trainer_state.json

{
  "best_global_step": 23,
  "best_metric": 2.332362413406372,
  "best_model_checkpoint": "tiny_bert_rand_20_v2_stsb/checkpoint-23",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 138,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 5.404026508331299,
      "learning_rate": 4.9e-05,
      "loss": 3.3593,
      "step": 23
    },
    {
      "epoch": 1.0,
      "eval_combined_score": 0.08981603342215606,
      "eval_loss": 2.332362413406372,
      "eval_pearson": 0.096510271927189,
      "eval_runtime": 0.4526,
      "eval_samples_per_second": 3314.52,
      "eval_spearmanr": 0.08312179491712313,
      "eval_steps_per_second": 13.258,
      "step": 23
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.6226043701171875,
      "learning_rate": 4.8e-05,
      "loss": 2.0103,
      "step": 46
    },
    {
      "epoch": 2.0,
      "eval_combined_score": 0.10747806768581006,
      "eval_loss": 2.6266026496887207,
      "eval_pearson": 0.11564175443465152,
      "eval_runtime": 0.4548,
      "eval_samples_per_second": 3297.834,
      "eval_spearmanr": 0.09931438093696858,
      "eval_steps_per_second": 13.191,
      "step": 46
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.7208993434906006,
      "learning_rate": 4.7e-05,
      "loss": 1.8753,
      "step": 69
    },
    {
      "epoch": 3.0,
      "eval_combined_score": 0.14675188897568292,
      "eval_loss": 2.426100015640259,
      "eval_pearson": 0.15510937445799697,
      "eval_runtime": 0.5263,
      "eval_samples_per_second": 2850.152,
      "eval_spearmanr": 0.13839440349336884,
      "eval_steps_per_second": 11.401,
      "step": 69
    },
    {
      "epoch": 4.0,
      "grad_norm": 6.908268928527832,
      "learning_rate": 4.600000000000001e-05,
      "loss": 1.6951,
      "step": 92
    },
    {
      "epoch": 4.0,
      "eval_combined_score": 0.2058659455787042,
      "eval_loss": 2.4686241149902344,
      "eval_pearson": 0.2094385416597008,
      "eval_runtime": 0.4474,
      "eval_samples_per_second": 3352.645,
      "eval_spearmanr": 0.20229334949770755,
      "eval_steps_per_second": 13.411,
      "step": 92
    },
    {
      "epoch": 5.0,
      "grad_norm": 8.469391822814941,
      "learning_rate": 4.5e-05,
      "loss": 1.431,
      "step": 115
    },
    {
      "epoch": 5.0,
      "eval_combined_score": 0.2136916379006699,
      "eval_loss": 2.4816925525665283,
      "eval_pearson": 0.2170627870847877,
      "eval_runtime": 0.4639,
      "eval_samples_per_second": 3233.422,
      "eval_spearmanr": 0.21032048871655207,
      "eval_steps_per_second": 12.934,
      "step": 115
    },
    {
      "epoch": 6.0,
      "grad_norm": 8.159451484680176,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 1.3014,
      "step": 138
    },
    {
      "epoch": 6.0,
      "eval_combined_score": 0.22419636391194636,
      "eval_loss": 2.6271088123321533,
      "eval_pearson": 0.2217179151568541,
      "eval_runtime": 0.4815,
      "eval_samples_per_second": 3115.335,
      "eval_spearmanr": 0.22667481266703865,
      "eval_steps_per_second": 12.461,
      "step": 138
    },
    {
      "epoch": 6.0,
      "step": 138,
      "total_flos": 904528923227136.0,
      "train_loss": 1.9453966168389805,
      "train_runtime": 25.9633,
      "train_samples_per_second": 11071.387,
      "train_steps_per_second": 44.293
    }
  ],
  "logging_steps": 1,
  "max_steps": 1150,
  "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": 904528923227136.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}