812,
  "best_model_checkpoint": "/content/gemma_jigsaw_binary/checkpoint-120",
  "epoch": 2.3529411764705883,
  "eval_steps": 20,
  "global_step": 120,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.39215686274509803,
      "grad_norm": 163.88746643066406,
      "learning_rate": 3.8e-07,
      "loss": 2.5213,
      "step": 20
    },
    {
      "epoch": 0.39215686274509803,
      "eval_accuracy": 0.49261083743842365,
      "eval_f1": 0.0,
      "eval_loss": 2.186915159225464,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 3.9996,
      "eval_samples_per_second": 50.755,
      "eval_steps_per_second": 1.75,
      "step": 20
    },
    {
      "epoch": 0.7843137254901961,
      "grad_norm": 144.07699584960938,
      "learning_rate": 7.8e-07,
      "loss": 1.4747,
      "step": 40
    },
    {
      "epoch": 0.7843137254901961,
      "eval_accuracy": 0.5270935960591133,
      "eval_f1": 0.673469387755102,
      "eval_loss": 1.1523324251174927,
      "eval_precision": 0.518324607329843,
      "eval_recall": 0.9611650485436893,
      "eval_runtime": 3.989,
      "eval_samples_per_second": 50.89,
      "eval_steps_per_second": 1.755,
      "step": 40
    },
    {
      "epoch": 1.1764705882352942,
      "grad_norm": 39.203147888183594,
      "learning_rate": 1.1800000000000001e-06,
      "loss": 0.7666,
      "step": 60
    },
    {
      "epoch": 1.1764705882352942,
      "eval_accuracy": 0.6748768472906403,
      "eval_f1": 0.7272727272727273,
      "eval_loss": 0.6169362664222717,
      "eval_precision": 0.6330935251798561,
      "eval_recall": 0.8543689320388349,
      "eval_runtime": 3.9915,
      "eval_samples_per_second": 50.858,
      "eval_steps_per_second": 1.754,
      "step": 60
    },
    {
      "epoch": 1.5686274509803921,
      "grad_norm": 22.82895851135254,
      "learning_rate": 1.5800000000000001e-06,
      "loss": 0.6096,
      "step": 80
    },
    {
      "epoch": 1.5686274509803921,
      "eval_accuracy": 0.6847290640394089,
      "eval_f1": 0.75,
      "eval_loss": 0.5973589420318604,
      "eval_precision": 0.6274509803921569,
      "eval_recall": 0.9320388349514563,
      "eval_runtime": 3.9967,
      "eval_samples_per_second": 50.792,
      "eval_steps_per_second": 1.751,
      "step": 80
    },
    {
      "epoch": 1.9607843137254903,
      "grad_norm": 31.963476181030273,
      "learning_rate": 1.98e-06,
      "loss": 0.5868,
      "step": 100
    },
    {
      "epoch": 1.9607843137254903,
      "eval_accuracy": 0.7586206896551724,
      "eval_f1": 0.7586206896551724,
      "eval_loss": 0.5014973878860474,
      "eval_precision": 0.77,
      "eval_recall": 0.7475728155339806,
      "eval_runtime": 3.9907,
      "eval_samples_per_second": 50.869,
      "eval_steps_per_second": 1.754,
      "step": 100
    },
    {
      "epoch": 2.3529411764705883,
      "grad_norm": 21.746936798095703,
      "learning_rate": 2.38e-06,
      "loss": 0.5258,
      "step": 120
    },
    {
      "epoch": 2.3529411764705883,
      "eval_accuracy": 0.7684729064039408,
      "eval_f1": 0.7853881278538812,
      "eval_loss": 0.4798792600631714,
      "eval_precision": 0.7413793103448276,
      "eval_recall": 0.8349514563106796,
      "eval_runtime": 3.9892,
      "eval_samples_per_second": 50.887,
      "eval_steps_per_second": 1.755,
      "step": 120
    }
  ],
  "logging_steps": 20,
  "max_steps": 153,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 20,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.5720898182447104e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}