trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5961782336235046,
  "best_model_checkpoint": "distilbert_km_5_v2_mrpc/checkpoint-30",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 105,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.4399302005767822,
      "learning_rate": 4.9e-05,
      "loss": 0.6185,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.678921568627451,
      "eval_combined_score": 0.7346607843137255,
      "eval_f1": 0.7904,
      "eval_loss": 0.6003154516220093,
      "eval_runtime": 0.1953,
      "eval_samples_per_second": 2089.263,
      "eval_steps_per_second": 10.241,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.7464332580566406,
      "learning_rate": 4.8e-05,
      "loss": 0.5391,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6862745098039216,
      "eval_combined_score": 0.739911448450348,
      "eval_f1": 0.7935483870967742,
      "eval_loss": 0.5961782336235046,
      "eval_runtime": 0.1972,
      "eval_samples_per_second": 2068.74,
      "eval_steps_per_second": 10.141,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.069668769836426,
      "learning_rate": 4.7e-05,
      "loss": 0.4435,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.696078431372549,
      "eval_combined_score": 0.7453902090637579,
      "eval_f1": 0.7947019867549668,
      "eval_loss": 0.6492855548858643,
      "eval_runtime": 0.1944,
      "eval_samples_per_second": 2098.371,
      "eval_steps_per_second": 10.286,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.9558331966400146,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.3126,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6764705882352942,
      "eval_combined_score": 0.7256073077695242,
      "eval_f1": 0.7747440273037542,
      "eval_loss": 0.7969324588775635,
      "eval_runtime": 0.1902,
      "eval_samples_per_second": 2145.222,
      "eval_steps_per_second": 10.516,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.951694965362549,
      "learning_rate": 4.5e-05,
      "loss": 0.1757,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6544117647058824,
      "eval_combined_score": 0.7024271212909943,
      "eval_f1": 0.7504424778761062,
      "eval_loss": 1.0384982824325562,
      "eval_runtime": 0.1911,
      "eval_samples_per_second": 2134.812,
      "eval_steps_per_second": 10.465,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.9431947469711304,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.0809,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6838235294117647,
      "eval_combined_score": 0.7316553544494722,
      "eval_f1": 0.7794871794871795,
      "eval_loss": 1.346550464630127,
      "eval_runtime": 0.217,
      "eval_samples_per_second": 1880.019,
      "eval_steps_per_second": 9.216,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.6144425868988037,
      "learning_rate": 4.3e-05,
      "loss": 0.0458,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6838235294117647,
      "eval_combined_score": 0.7293463196797043,
      "eval_f1": 0.774869109947644,
      "eval_loss": 1.5137430429458618,
      "eval_runtime": 0.1879,
      "eval_samples_per_second": 2171.13,
      "eval_steps_per_second": 10.643,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.31660059009279523,
      "train_runtime": 34.3244,
      "train_samples_per_second": 5343.141,
      "train_steps_per_second": 21.85
    }
  ],
  "logging_steps": 1,
  "max_steps": 750,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 50,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 5,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 5
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1700616463945728.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}