trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5913755893707275,
  "best_model_checkpoint": "distilbert_rand_20_v2_mrpc/checkpoint-30",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 105,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.4107073545455933,
      "learning_rate": 4.9e-05,
      "loss": 0.6287,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6740196078431373,
      "eval_combined_score": 0.727454614465226,
      "eval_f1": 0.7808896210873146,
      "eval_loss": 0.6147627830505371,
      "eval_runtime": 0.1825,
      "eval_samples_per_second": 2235.542,
      "eval_steps_per_second": 10.959,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.8617336750030518,
      "learning_rate": 4.8e-05,
      "loss": 0.579,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6813725490196079,
      "eval_combined_score": 0.7351667939903235,
      "eval_f1": 0.788961038961039,
      "eval_loss": 0.5913755893707275,
      "eval_runtime": 0.257,
      "eval_samples_per_second": 1587.518,
      "eval_steps_per_second": 7.782,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.4047906398773193,
      "learning_rate": 4.7e-05,
      "loss": 0.5138,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.696078431372549,
      "eval_combined_score": 0.7486802413273002,
      "eval_f1": 0.8012820512820513,
      "eval_loss": 0.620290994644165,
      "eval_runtime": 0.1934,
      "eval_samples_per_second": 2109.66,
      "eval_steps_per_second": 10.341,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.039705753326416,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4172,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6764705882352942,
      "eval_combined_score": 0.7282352941176471,
      "eval_f1": 0.78,
      "eval_loss": 0.7218132615089417,
      "eval_runtime": 0.1838,
      "eval_samples_per_second": 2220.369,
      "eval_steps_per_second": 10.884,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.445235252380371,
      "learning_rate": 4.5e-05,
      "loss": 0.2715,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6127450980392157,
      "eval_combined_score": 0.6508607379959859,
      "eval_f1": 0.6889763779527559,
      "eval_loss": 1.0093188285827637,
      "eval_runtime": 0.1838,
      "eval_samples_per_second": 2219.565,
      "eval_steps_per_second": 10.88,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.9740326404571533,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1628,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6102941176470589,
      "eval_combined_score": 0.6519678680720844,
      "eval_f1": 0.6936416184971098,
      "eval_loss": 1.115601897239685,
      "eval_runtime": 0.1847,
      "eval_samples_per_second": 2209.278,
      "eval_steps_per_second": 10.83,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 6.120091438293457,
      "learning_rate": 4.3e-05,
      "loss": 0.0908,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6495098039215687,
      "eval_combined_score": 0.7008380908689299,
      "eval_f1": 0.7521663778162911,
      "eval_loss": 1.407080054283142,
      "eval_runtime": 0.2105,
      "eval_samples_per_second": 1938.306,
      "eval_steps_per_second": 9.501,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.3805481899352301,
      "train_runtime": 34.6829,
      "train_samples_per_second": 5287.907,
      "train_steps_per_second": 21.624
    }
  ],
  "logging_steps": 1,
  "max_steps": 750,
  "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": 1700616463945728.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}