trainer_state.json

{
  "best_global_step": 264,
  "best_metric": 0.4648650288581848,
  "best_model_checkpoint": "tiny_bert_km_20_v2_sst2/checkpoint-264",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 1584,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 7.1825785636901855,
      "learning_rate": 4.9e-05,
      "loss": 0.448,
      "step": 264
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.783256880733945,
      "eval_loss": 0.4648650288581848,
      "eval_runtime": 0.3683,
      "eval_samples_per_second": 2367.809,
      "eval_steps_per_second": 10.862,
      "step": 264
    },
    {
      "epoch": 2.0,
      "grad_norm": 11.759496688842773,
      "learning_rate": 4.8e-05,
      "loss": 0.249,
      "step": 528
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7889908256880734,
      "eval_loss": 0.5035269260406494,
      "eval_runtime": 0.3562,
      "eval_samples_per_second": 2448.119,
      "eval_steps_per_second": 11.23,
      "step": 528
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.369842052459717,
      "learning_rate": 4.7e-05,
      "loss": 0.1908,
      "step": 792
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.805045871559633,
      "eval_loss": 0.5272838473320007,
      "eval_runtime": 0.2646,
      "eval_samples_per_second": 3295.169,
      "eval_steps_per_second": 15.115,
      "step": 792
    },
    {
      "epoch": 4.0,
      "grad_norm": 8.15778636932373,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.1571,
      "step": 1056
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7844036697247706,
      "eval_loss": 0.5577961802482605,
      "eval_runtime": 0.2601,
      "eval_samples_per_second": 3351.916,
      "eval_steps_per_second": 15.376,
      "step": 1056
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.027233600616455,
      "learning_rate": 4.5e-05,
      "loss": 0.1298,
      "step": 1320
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7855504587155964,
      "eval_loss": 0.6700541377067566,
      "eval_runtime": 0.2805,
      "eval_samples_per_second": 3108.32,
      "eval_steps_per_second": 14.258,
      "step": 1320
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.197101593017578,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1096,
      "step": 1584
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7694954128440367,
      "eval_loss": 0.7885540127754211,
      "eval_runtime": 0.2684,
      "eval_samples_per_second": 3248.291,
      "eval_steps_per_second": 14.9,
      "step": 1584
    },
    {
      "epoch": 6.0,
      "step": 1584,
      "total_flos": 1.0596790573590528e+16,
      "train_loss": 0.21403723894947707,
      "train_runtime": 222.2321,
      "train_samples_per_second": 15152.854,
      "train_steps_per_second": 59.397
    }
  ],
  "logging_steps": 1,
  "max_steps": 13200,
  "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": 1.0596790573590528e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}