trainer_state.json

{
  "best_metric": 0.5924765467643738,
  "best_model_checkpoint": "bert_tiny_lda_100_v1_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": 0.7286933064460754,
      "learning_rate": 4.9e-05,
      "loss": 0.631,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.696078431372549,
      "eval_combined_score": 0.756593787957661,
      "eval_f1": 0.8171091445427728,
      "eval_loss": 0.5995902419090271,
      "eval_runtime": 0.1333,
      "eval_samples_per_second": 3060.851,
      "eval_steps_per_second": 15.004,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.1771771907806396,
      "learning_rate": 4.8e-05,
      "loss": 0.5947,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6813725490196079,
      "eval_combined_score": 0.7365196078431373,
      "eval_f1": 0.7916666666666669,
      "eval_loss": 0.5924765467643738,
      "eval_runtime": 0.1558,
      "eval_samples_per_second": 2619.349,
      "eval_steps_per_second": 12.84,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.1937808990478516,
      "learning_rate": 4.7e-05,
      "loss": 0.5708,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7009803921568627,
      "eval_combined_score": 0.7572180248246088,
      "eval_f1": 0.8134556574923548,
      "eval_loss": 0.5934233069419861,
      "eval_runtime": 0.1333,
      "eval_samples_per_second": 3060.983,
      "eval_steps_per_second": 15.005,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.4581093788146973,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.5419,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6911764705882353,
      "eval_combined_score": 0.7436464877213022,
      "eval_f1": 0.796116504854369,
      "eval_loss": 0.5989792943000793,
      "eval_runtime": 0.1415,
      "eval_samples_per_second": 2883.208,
      "eval_steps_per_second": 14.133,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.7154335975646973,
      "learning_rate": 4.5e-05,
      "loss": 0.4984,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.678921568627451,
      "eval_combined_score": 0.7369568719506583,
      "eval_f1": 0.7949921752738655,
      "eval_loss": 0.6379741430282593,
      "eval_runtime": 0.1385,
      "eval_samples_per_second": 2945.286,
      "eval_steps_per_second": 14.438,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.438758134841919,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.4277,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6495098039215687,
      "eval_combined_score": 0.6940419220704734,
      "eval_f1": 0.7385740402193783,
      "eval_loss": 0.7019862532615662,
      "eval_runtime": 0.1332,
      "eval_samples_per_second": 3061.947,
      "eval_steps_per_second": 15.01,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.9900314807891846,
      "learning_rate": 4.3e-05,
      "loss": 0.3467,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6299019607843137,
      "eval_combined_score": 0.6808479608539686,
      "eval_f1": 0.7317939609236234,
      "eval_loss": 0.8055003881454468,
      "eval_runtime": 0.1441,
      "eval_samples_per_second": 2832.183,
      "eval_steps_per_second": 13.883,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 673316591603712.0,
      "train_loss": 0.515889158703032,
      "train_runtime": 20.6256,
      "train_samples_per_second": 8891.875,
      "train_steps_per_second": 36.363
    }
  ],
  "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": 673316591603712.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}