trainer_state.json

{
  "best_metric": 0.02908700704574585,
  "best_model_checkpoint": "results_bert-base-italian-xxl-cased/epoch20_bs64/checkpoint-72",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 252,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.8602136373519897,
      "learning_rate": 4.75e-05,
      "loss": 0.1559,
      "step": 36
    },
    {
      "epoch": 1.0,
      "eval_accurracy": {
        "accuracy": 0.9891317050368119
      },
      "eval_f1": [
        0.9865353988707108,
        0.8812260536398467,
        1.0
      ],
      "eval_loss": 0.030727442353963852,
      "eval_precision": [
        0.990119151409474,
        0.8539603960396039,
        1.0
      ],
      "eval_recall": [
        0.9829774956722447,
        0.9102902374670184,
        1.0
      ],
      "eval_runtime": 3.4924,
      "eval_samples_per_second": 56.981,
      "eval_steps_per_second": 1.145,
      "step": 36
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.7721014618873596,
      "learning_rate": 4.5e-05,
      "loss": 0.0329,
      "step": 72
    },
    {
      "epoch": 2.0,
      "eval_accurracy": {
        "accuracy": 0.9892485684235129
      },
      "eval_f1": [
        0.9867511520737328,
        0.8766756032171582,
        1.0
      ],
      "eval_loss": 0.02908700704574585,
      "eval_precision": [
        0.9850488786659,
        0.8910081743869209,
        1.0
      ],
      "eval_recall": [
        0.9884593190998269,
        0.862796833773087,
        1.0
      ],
      "eval_runtime": 3.6114,
      "eval_samples_per_second": 55.104,
      "eval_steps_per_second": 1.108,
      "step": 72
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.09961529076099396,
      "learning_rate": 4.25e-05,
      "loss": 0.0245,
      "step": 108
    },
    {
      "epoch": 3.0,
      "eval_accurracy": {
        "accuracy": 0.9887811148767092
      },
      "eval_f1": [
        0.9860909881193857,
        0.8781725888324873,
        1.0
      ],
      "eval_loss": 0.030218517407774925,
      "eval_precision": [
        0.990395809080326,
        0.8459657701711492,
        1.0
      ],
      "eval_recall": [
        0.9818234275822274,
        0.9129287598944591,
        1.0
      ],
      "eval_runtime": 4.3781,
      "eval_samples_per_second": 45.454,
      "eval_steps_per_second": 0.914,
      "step": 108
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.6222161054611206,
      "learning_rate": 4e-05,
      "loss": 0.0193,
      "step": 144
    },
    {
      "epoch": 4.0,
      "eval_accurracy": {
        "accuracy": 0.9899497487437185
      },
      "eval_f1": [
        0.9876187733947596,
        0.8844086021505376,
        1.0
      ],
      "eval_loss": 0.03514599800109863,
      "eval_precision": [
        0.985632183908046,
        0.9013698630136986,
        1.0
      ],
      "eval_recall": [
        0.9896133871898442,
        0.8680738786279684,
        1.0
      ],
      "eval_runtime": 4.0342,
      "eval_samples_per_second": 49.328,
      "eval_steps_per_second": 0.992,
      "step": 144
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.7397810220718384,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.016,
      "step": 180
    },
    {
      "epoch": 5.0,
      "eval_accurracy": {
        "accuracy": 0.9892485684235129
      },
      "eval_f1": [
        0.9867052023121388,
        0.8805194805194805,
        1.0
      ],
      "eval_loss": 0.03297892585396767,
      "eval_precision": [
        0.9884192240880139,
        0.8670076726342711,
        1.0
      ],
      "eval_recall": [
        0.9849971148297749,
        0.8944591029023746,
        1.0
      ],
      "eval_runtime": 3.7497,
      "eval_samples_per_second": 53.071,
      "eval_steps_per_second": 1.067,
      "step": 180
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.4494711756706238,
      "learning_rate": 3.5e-05,
      "loss": 0.0125,
      "step": 216
    },
    {
      "epoch": 6.0,
      "eval_accurracy": {
        "accuracy": 0.9891317050368119
      },
      "eval_f1": [
        0.9865781498051667,
        0.8777923784494087,
        1.0
      ],
      "eval_loss": 0.03959830850362778,
      "eval_precision": [
        0.9870054865723361,
        0.8743455497382199,
        1.0
      ],
      "eval_recall": [
        0.9861511829197923,
        0.8812664907651715,
        1.0
      ],
      "eval_runtime": 3.3105,
      "eval_samples_per_second": 60.112,
      "eval_steps_per_second": 1.208,
      "step": 216
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.7388216853141785,
      "learning_rate": 3.2500000000000004e-05,
      "loss": 0.0111,
      "step": 252
    },
    {
      "epoch": 7.0,
      "eval_accurracy": {
        "accuracy": 0.9893654318102139
      },
      "eval_f1": [
        0.9868211440984793,
        0.8840764331210191,
        1.0
      ],
      "eval_loss": 0.04391070082783699,
      "eval_precision": [
        0.9906949694678686,
        0.854679802955665,
        1.0
      ],
      "eval_recall": [
        0.9829774956722447,
        0.9155672823218998,
        1.0
      ],
      "eval_runtime": 3.3481,
      "eval_samples_per_second": 59.437,
      "eval_steps_per_second": 1.195,
      "step": 252
    }
  ],
  "logging_steps": 500,
  "max_steps": 720,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 328679588043576.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}