trainer_state.json

{
  "best_global_step": 20,
  "best_metric": 0.7161200046539307,
  "best_model_checkpoint": "bert_base_km_50_v1_rte/checkpoint-20",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 70,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.479082703590393,
      "learning_rate": 4.9e-05,
      "loss": 0.7083,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.4620938628158845,
      "eval_loss": 0.7425541281700134,
      "eval_runtime": 0.1978,
      "eval_samples_per_second": 1400.515,
      "eval_steps_per_second": 10.112,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.981489360332489,
      "learning_rate": 4.8e-05,
      "loss": 0.6628,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.4981949458483754,
      "eval_loss": 0.7161200046539307,
      "eval_runtime": 0.368,
      "eval_samples_per_second": 752.766,
      "eval_steps_per_second": 5.435,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.0692100524902344,
      "learning_rate": 4.7e-05,
      "loss": 0.6176,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.45126353790613716,
      "eval_loss": 0.7602168917655945,
      "eval_runtime": 0.1925,
      "eval_samples_per_second": 1438.78,
      "eval_steps_per_second": 10.388,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.2343248128890991,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.5475,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.49458483754512633,
      "eval_loss": 0.7970010638237,
      "eval_runtime": 0.2032,
      "eval_samples_per_second": 1362.881,
      "eval_steps_per_second": 9.84,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.726162075996399,
      "learning_rate": 4.5e-05,
      "loss": 0.4412,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4729241877256318,
      "eval_loss": 0.9100319743156433,
      "eval_runtime": 0.1982,
      "eval_samples_per_second": 1397.899,
      "eval_steps_per_second": 10.093,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 4.768690586090088,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.3187,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.45126353790613716,
      "eval_loss": 1.1240915060043335,
      "eval_runtime": 0.1928,
      "eval_samples_per_second": 1436.967,
      "eval_steps_per_second": 10.375,
      "step": 60
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.027280807495117,
      "learning_rate": 4.3e-05,
      "loss": 0.2106,
      "step": 70
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.4981949458483754,
      "eval_loss": 1.2648439407348633,
      "eval_runtime": 0.1921,
      "eval_samples_per_second": 1442.251,
      "eval_steps_per_second": 10.413,
      "step": 70
    },
    {
      "epoch": 7.0,
      "step": 70,
      "total_flos": 2293012847462400.0,
      "train_loss": 0.5009404557091849,
      "train_runtime": 43.4592,
      "train_samples_per_second": 2864.754,
      "train_steps_per_second": 11.505
    }
  ],
  "logging_steps": 1,
  "max_steps": 500,
  "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": 2293012847462400.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}