{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 1000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "grad_norm": 0.4471716284751892,
      "learning_rate": 1.8e-05,
      "loss": 1.287,
      "step": 10
    },
    {
      "grad_norm": 0.20870500802993774,
      "learning_rate": 3.8e-05,
      "loss": 1.202,
      "step": 20
    },
    {
      "grad_norm": 0.15201318264007568,
      "learning_rate": 5.8e-05,
      "loss": 1.2002,
      "step": 30
    },
    {
      "grad_norm": 0.16838377714157104,
      "learning_rate": 7.800000000000001e-05,
      "loss": 1.1862,
      "step": 40
    },
    {
      "grad_norm": 0.8558338284492493,
      "learning_rate": 9.8e-05,
      "loss": 1.1176,
      "step": 50
    },
    {
      "grad_norm": 0.3106129765510559,
      "learning_rate": 9.997785653888835e-05,
      "loss": 1.0896,
      "step": 60
    },
    {
      "grad_norm": 0.18132999539375305,
      "learning_rate": 9.990133642141359e-05,
      "loss": 1.0682,
      "step": 70
    },
    {
      "grad_norm": 0.20137448608875275,
      "learning_rate": 9.977024992520602e-05,
      "loss": 1.0698,
      "step": 80
    },
    {
      "grad_norm": 0.4021719992160797,
      "learning_rate": 9.95847403914247e-05,
      "loss": 1.0598,
      "step": 90
    },
    {
      "grad_norm": 0.25229543447494507,
      "learning_rate": 9.934501067202117e-05,
      "loss": 1.0643,
      "step": 100
    },
    {
      "grad_norm": 0.41172972321510315,
      "learning_rate": 9.905132290792394e-05,
      "loss": 1.0437,
      "step": 110
    },
    {
      "grad_norm": 0.3424518406391144,
      "learning_rate": 9.870399824239117e-05,
      "loss": 1.0168,
      "step": 120
    },
    {
      "grad_norm": 0.5512681007385254,
      "learning_rate": 9.830341646984521e-05,
      "loss": 0.9655,
      "step": 130
    },
    {
      "grad_norm": 0.6318266987800598,
      "learning_rate": 9.785001562057309e-05,
      "loss": 0.9044,
      "step": 140
    },
    {
      "grad_norm": 0.68224036693573,
      "learning_rate": 9.734429148174675e-05,
      "loss": 0.825,
      "step": 150
    },
    {
      "grad_norm": 0.7105351090431213,
      "learning_rate": 9.6786797055287e-05,
      "loss": 0.7075,
      "step": 160
    },
    {
      "grad_norm": 0.9974402189254761,
      "learning_rate": 9.617814195316411e-05,
      "loss": 0.6054,
      "step": 170
    },
    {
      "grad_norm": 0.7529901266098022,
      "learning_rate": 9.551899173079607e-05,
      "loss": 0.5167,
      "step": 180
    },
    {
      "grad_norm": 0.7296067476272583,
      "learning_rate": 9.481006715927351e-05,
      "loss": 0.4093,
      "step": 190
    },
    {
      "grad_norm": 0.8161115050315857,
      "learning_rate": 9.405214343720707e-05,
      "loss": 0.3432,
      "step": 200
    },
    {
      "grad_norm": 0.8231457471847534,
      "learning_rate": 9.32460493430591e-05,
      "loss": 0.3133,
      "step": 210
    },
    {
      "grad_norm": 0.6054251790046692,
      "learning_rate": 9.239266632888659e-05,
      "loss": 0.269,
      "step": 220
    },
    {
      "grad_norm": 0.8241866230964661,
      "learning_rate": 9.14929275564863e-05,
      "loss": 0.2279,
      "step": 230
    },
    {
      "grad_norm": 0.6421325206756592,
      "learning_rate": 9.0547816876996e-05,
      "loss": 0.2042,
      "step": 240
    },
    {
      "grad_norm": 0.7083246111869812,
      "learning_rate": 8.955836775506776e-05,
      "loss": 0.2063,
      "step": 250
    },
    {
      "grad_norm": 0.6825037002563477,
      "learning_rate": 8.852566213878947e-05,
      "loss": 0.1857,
      "step": 260
    },
    {
      "grad_norm": 0.9095971584320068,
      "learning_rate": 8.745082927659047e-05,
      "loss": 0.1583,
      "step": 270
    },
    {
      "grad_norm": 0.7291728258132935,
      "learning_rate": 8.633504448242505e-05,
      "loss": 0.168,
      "step": 280
    },
    {
      "grad_norm": 0.5694337487220764,
      "learning_rate": 8.517952785058385e-05,
      "loss": 0.1521,
      "step": 290
    },
    {
      "grad_norm": 0.6113538146018982,
      "learning_rate": 8.398554292153866e-05,
      "loss": 0.1446,
      "step": 300
    },
    {
      "grad_norm": 0.7520995140075684,
      "learning_rate": 8.275439530027948e-05,
      "loss": 0.1436,
      "step": 310
    },
    {
      "grad_norm": 0.5477866530418396,
      "learning_rate": 8.148743122865463e-05,
      "loss": 0.1256,
      "step": 320
    },
    {
      "grad_norm": 0.7027655839920044,
      "learning_rate": 8.018603611327504e-05,
      "loss": 0.1078,
      "step": 330
    },
    {
      "grad_norm": 0.6788212060928345,
      "learning_rate": 7.88516330105925e-05,
      "loss": 0.0999,
      "step": 340
    },
    {
      "grad_norm": 0.8349484801292419,
      "learning_rate": 7.748568107080832e-05,
      "loss": 0.0976,
      "step": 350
    },
    {
      "grad_norm": 0.6667965650558472,
      "learning_rate": 7.608967394231387e-05,
      "loss": 0.0889,
      "step": 360
    },
    {
      "grad_norm": 0.5367528200149536,
      "learning_rate": 7.466513813840825e-05,
      "loss": 0.0831,
      "step": 370
    },
    {
      "grad_norm": 0.6868879199028015,
      "learning_rate": 7.32136313680782e-05,
      "loss": 0.0782,
      "step": 380
    },
    {
      "grad_norm": 0.5659252405166626,
      "learning_rate": 7.173674083266624e-05,
      "loss": 0.0844,
      "step": 390
    },
    {
      "grad_norm": 0.5637879371643066,
      "learning_rate": 7.023608149028937e-05,
      "loss": 0.0857,
      "step": 400
    },
    {
      "grad_norm": 0.5377715826034546,
      "learning_rate": 6.871329428990602e-05,
      "loss": 0.0887,
      "step": 410
    },
    {
      "grad_norm": 0.6541546583175659,
      "learning_rate": 6.71700443769625e-05,
      "loss": 0.0708,
      "step": 420
    },
    {
      "grad_norm": 0.6478805541992188,
      "learning_rate": 6.56080192725808e-05,
      "loss": 0.0722,
      "step": 430
    },
    {
      "grad_norm": 0.5954216122627258,
      "learning_rate": 6.402892702827916e-05,
      "loss": 0.0765,
      "step": 440
    },
    {
      "grad_norm": 0.533097505569458,
      "learning_rate": 6.243449435824276e-05,
      "loss": 0.0793,
      "step": 450
    },
    {
      "grad_norm": 0.5573644042015076,
      "learning_rate": 6.0826464751186994e-05,
      "loss": 0.0747,
      "step": 460
    },
    {
      "grad_norm": 0.5666183829307556,
      "learning_rate": 5.9206596563878357e-05,
      "loss": 0.0804,
      "step": 470
    },
    {
      "grad_norm": 0.6182389259338379,
      "learning_rate": 5.757666109839702e-05,
      "loss": 0.0675,
      "step": 480
    },
    {
      "grad_norm": 0.41103699803352356,
      "learning_rate": 5.5938440665244006e-05,
      "loss": 0.0638,
      "step": 490
    },
    {
      "grad_norm": 0.5240882635116577,
      "learning_rate": 5.4293726634410855e-05,
      "loss": 0.064,
      "step": 500
    },
    {
      "grad_norm": 0.5724607110023499,
      "learning_rate": 5.264431747654284e-05,
      "loss": 0.0615,
      "step": 510
    },
    {
      "grad_norm": 0.4552757441997528,
      "learning_rate": 5.0992016796337686e-05,
      "loss": 0.0627,
      "step": 520
    },
    {
      "grad_norm": 0.5225529670715332,
      "learning_rate": 4.93386313603304e-05,
      "loss": 0.0608,
      "step": 530
    },
    {
      "grad_norm": 0.491335391998291,
      "learning_rate": 4.7685969121220456e-05,
      "loss": 0.0693,
      "step": 540
    },
    {
      "grad_norm": 0.48702317476272583,
      "learning_rate": 4.60358372409022e-05,
      "loss": 0.0595,
      "step": 550
    },
    {
      "grad_norm": 0.5675700902938843,
      "learning_rate": 4.439004011435979e-05,
      "loss": 0.0651,
      "step": 560
    },
    {
      "grad_norm": 0.5530186295509338,
      "learning_rate": 4.275037739658771e-05,
      "loss": 0.0701,
      "step": 570
    },
    {
      "grad_norm": 0.6130643486976624,
      "learning_rate": 4.111864203469457e-05,
      "loss": 0.0589,
      "step": 580
    },
    {
      "grad_norm": 0.523269534111023,
      "learning_rate": 3.949661830734172e-05,
      "loss": 0.057,
      "step": 590
    },
    {
      "grad_norm": 0.5698956847190857,
      "learning_rate": 3.788607987366069e-05,
      "loss": 0.0677,
      "step": 600
    },
    {
      "grad_norm": 0.4857608675956726,
      "learning_rate": 3.628878783378302e-05,
      "loss": 0.0522,
      "step": 610
    },
    {
      "grad_norm": 0.5062139630317688,
      "learning_rate": 3.470648880310313e-05,
      "loss": 0.0565,
      "step": 620
    },
    {
      "grad_norm": 0.5451298356056213,
      "learning_rate": 3.3140913002379995e-05,
      "loss": 0.0526,
      "step": 630
    },
    {
      "grad_norm": 0.47404196858406067,
      "learning_rate": 3.1593772365766105e-05,
      "loss": 0.0498,
      "step": 640
    },
    {
      "grad_norm": 0.4543800354003906,
      "learning_rate": 3.006675866883275e-05,
      "loss": 0.0501,
      "step": 650
    },
    {
      "grad_norm": 0.4308624863624573,
      "learning_rate": 2.8561541678638142e-05,
      "loss": 0.054,
      "step": 660
    },
    {
      "grad_norm": 0.4590848386287689,
      "learning_rate": 2.707976732786166e-05,
      "loss": 0.0527,
      "step": 670
    },
    {
      "grad_norm": 0.46471887826919556,
      "learning_rate": 2.562305591500069e-05,
      "loss": 0.0515,
      "step": 680
    },
    {
      "grad_norm": 0.5157451033592224,
      "learning_rate": 2.419300033259798e-05,
      "loss": 0.0496,
      "step": 690
    },
    {
      "grad_norm": 0.3607851564884186,
      "learning_rate": 2.279116432543705e-05,
      "loss": 0.0495,
      "step": 700
    },
    {
      "grad_norm": 0.45652496814727783,
      "learning_rate": 2.1419080780610123e-05,
      "loss": 0.0623,
      "step": 710
    },
    {
      "grad_norm": 0.42067599296569824,
      "learning_rate": 2.0078250051328784e-05,
      "loss": 0.0515,
      "step": 720
    },
    {
      "grad_norm": 0.447702556848526,
      "learning_rate": 1.877013831630961e-05,
      "loss": 0.0409,
      "step": 730
    },
    {
      "grad_norm": 0.46375778317451477,
      "learning_rate": 1.749617597652934e-05,
      "loss": 0.0568,
      "step": 740
    },
    {
      "grad_norm": 0.40698182582855225,
      "learning_rate": 1.62577560911024e-05,
      "loss": 0.0549,
      "step": 750
    },
    {
      "grad_norm": 0.36565569043159485,
      "learning_rate": 1.5056232853991209e-05,
      "loss": 0.0619,
      "step": 760
    },
    {
      "grad_norm": 0.3338938355445862,
      "learning_rate": 1.389292011321498e-05,
      "loss": 0.0471,
      "step": 770
    },
    {
      "grad_norm": 0.391041100025177,
      "learning_rate": 1.2769089934176126e-05,
      "loss": 0.0513,
      "step": 780
    },
    {
      "grad_norm": 0.2763245701789856,
      "learning_rate": 1.1685971208675539e-05,
      "loss": 0.0551,
      "step": 790
    },
    {
      "grad_norm": 0.30885049700737,
      "learning_rate": 1.0644748311137376e-05,
      "loss": 0.0513,
      "step": 800
    },
    {
      "grad_norm": 0.28447505831718445,
      "learning_rate": 9.646559803512994e-06,
      "loss": 0.0481,
      "step": 810
    },
    {
      "grad_norm": 0.34004244208335876,
      "learning_rate": 8.692497190280224e-06,
      "loss": 0.0498,
      "step": 820
    },
    {
      "grad_norm": 0.27130886912345886,
      "learning_rate": 7.783603724899257e-06,
      "loss": 0.0462,
      "step": 830
    },
    {
      "grad_norm": 0.2683325707912445,
      "learning_rate": 6.92087326903022e-06,
      "loss": 0.046,
      "step": 840
    },
    {
      "grad_norm": 0.3312835991382599,
      "learning_rate": 6.1052492057601275e-06,
      "loss": 0.0483,
      "step": 850
    },
    {
      "grad_norm": 0.35007426142692566,
      "learning_rate": 5.337623408027293e-06,
      "loss": 0.056,
      "step": 860
    },
    {
      "grad_norm": 0.2915596067905426,
      "learning_rate": 4.618835263371396e-06,
      "loss": 0.0485,
      "step": 870
    },
    {
      "grad_norm": 0.3173588812351227,
      "learning_rate": 3.949670756075447e-06,
      "loss": 0.0473,
      "step": 880
    },
    {
      "grad_norm": 0.3005145788192749,
      "learning_rate": 3.3308616077036115e-06,
      "loss": 0.0479,
      "step": 890
    },
    {
      "grad_norm": 0.28296926617622375,
      "learning_rate": 2.7630844769743757e-06,
      "loss": 0.0422,
      "step": 900
    },
    {
      "grad_norm": 0.2193201184272766,
      "learning_rate": 2.2469602198441573e-06,
      "loss": 0.0421,
      "step": 910
    },
    {
      "grad_norm": 0.1950123906135559,
      "learning_rate": 1.7830532106104747e-06,
      "loss": 0.0447,
      "step": 920
    },
    {
      "grad_norm": 0.32784703373908997,
      "learning_rate": 1.3718707247769135e-06,
      "loss": 0.0448,
      "step": 930
    },
    {
      "grad_norm": 0.20192043483257294,
      "learning_rate": 1.0138623843548078e-06,
      "loss": 0.0414,
      "step": 940
    },
    {
      "grad_norm": 0.23864571750164032,
      "learning_rate": 7.094196662081831e-07,
      "loss": 0.0476,
      "step": 950
    },
    {
      "grad_norm": 0.22843579947948456,
      "learning_rate": 4.5887547397955864e-07,
      "loss": 0.0449,
      "step": 960
    },
    {
      "grad_norm": 0.24962212145328522,
      "learning_rate": 2.625037740646763e-07,
      "loss": 0.0455,
      "step": 970
    },
    {
      "grad_norm": 0.316610187292099,
      "learning_rate": 1.2051929603428825e-07,
      "loss": 0.0467,
      "step": 980
    },
    {
      "grad_norm": 0.32058215141296387,
      "learning_rate": 3.3077297830541584e-08,
      "loss": 0.0506,
      "step": 990
    },
    {
      "grad_norm": 0.2989563047885895,
      "learning_rate": 2.7339599464326627e-10,
      "loss": 0.0532,
      "step": 1000
    }
  ],
  "logging_steps": 10,
  "max_steps": 1000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 48,
  "trial_name": null,
  "trial_params": null
}