{
  "best_global_step": 14000,
  "best_metric": 0.8463211795322131,
  "best_model_checkpoint": "xnli-xlm-roberta-base/checkpoint-10000",
  "epoch": 3.0,
  "eval_steps": 1000,
  "global_step": 18408,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.16297262059973924,
      "grad_norm": 10.51728343963623,
      "learning_rate": 1.8081447963800906e-05,
      "loss": 0.9575,
      "step": 1000
    },
    {
      "epoch": 0.16297262059973924,
      "eval_accuracy": 0.7164658634538152,
      "eval_loss": 0.6793871521949768,
      "eval_macro_f1": 0.7179694771560841,
      "eval_macro_precision": 0.7238898844951175,
      "eval_macro_recall": 0.7164658634538151,
      "eval_runtime": 4.5345,
      "eval_samples_per_second": 549.125,
      "eval_steps_per_second": 8.601,
      "eval_weighted_f1": 0.7179694771560841,
      "step": 1000
    },
    {
      "epoch": 0.3259452411994785,
      "grad_norm": 6.724602699279785,
      "learning_rate": 1.8966653181529214e-05,
      "loss": 0.6483,
      "step": 2000
    },
    {
      "epoch": 0.3259452411994785,
      "eval_accuracy": 0.7967871485943775,
      "eval_loss": 0.5350644588470459,
      "eval_macro_f1": 0.7972864852344191,
      "eval_macro_precision": 0.8018465694414347,
      "eval_macro_recall": 0.7967871485943775,
      "eval_runtime": 4.5631,
      "eval_samples_per_second": 545.684,
      "eval_steps_per_second": 8.547,
      "eval_weighted_f1": 0.7972864852344191,
      "step": 2000
    },
    {
      "epoch": 0.48891786179921776,
      "grad_norm": 6.1638946533203125,
      "learning_rate": 1.781078425706525e-05,
      "loss": 0.5616,
      "step": 3000
    },
    {
      "epoch": 0.48891786179921776,
      "eval_accuracy": 0.7947791164658634,
      "eval_loss": 0.509804368019104,
      "eval_macro_f1": 0.7951670773482286,
      "eval_macro_precision": 0.8064539329297533,
      "eval_macro_recall": 0.7947791164658634,
      "eval_runtime": 4.6567,
      "eval_samples_per_second": 534.712,
      "eval_steps_per_second": 8.375,
      "eval_weighted_f1": 0.7951670773482287,
      "step": 3000
    },
    {
      "epoch": 0.651890482398957,
      "grad_norm": 5.407867908477783,
      "learning_rate": 1.6654915332601286e-05,
      "loss": 0.5226,
      "step": 4000
    },
    {
      "epoch": 0.651890482398957,
      "eval_accuracy": 0.8004016064257028,
      "eval_loss": 0.49269911646842957,
      "eval_macro_f1": 0.8015680893541245,
      "eval_macro_precision": 0.8123398526926758,
      "eval_macro_recall": 0.8004016064257028,
      "eval_runtime": 4.6437,
      "eval_samples_per_second": 536.212,
      "eval_steps_per_second": 8.399,
      "eval_weighted_f1": 0.8015680893541245,
      "step": 4000
    },
    {
      "epoch": 0.8148631029986962,
      "grad_norm": 8.23536205291748,
      "learning_rate": 1.549904640813732e-05,
      "loss": 0.4966,
      "step": 5000
    },
    {
      "epoch": 0.8148631029986962,
      "eval_accuracy": 0.8124497991967872,
      "eval_loss": 0.480131596326828,
      "eval_macro_f1": 0.8128197553956924,
      "eval_macro_precision": 0.8202401283021422,
      "eval_macro_recall": 0.8124497991967873,
      "eval_runtime": 4.5688,
      "eval_samples_per_second": 545.006,
      "eval_steps_per_second": 8.536,
      "eval_weighted_f1": 0.8128197553956925,
      "step": 5000
    },
    {
      "epoch": 0.9778357235984355,
      "grad_norm": 9.736433029174805,
      "learning_rate": 1.4343177483673352e-05,
      "loss": 0.4799,
      "step": 6000
    },
    {
      "epoch": 0.9778357235984355,
      "eval_accuracy": 0.8092369477911646,
      "eval_loss": 0.4666927456855774,
      "eval_macro_f1": 0.8108832750227813,
      "eval_macro_precision": 0.8254810350822899,
      "eval_macro_recall": 0.8092369477911646,
      "eval_runtime": 4.542,
      "eval_samples_per_second": 548.221,
      "eval_steps_per_second": 8.587,
      "eval_weighted_f1": 0.8108832750227812,
      "step": 6000
    },
    {
      "epoch": 1.1408083441981747,
      "grad_norm": 8.493484497070312,
      "learning_rate": 1.3187308559209387e-05,
      "loss": 0.4251,
      "step": 7000
    },
    {
      "epoch": 1.1408083441981747,
      "eval_accuracy": 0.8180722891566266,
      "eval_loss": 0.45433178544044495,
      "eval_macro_f1": 0.8191191956693326,
      "eval_macro_precision": 0.8269376625195844,
      "eval_macro_recall": 0.8180722891566266,
      "eval_runtime": 4.4157,
      "eval_samples_per_second": 563.894,
      "eval_steps_per_second": 8.832,
      "eval_weighted_f1": 0.8191191956693327,
      "step": 7000
    },
    {
      "epoch": 1.303780964797914,
      "grad_norm": 9.21502685546875,
      "learning_rate": 1.2031439634745421e-05,
      "loss": 0.4186,
      "step": 8000
    },
    {
      "epoch": 1.303780964797914,
      "eval_accuracy": 0.8176706827309237,
      "eval_loss": 0.46454402804374695,
      "eval_macro_f1": 0.8173393684861902,
      "eval_macro_precision": 0.8258790100555838,
      "eval_macro_recall": 0.8176706827309238,
      "eval_runtime": 4.6094,
      "eval_samples_per_second": 540.202,
      "eval_steps_per_second": 8.461,
      "eval_weighted_f1": 0.8173393684861902,
      "step": 8000
    },
    {
      "epoch": 1.4667535853976532,
      "grad_norm": 7.096667289733887,
      "learning_rate": 1.0875570710281456e-05,
      "loss": 0.4129,
      "step": 9000
    },
    {
      "epoch": 1.4667535853976532,
      "eval_accuracy": 0.8257028112449799,
      "eval_loss": 0.44398775696754456,
      "eval_macro_f1": 0.8268290877462207,
      "eval_macro_precision": 0.8358894256319416,
      "eval_macro_recall": 0.8257028112449799,
      "eval_runtime": 4.7765,
      "eval_samples_per_second": 521.303,
      "eval_steps_per_second": 8.165,
      "eval_weighted_f1": 0.8268290877462207,
      "step": 9000
    },
    {
      "epoch": 1.6297262059973925,
      "grad_norm": 6.999719142913818,
      "learning_rate": 9.719701785817489e-06,
      "loss": 0.4053,
      "step": 10000
    },
    {
      "epoch": 1.6297262059973925,
      "eval_accuracy": 0.8309236947791164,
      "eval_loss": 0.4421282708644867,
      "eval_macro_f1": 0.8314919406349293,
      "eval_macro_precision": 0.836586446573221,
      "eval_macro_recall": 0.8309236947791164,
      "eval_runtime": 4.5642,
      "eval_samples_per_second": 545.553,
      "eval_steps_per_second": 8.545,
      "eval_weighted_f1": 0.8314919406349294,
      "step": 10000
    },
    {
      "epoch": 1.7926988265971318,
      "grad_norm": 7.709704399108887,
      "learning_rate": 8.563832861353523e-06,
      "loss": 0.3971,
      "step": 11000
    },
    {
      "epoch": 1.7926988265971318,
      "eval_accuracy": 0.8248995983935743,
      "eval_loss": 0.4437423646450043,
      "eval_macro_f1": 0.825940758187211,
      "eval_macro_precision": 0.8357952618825585,
      "eval_macro_recall": 0.8248995983935744,
      "eval_runtime": 4.5323,
      "eval_samples_per_second": 549.394,
      "eval_steps_per_second": 8.605,
      "eval_weighted_f1": 0.8259407581872109,
      "step": 11000
    },
    {
      "epoch": 1.9556714471968708,
      "grad_norm": 5.534423828125,
      "learning_rate": 7.407963936889557e-06,
      "loss": 0.3947,
      "step": 12000
    },
    {
      "epoch": 1.9556714471968708,
      "eval_accuracy": 0.834136546184739,
      "eval_loss": 0.4154554307460785,
      "eval_macro_f1": 0.8350152412898564,
      "eval_macro_precision": 0.8416677147812893,
      "eval_macro_recall": 0.8341365461847389,
      "eval_runtime": 4.5044,
      "eval_samples_per_second": 552.797,
      "eval_steps_per_second": 8.658,
      "eval_weighted_f1": 0.8350152412898562,
      "step": 12000
    },
    {
      "epoch": 2.1186440677966103,
      "grad_norm": 6.122025489807129,
      "learning_rate": 6.252095012425592e-06,
      "loss": 0.3437,
      "step": 13000
    },
    {
      "epoch": 2.1186440677966103,
      "eval_accuracy": 0.8236947791164658,
      "eval_loss": 0.4527803659439087,
      "eval_macro_f1": 0.8249615968184654,
      "eval_macro_precision": 0.8339980704173974,
      "eval_macro_recall": 0.8236947791164658,
      "eval_runtime": 4.5511,
      "eval_samples_per_second": 547.117,
      "eval_steps_per_second": 8.569,
      "eval_weighted_f1": 0.8249615968184654,
      "step": 13000
    },
    {
      "epoch": 2.2816166883963493,
      "grad_norm": 10.657154083251953,
      "learning_rate": 5.096226087961625e-06,
      "loss": 0.3296,
      "step": 14000
    },
    {
      "epoch": 2.2816166883963493,
      "eval_accuracy": 0.8461847389558232,
      "eval_loss": 0.4210074543952942,
      "eval_macro_f1": 0.8463211795322131,
      "eval_macro_precision": 0.8503350839590045,
      "eval_macro_recall": 0.8461847389558232,
      "eval_runtime": 4.5213,
      "eval_samples_per_second": 550.73,
      "eval_steps_per_second": 8.626,
      "eval_weighted_f1": 0.846321179532213,
      "step": 14000
    },
    {
      "epoch": 2.444589308996089,
      "grad_norm": 7.961305141448975,
      "learning_rate": 3.940357163497659e-06,
      "loss": 0.3219,
      "step": 15000
    },
    {
      "epoch": 2.444589308996089,
      "eval_accuracy": 0.8337349397590361,
      "eval_loss": 0.4486163258552551,
      "eval_macro_f1": 0.8349888002365645,
      "eval_macro_precision": 0.8443748814531276,
      "eval_macro_recall": 0.8337349397590362,
      "eval_runtime": 4.5481,
      "eval_samples_per_second": 547.482,
      "eval_steps_per_second": 8.575,
      "eval_weighted_f1": 0.8349888002365645,
      "step": 15000
    },
    {
      "epoch": 2.607561929595828,
      "grad_norm": 7.545483589172363,
      "learning_rate": 2.7844882390336937e-06,
      "loss": 0.3241,
      "step": 16000
    },
    {
      "epoch": 2.607561929595828,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 0.46160411834716797,
      "eval_macro_f1": 0.8337624688566216,
      "eval_macro_precision": 0.8404933913764827,
      "eval_macro_recall": 0.8333333333333334,
      "eval_runtime": 4.5462,
      "eval_samples_per_second": 547.708,
      "eval_steps_per_second": 8.579,
      "eval_weighted_f1": 0.8337624688566218,
      "step": 16000
    },
    {
      "epoch": 2.770534550195567,
      "grad_norm": 7.061276435852051,
      "learning_rate": 1.6286193145697278e-06,
      "loss": 0.326,
      "step": 17000
    },
    {
      "epoch": 2.770534550195567,
      "eval_accuracy": 0.8345381526104417,
      "eval_loss": 0.4480307400226593,
      "eval_macro_f1": 0.8351454373443175,
      "eval_macro_precision": 0.8421626579409135,
      "eval_macro_recall": 0.8345381526104418,
      "eval_runtime": 4.5496,
      "eval_samples_per_second": 547.299,
      "eval_steps_per_second": 8.572,
      "eval_weighted_f1": 0.8351454373443173,
      "step": 17000
    },
    {
      "epoch": 2.9335071707953064,
      "grad_norm": 6.060895919799805,
      "learning_rate": 4.72750390105762e-07,
      "loss": 0.3128,
      "step": 18000
    },
    {
      "epoch": 2.9335071707953064,
      "eval_accuracy": 0.834136546184739,
      "eval_loss": 0.45053839683532715,
      "eval_macro_f1": 0.8350207810106501,
      "eval_macro_precision": 0.8411309111788473,
      "eval_macro_recall": 0.8341365461847391,
      "eval_runtime": 4.5608,
      "eval_samples_per_second": 545.951,
      "eval_steps_per_second": 8.551,
      "eval_weighted_f1": 0.8350207810106502,
      "step": 18000
    },
    {
      "epoch": 3.0,
      "step": 18408,
      "total_flos": 3.099754961058632e+17,
      "train_loss": 0.4458606608895207,
      "train_runtime": 5493.4622,
      "train_samples_per_second": 214.456,
      "train_steps_per_second": 3.351
    }
  ],
  "logging_steps": 1000,
  "max_steps": 18408,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 10000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 3.099754961058632e+17,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}