trainer_state.json

{
  "best_metric": 1.0895687341690063,
  "best_model_checkpoint": "hBERTv1_new_pretrain_w_init__mnli/checkpoint-15340",
  "epoch": 10.0,
  "global_step": 30680,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 3.9200000000000004e-05,
      "loss": 1.0911,
      "step": 3068
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.33397860417727965,
      "eval_loss": 1.094429612159729,
      "eval_runtime": 17.6447,
      "eval_samples_per_second": 556.258,
      "eval_steps_per_second": 4.364,
      "step": 3068
    },
    {
      "epoch": 2.0,
      "learning_rate": 3.8400000000000005e-05,
      "loss": 1.0984,
      "step": 6136
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.31818644931227713,
      "eval_loss": 1.096695065498352,
      "eval_runtime": 17.6375,
      "eval_samples_per_second": 556.485,
      "eval_steps_per_second": 4.366,
      "step": 6136
    },
    {
      "epoch": 3.0,
      "learning_rate": 3.76e-05,
      "loss": 1.0988,
      "step": 9204
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.3544574630667346,
      "eval_loss": 1.0961673259735107,
      "eval_runtime": 17.6642,
      "eval_samples_per_second": 555.643,
      "eval_steps_per_second": 4.359,
      "step": 9204
    },
    {
      "epoch": 4.0,
      "learning_rate": 3.680000000000001e-05,
      "loss": 1.0989,
      "step": 12272
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.31818644931227713,
      "eval_loss": 1.097014307975769,
      "eval_runtime": 17.6481,
      "eval_samples_per_second": 556.152,
      "eval_steps_per_second": 4.363,
      "step": 12272
    },
    {
      "epoch": 5.0,
      "learning_rate": 3.6e-05,
      "loss": 1.0986,
      "step": 15340
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.3515028018339277,
      "eval_loss": 1.0895687341690063,
      "eval_runtime": 17.6444,
      "eval_samples_per_second": 556.267,
      "eval_steps_per_second": 4.364,
      "step": 15340
    },
    {
      "epoch": 6.0,
      "learning_rate": 3.52e-05,
      "loss": 1.0984,
      "step": 18408
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.3544574630667346,
      "eval_loss": 1.098585605621338,
      "eval_runtime": 17.6522,
      "eval_samples_per_second": 556.022,
      "eval_steps_per_second": 4.362,
      "step": 18408
    },
    {
      "epoch": 7.0,
      "learning_rate": 3.44e-05,
      "loss": 1.0992,
      "step": 21476
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.3273560876209883,
      "eval_loss": 1.0993512868881226,
      "eval_runtime": 17.6446,
      "eval_samples_per_second": 556.262,
      "eval_steps_per_second": 4.364,
      "step": 21476
    },
    {
      "epoch": 8.0,
      "learning_rate": 3.3600000000000004e-05,
      "loss": 1.0993,
      "step": 24544
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.3544574630667346,
      "eval_loss": 1.099851131439209,
      "eval_runtime": 17.6553,
      "eval_samples_per_second": 555.922,
      "eval_steps_per_second": 4.361,
      "step": 24544
    },
    {
      "epoch": 9.0,
      "learning_rate": 3.28e-05,
      "loss": 1.099,
      "step": 27612
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.31818644931227713,
      "eval_loss": 1.0964539051055908,
      "eval_runtime": 17.6517,
      "eval_samples_per_second": 556.037,
      "eval_steps_per_second": 4.362,
      "step": 27612
    },
    {
      "epoch": 10.0,
      "learning_rate": 3.2000000000000005e-05,
      "loss": 1.0991,
      "step": 30680
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.31818644931227713,
      "eval_loss": 1.1004576683044434,
      "eval_runtime": 17.6478,
      "eval_samples_per_second": 556.161,
      "eval_steps_per_second": 4.363,
      "step": 30680
    },
    {
      "epoch": 10.0,
      "step": 30680,
      "total_flos": 5.81583626085335e+17,
      "train_loss": 1.09808146689364,
      "train_runtime": 20906.5595,
      "train_samples_per_second": 939.184,
      "train_steps_per_second": 7.337
    }
  ],
  "max_steps": 153400,
  "num_train_epochs": 50,
  "total_flos": 5.81583626085335e+17,
  "trial_name": null,
  "trial_params": null
}