trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.6916264295578003,
  "best_model_checkpoint": "tiny_bert_rand_100_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": 0.9864356517791748,
      "learning_rate": 4.9e-05,
      "loss": 0.6975,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.6916264295578003,
      "eval_runtime": 0.0317,
      "eval_samples_per_second": 2239.435,
      "eval_steps_per_second": 31.541,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.9560456871986389,
      "learning_rate": 4.8e-05,
      "loss": 0.703,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.6933869123458862,
      "eval_runtime": 0.0352,
      "eval_samples_per_second": 2019.528,
      "eval_steps_per_second": 28.444,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.5352477431297302,
      "learning_rate": 4.7e-05,
      "loss": 0.6998,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7101672291755676,
      "eval_runtime": 0.0321,
      "eval_samples_per_second": 2214.39,
      "eval_steps_per_second": 31.189,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.5070664286613464,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6953,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.39436619718309857,
      "eval_loss": 0.7075814008712769,
      "eval_runtime": 0.0309,
      "eval_samples_per_second": 2299.313,
      "eval_steps_per_second": 32.385,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.66490238904953,
      "learning_rate": 4.5e-05,
      "loss": 0.6941,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4647887323943662,
      "eval_loss": 0.7023547291755676,
      "eval_runtime": 0.0406,
      "eval_samples_per_second": 1746.94,
      "eval_steps_per_second": 24.605,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.5526628494262695,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6933,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.4647887323943662,
      "eval_loss": 0.7030148506164551,
      "eval_runtime": 0.0313,
      "eval_samples_per_second": 2267.088,
      "eval_steps_per_second": 31.931,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 99911832606720.0,
      "train_loss": 0.6971826685799493,
      "train_runtime": 6.289,
      "train_samples_per_second": 5048.5,
      "train_steps_per_second": 23.851
    }
  ],
  "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": 99911832606720.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}