trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5883752703666687,
  "best_model_checkpoint": "distilbert_rand_10_v1_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": 1.2386474609375,
      "learning_rate": 4.9e-05,
      "loss": 0.6309,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6936274509803921,
      "eval_combined_score": 0.7474496555378909,
      "eval_f1": 0.8012718600953895,
      "eval_loss": 0.5975795984268188,
      "eval_runtime": 0.1902,
      "eval_samples_per_second": 2145.384,
      "eval_steps_per_second": 10.517,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.6848299503326416,
      "learning_rate": 4.8e-05,
      "loss": 0.5756,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_combined_score": 0.7207505518763797,
      "eval_f1": 0.7748344370860927,
      "eval_loss": 0.5883752703666687,
      "eval_runtime": 0.2083,
      "eval_samples_per_second": 1958.546,
      "eval_steps_per_second": 9.601,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.6218466758728027,
      "learning_rate": 4.7e-05,
      "loss": 0.507,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6985294117647058,
      "eval_combined_score": 0.7476118133203695,
      "eval_f1": 0.7966942148760331,
      "eval_loss": 0.616435706615448,
      "eval_runtime": 0.2045,
      "eval_samples_per_second": 1994.657,
      "eval_steps_per_second": 9.778,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.5884528160095215,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4021,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6838235294117647,
      "eval_combined_score": 0.7285549984492918,
      "eval_f1": 0.773286467486819,
      "eval_loss": 0.7436296343803406,
      "eval_runtime": 0.1815,
      "eval_samples_per_second": 2247.645,
      "eval_steps_per_second": 11.018,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.304786682128906,
      "learning_rate": 4.5e-05,
      "loss": 0.2627,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6127450980392157,
      "eval_combined_score": 0.6567513368983957,
      "eval_f1": 0.7007575757575758,
      "eval_loss": 0.9783936142921448,
      "eval_runtime": 0.1818,
      "eval_samples_per_second": 2243.949,
      "eval_steps_per_second": 11.0,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 5.823033809661865,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1522,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6397058823529411,
      "eval_combined_score": 0.6824697636063771,
      "eval_f1": 0.7252336448598131,
      "eval_loss": 1.2597531080245972,
      "eval_runtime": 0.1808,
      "eval_samples_per_second": 2257.176,
      "eval_steps_per_second": 11.065,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 3.5266897678375244,
      "learning_rate": 4.3e-05,
      "loss": 0.0925,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6544117647058824,
      "eval_combined_score": 0.7050221735141197,
      "eval_f1": 0.755632582322357,
      "eval_loss": 1.4431391954421997,
      "eval_runtime": 0.1779,
      "eval_samples_per_second": 2293.853,
      "eval_steps_per_second": 11.244,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.37471781231108164,
      "train_runtime": 33.5374,
      "train_samples_per_second": 5468.518,
      "train_steps_per_second": 22.363
    }
  ],
  "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
}