{
  "best_global_step": 2300,
  "best_metric": 0.8162891986062718,
  "best_model_checkpoint": "./hpo_microsoft_mdeberta-v3-base/trial_10/checkpoint-2300",
  "epoch": 4.307116104868914,
  "eval_steps": 100,
  "global_step": 2300,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.18726591760299627,
      "grad_norm": 3.1831746101379395,
      "learning_rate": 8.172021938371146e-07,
      "loss": 0.7259,
      "step": 100
    },
    {
      "epoch": 0.18726591760299627,
      "eval_accuracy": 0.36936936936936937,
      "eval_f1": 0.26973684210526316,
      "eval_loss": 0.7737258076667786,
      "eval_precision": 0.18468468468468469,
      "eval_recall": 0.5,
      "eval_runtime": 0.9378,
      "eval_samples_per_second": 236.735,
      "eval_steps_per_second": 29.859,
      "step": 100
    },
    {
      "epoch": 0.37453183520599254,
      "grad_norm": 10.540749549865723,
      "learning_rate": 1.6344043876742292e-06,
      "loss": 0.6787,
      "step": 200
    },
    {
      "epoch": 0.37453183520599254,
      "eval_accuracy": 0.6981981981981982,
      "eval_f1": 0.6232809057062533,
      "eval_loss": 0.6247357130050659,
      "eval_precision": 0.6922921439159143,
      "eval_recall": 0.6243031358885017,
      "eval_runtime": 0.9392,
      "eval_samples_per_second": 236.362,
      "eval_steps_per_second": 29.811,
      "step": 200
    },
    {
      "epoch": 0.5617977528089888,
      "grad_norm": 11.673765182495117,
      "learning_rate": 2.4433520138968275e-06,
      "loss": 0.6015,
      "step": 300
    },
    {
      "epoch": 0.5617977528089888,
      "eval_accuracy": 0.7207207207207207,
      "eval_f1": 0.687891156462585,
      "eval_loss": 0.5849120616912842,
      "eval_precision": 0.7001527883880825,
      "eval_recall": 0.6825783972125437,
      "eval_runtime": 0.9386,
      "eval_samples_per_second": 236.521,
      "eval_steps_per_second": 29.831,
      "step": 300
    },
    {
      "epoch": 0.7490636704119851,
      "grad_norm": 13.403273582458496,
      "learning_rate": 3.2688087753484585e-06,
      "loss": 0.5959,
      "step": 400
    },
    {
      "epoch": 0.7490636704119851,
      "eval_accuracy": 0.7342342342342343,
      "eval_f1": 0.6956996491880213,
      "eval_loss": 0.5658894181251526,
      "eval_precision": 0.7200386925974951,
      "eval_recall": 0.6882404181184669,
      "eval_runtime": 0.94,
      "eval_samples_per_second": 236.168,
      "eval_steps_per_second": 29.787,
      "step": 400
    },
    {
      "epoch": 0.9363295880149812,
      "grad_norm": 9.928894996643066,
      "learning_rate": 4.09426553680009e-06,
      "loss": 0.5559,
      "step": 500
    },
    {
      "epoch": 0.9363295880149812,
      "eval_accuracy": 0.7432432432432432,
      "eval_f1": 0.6940744143316491,
      "eval_loss": 0.5582005381584167,
      "eval_precision": 0.7438940488476092,
      "eval_recall": 0.6852787456445993,
      "eval_runtime": 0.9379,
      "eval_samples_per_second": 236.71,
      "eval_steps_per_second": 29.855,
      "step": 500
    },
    {
      "epoch": 1.1235955056179776,
      "grad_norm": 4.285970211029053,
      "learning_rate": 4.91972229825172e-06,
      "loss": 0.4909,
      "step": 600
    },
    {
      "epoch": 1.1235955056179776,
      "eval_accuracy": 0.7612612612612613,
      "eval_f1": 0.7155428764838141,
      "eval_loss": 0.5819087624549866,
      "eval_precision": 0.7693498452012384,
      "eval_recall": 0.7046167247386759,
      "eval_runtime": 0.9385,
      "eval_samples_per_second": 236.556,
      "eval_steps_per_second": 29.836,
      "step": 600
    },
    {
      "epoch": 1.3108614232209739,
      "grad_norm": 12.254231452941895,
      "learning_rate": 5.745179059703351e-06,
      "loss": 0.4816,
      "step": 700
    },
    {
      "epoch": 1.3108614232209739,
      "eval_accuracy": 0.7657657657657657,
      "eval_f1": 0.7445339470655927,
      "eval_loss": 0.5305711627006531,
      "eval_precision": 0.7493691420331651,
      "eval_recall": 0.74102787456446,
      "eval_runtime": 0.9437,
      "eval_samples_per_second": 235.245,
      "eval_steps_per_second": 29.671,
      "step": 700
    },
    {
      "epoch": 1.4981273408239701,
      "grad_norm": 12.296418190002441,
      "learning_rate": 6.570635821154982e-06,
      "loss": 0.4404,
      "step": 800
    },
    {
      "epoch": 1.4981273408239701,
      "eval_accuracy": 0.7702702702702703,
      "eval_f1": 0.7329402052128788,
      "eval_loss": 0.5971503257751465,
      "eval_precision": 0.7708133971291866,
      "eval_recall": 0.7218641114982578,
      "eval_runtime": 0.9362,
      "eval_samples_per_second": 237.123,
      "eval_steps_per_second": 29.907,
      "step": 800
    },
    {
      "epoch": 1.6853932584269664,
      "grad_norm": 32.506744384765625,
      "learning_rate": 7.396092582606613e-06,
      "loss": 0.4179,
      "step": 900
    },
    {
      "epoch": 1.6853932584269664,
      "eval_accuracy": 0.7567567567567568,
      "eval_f1": 0.75159552424368,
      "eval_loss": 0.6150170564651489,
      "eval_precision": 0.753411306042885,
      "eval_recall": 0.7717770034843205,
      "eval_runtime": 0.9387,
      "eval_samples_per_second": 236.485,
      "eval_steps_per_second": 29.827,
      "step": 900
    },
    {
      "epoch": 1.8726591760299627,
      "grad_norm": 21.407085418701172,
      "learning_rate": 7.76395084943973e-06,
      "loss": 0.4303,
      "step": 1000
    },
    {
      "epoch": 1.8726591760299627,
      "eval_accuracy": 0.7882882882882883,
      "eval_f1": 0.7721983276204616,
      "eval_loss": 0.5324995517730713,
      "eval_precision": 0.7729183083793012,
      "eval_recall": 0.7715156794425087,
      "eval_runtime": 0.9409,
      "eval_samples_per_second": 235.951,
      "eval_steps_per_second": 29.76,
      "step": 1000
    },
    {
      "epoch": 2.059925093632959,
      "grad_norm": 46.14255905151367,
      "learning_rate": 7.653274429633035e-06,
      "loss": 0.3717,
      "step": 1100
    },
    {
      "epoch": 2.059925093632959,
      "eval_accuracy": 0.7657657657657657,
      "eval_f1": 0.7614087301587301,
      "eval_loss": 0.7195941805839539,
      "eval_precision": 0.7646103896103896,
      "eval_recall": 0.7839721254355401,
      "eval_runtime": 0.9377,
      "eval_samples_per_second": 236.755,
      "eval_steps_per_second": 29.861,
      "step": 1100
    },
    {
      "epoch": 2.247191011235955,
      "grad_norm": 9.187861442565918,
      "learning_rate": 7.5425980098263385e-06,
      "loss": 0.3651,
      "step": 1200
    },
    {
      "epoch": 2.247191011235955,
      "eval_accuracy": 0.7702702702702703,
      "eval_f1": 0.7604265854123025,
      "eval_loss": 0.5944370627403259,
      "eval_precision": 0.7567757977621219,
      "eval_recall": 0.7698606271777004,
      "eval_runtime": 0.9386,
      "eval_samples_per_second": 236.528,
      "eval_steps_per_second": 29.832,
      "step": 1200
    },
    {
      "epoch": 2.4344569288389515,
      "grad_norm": 55.409934997558594,
      "learning_rate": 7.431921590019642e-06,
      "loss": 0.3586,
      "step": 1300
    },
    {
      "epoch": 2.4344569288389515,
      "eval_accuracy": 0.7792792792792793,
      "eval_f1": 0.7736532179196405,
      "eval_loss": 0.8182331919670105,
      "eval_precision": 0.7730158730158729,
      "eval_recall": 0.7921602787456445,
      "eval_runtime": 0.937,
      "eval_samples_per_second": 236.917,
      "eval_steps_per_second": 29.881,
      "step": 1300
    },
    {
      "epoch": 2.6217228464419478,
      "grad_norm": 4.450952053070068,
      "learning_rate": 7.321245170212946e-06,
      "loss": 0.3967,
      "step": 1400
    },
    {
      "epoch": 2.6217228464419478,
      "eval_accuracy": 0.7567567567567568,
      "eval_f1": 0.75159552424368,
      "eval_loss": 0.6650905013084412,
      "eval_precision": 0.753411306042885,
      "eval_recall": 0.7717770034843205,
      "eval_runtime": 0.938,
      "eval_samples_per_second": 236.671,
      "eval_steps_per_second": 29.85,
      "step": 1400
    },
    {
      "epoch": 2.808988764044944,
      "grad_norm": 0.43200916051864624,
      "learning_rate": 7.210568750406251e-06,
      "loss": 0.3031,
      "step": 1500
    },
    {
      "epoch": 2.808988764044944,
      "eval_accuracy": 0.7837837837837838,
      "eval_f1": 0.7667454688731284,
      "eval_loss": 0.6875036954879761,
      "eval_precision": 0.7682218309859155,
      "eval_recall": 0.7654181184668989,
      "eval_runtime": 0.9404,
      "eval_samples_per_second": 236.06,
      "eval_steps_per_second": 29.773,
      "step": 1500
    },
    {
      "epoch": 2.9962546816479403,
      "grad_norm": 53.977630615234375,
      "learning_rate": 7.099892330599554e-06,
      "loss": 0.3883,
      "step": 1600
    },
    {
      "epoch": 2.9962546816479403,
      "eval_accuracy": 0.7972972972972973,
      "eval_f1": 0.7795431976166831,
      "eval_loss": 0.7121178507804871,
      "eval_precision": 0.783922973578146,
      "eval_recall": 0.7761324041811846,
      "eval_runtime": 0.9404,
      "eval_samples_per_second": 236.067,
      "eval_steps_per_second": 29.774,
      "step": 1600
    },
    {
      "epoch": 3.1835205992509366,
      "grad_norm": 6.338677406311035,
      "learning_rate": 6.989215910792859e-06,
      "loss": 0.2718,
      "step": 1700
    },
    {
      "epoch": 3.1835205992509366,
      "eval_accuracy": 0.8018018018018018,
      "eval_f1": 0.7872822299651567,
      "eval_loss": 0.8387073278427124,
      "eval_precision": 0.7872822299651567,
      "eval_recall": 0.7872822299651567,
      "eval_runtime": 0.9424,
      "eval_samples_per_second": 235.568,
      "eval_steps_per_second": 29.711,
      "step": 1700
    },
    {
      "epoch": 3.370786516853933,
      "grad_norm": 0.23445236682891846,
      "learning_rate": 6.87964625518423e-06,
      "loss": 0.259,
      "step": 1800
    },
    {
      "epoch": 3.370786516853933,
      "eval_accuracy": 0.7792792792792793,
      "eval_f1": 0.7472759798341193,
      "eval_loss": 1.0221707820892334,
      "eval_precision": 0.7765502494654313,
      "eval_recall": 0.7365853658536585,
      "eval_runtime": 0.9403,
      "eval_samples_per_second": 236.09,
      "eval_steps_per_second": 29.777,
      "step": 1800
    },
    {
      "epoch": 3.558052434456929,
      "grad_norm": 5.759720325469971,
      "learning_rate": 6.768969835377533e-06,
      "loss": 0.2144,
      "step": 1900
    },
    {
      "epoch": 3.558052434456929,
      "eval_accuracy": 0.7837837837837838,
      "eval_f1": 0.7765288146967537,
      "eval_loss": 0.9993805885314941,
      "eval_precision": 0.7735245901639345,
      "eval_recall": 0.7906794425087107,
      "eval_runtime": 0.9382,
      "eval_samples_per_second": 236.633,
      "eval_steps_per_second": 29.846,
      "step": 1900
    },
    {
      "epoch": 3.7453183520599254,
      "grad_norm": 22.900344848632812,
      "learning_rate": 6.658293415570837e-06,
      "loss": 0.2209,
      "step": 2000
    },
    {
      "epoch": 3.7453183520599254,
      "eval_accuracy": 0.7792792792792793,
      "eval_f1": 0.7706756614314325,
      "eval_loss": 1.0168108940124512,
      "eval_precision": 0.7670515463917525,
      "eval_recall": 0.78205574912892,
      "eval_runtime": 0.9394,
      "eval_samples_per_second": 236.31,
      "eval_steps_per_second": 29.805,
      "step": 2000
    },
    {
      "epoch": 3.932584269662921,
      "grad_norm": 0.10348391532897949,
      "learning_rate": 6.54761699576414e-06,
      "loss": 0.2178,
      "step": 2100
    },
    {
      "epoch": 3.932584269662921,
      "eval_accuracy": 0.8198198198198198,
      "eval_f1": 0.8066202090592334,
      "eval_loss": 0.9587396383285522,
      "eval_precision": 0.8066202090592334,
      "eval_recall": 0.8066202090592334,
      "eval_runtime": 0.9442,
      "eval_samples_per_second": 235.117,
      "eval_steps_per_second": 29.654,
      "step": 2100
    },
    {
      "epoch": 4.119850187265918,
      "grad_norm": 0.5545433759689331,
      "learning_rate": 6.436940575957445e-06,
      "loss": 0.2408,
      "step": 2200
    },
    {
      "epoch": 4.119850187265918,
      "eval_accuracy": 0.8243243243243243,
      "eval_f1": 0.8145363408521302,
      "eval_loss": 0.9260944724082947,
      "eval_precision": 0.810636140914083,
      "eval_recall": 0.8202961672473867,
      "eval_runtime": 0.9409,
      "eval_samples_per_second": 235.953,
      "eval_steps_per_second": 29.76,
      "step": 2200
    },
    {
      "epoch": 4.307116104868914,
      "grad_norm": 0.34004443883895874,
      "learning_rate": 6.326264156150748e-06,
      "loss": 0.1416,
      "step": 2300
    },
    {
      "epoch": 4.307116104868914,
      "eval_accuracy": 0.8288288288288288,
      "eval_f1": 0.8162891986062718,
      "eval_loss": 1.0393856763839722,
      "eval_precision": 0.8162891986062718,
      "eval_recall": 0.8162891986062718,
      "eval_runtime": 0.9412,
      "eval_samples_per_second": 235.861,
      "eval_steps_per_second": 29.748,
      "step": 2300
    }
  ],
  "logging_steps": 100,
  "max_steps": 8010,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "save_steps": 100,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 4,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 4838172851429376.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}