trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.755666971206665,
  "best_model_checkpoint": "bert_base_km_5_v2_wnli/checkpoint-3",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.9213811159133911,
      "learning_rate": 4.9e-05,
      "loss": 0.7687,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.2676056338028169,
      "eval_loss": 0.755666971206665,
      "eval_runtime": 0.0609,
      "eval_samples_per_second": 1165.691,
      "eval_steps_per_second": 16.418,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.1780509948730469,
      "learning_rate": 4.8e-05,
      "loss": 0.7017,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.2112676056338028,
      "eval_loss": 0.775858461856842,
      "eval_runtime": 0.0582,
      "eval_samples_per_second": 1219.155,
      "eval_steps_per_second": 17.171,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.1349819898605347,
      "learning_rate": 4.7e-05,
      "loss": 0.6899,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.16901408450704225,
      "eval_loss": 0.7978102564811707,
      "eval_runtime": 0.068,
      "eval_samples_per_second": 1043.941,
      "eval_steps_per_second": 14.703,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.666037917137146,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6828,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.15492957746478872,
      "eval_loss": 0.8442451357841492,
      "eval_runtime": 0.0684,
      "eval_samples_per_second": 1038.527,
      "eval_steps_per_second": 14.627,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.3007279634475708,
      "learning_rate": 4.5e-05,
      "loss": 0.6741,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.15492957746478872,
      "eval_loss": 0.899730384349823,
      "eval_runtime": 0.0666,
      "eval_samples_per_second": 1065.859,
      "eval_steps_per_second": 15.012,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.3727164268493652,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6788,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.15492957746478872,
      "eval_loss": 0.9396181702613831,
      "eval_runtime": 0.0586,
      "eval_samples_per_second": 1211.206,
      "eval_steps_per_second": 17.059,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 501226560460800.0,
      "train_loss": 0.6993648078706529,
      "train_runtime": 18.1786,
      "train_samples_per_second": 1746.558,
      "train_steps_per_second": 8.251
    }
  ],
  "logging_steps": 1,
  "max_steps": 150,
  "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": 501226560460800.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}