{
  "best_metric": 3.101016044616699,
  "best_model_checkpoint": "./output/models/parallel-gpt2-medium-wikitext/checkpoint-17500",
  "epoch": 4.998736842105263,
  "eval_steps": 500,
  "global_step": 17810,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.028070175438596492,
      "grad_norm": 2.028989553451538,
      "learning_rate": 5.614823133071308e-06,
      "loss": 9.2256,
      "step": 100
    },
    {
      "epoch": 0.056140350877192984,
      "grad_norm": 1.8919556140899658,
      "learning_rate": 1.1229646266142617e-05,
      "loss": 7.8733,
      "step": 200
    },
    {
      "epoch": 0.08421052631578947,
      "grad_norm": 1.433174967765808,
      "learning_rate": 1.6844469399213926e-05,
      "loss": 7.0853,
      "step": 300
    },
    {
      "epoch": 0.11228070175438597,
      "grad_norm": 1.4522157907485962,
      "learning_rate": 2.2459292532285233e-05,
      "loss": 6.6702,
      "step": 400
    },
    {
      "epoch": 0.14035087719298245,
      "grad_norm": 1.322840929031372,
      "learning_rate": 2.8074115665356544e-05,
      "loss": 6.4455,
      "step": 500
    },
    {
      "epoch": 0.14035087719298245,
      "eval_accuracy": 0.17660504282313108,
      "eval_bleu": 0.02573169351336114,
      "eval_loss": 6.331261157989502,
      "eval_perplexity": 561.8647474275982,
      "eval_runtime": 20.1474,
      "eval_samples_per_second": 56.633,
      "eval_steps_per_second": 3.574,
      "step": 500
    },
    {
      "epoch": 0.16842105263157894,
      "grad_norm": 1.5128698348999023,
      "learning_rate": 3.368893879842785e-05,
      "loss": 6.2694,
      "step": 600
    },
    {
      "epoch": 0.19649122807017544,
      "grad_norm": 1.4778157472610474,
      "learning_rate": 3.930376193149916e-05,
      "loss": 6.1217,
      "step": 700
    },
    {
      "epoch": 0.22456140350877193,
      "grad_norm": 1.1647950410842896,
      "learning_rate": 4.4918585064570466e-05,
      "loss": 5.9897,
      "step": 800
    },
    {
      "epoch": 0.25263157894736843,
      "grad_norm": 1.3264423608779907,
      "learning_rate": 5.053340819764178e-05,
      "loss": 5.8659,
      "step": 900
    },
    {
      "epoch": 0.2807017543859649,
      "grad_norm": 1.521269679069519,
      "learning_rate": 5.614823133071309e-05,
      "loss": 5.7254,
      "step": 1000
    },
    {
      "epoch": 0.2807017543859649,
      "eval_accuracy": 0.2136290386834618,
      "eval_bleu": 0.045427591918014404,
      "eval_loss": 5.623491287231445,
      "eval_perplexity": 276.85427569402634,
      "eval_runtime": 20.1737,
      "eval_samples_per_second": 56.559,
      "eval_steps_per_second": 3.569,
      "step": 1000
    },
    {
      "epoch": 0.3087719298245614,
      "grad_norm": 1.1612671613693237,
      "learning_rate": 6.176305446378439e-05,
      "loss": 5.5759,
      "step": 1100
    },
    {
      "epoch": 0.3368421052631579,
      "grad_norm": 1.1074801683425903,
      "learning_rate": 6.73778775968557e-05,
      "loss": 5.4609,
      "step": 1200
    },
    {
      "epoch": 0.3649122807017544,
      "grad_norm": 1.0632935762405396,
      "learning_rate": 7.299270072992701e-05,
      "loss": 5.3257,
      "step": 1300
    },
    {
      "epoch": 0.3929824561403509,
      "grad_norm": 1.1046158075332642,
      "learning_rate": 7.860752386299832e-05,
      "loss": 5.1946,
      "step": 1400
    },
    {
      "epoch": 0.42105263157894735,
      "grad_norm": 1.1495596170425415,
      "learning_rate": 8.422234699606962e-05,
      "loss": 5.1084,
      "step": 1500
    },
    {
      "epoch": 0.42105263157894735,
      "eval_accuracy": 0.2575530545053601,
      "eval_bleu": 0.06490299322624599,
      "eval_loss": 4.982165813446045,
      "eval_perplexity": 145.78979346372705,
      "eval_runtime": 20.1855,
      "eval_samples_per_second": 56.526,
      "eval_steps_per_second": 3.567,
      "step": 1500
    },
    {
      "epoch": 0.44912280701754387,
      "grad_norm": 0.9580332040786743,
      "learning_rate": 8.983717012914093e-05,
      "loss": 4.9932,
      "step": 1600
    },
    {
      "epoch": 0.47719298245614034,
      "grad_norm": 0.9515564441680908,
      "learning_rate": 9.545199326221224e-05,
      "loss": 4.8819,
      "step": 1700
    },
    {
      "epoch": 0.5052631578947369,
      "grad_norm": 0.9855053424835205,
      "learning_rate": 9.98814648449685e-05,
      "loss": 4.7868,
      "step": 1800
    },
    {
      "epoch": 0.5333333333333333,
      "grad_norm": 1.0349925756454468,
      "learning_rate": 9.925759560796058e-05,
      "loss": 4.6832,
      "step": 1900
    },
    {
      "epoch": 0.5614035087719298,
      "grad_norm": 1.0200501680374146,
      "learning_rate": 9.863372637095265e-05,
      "loss": 4.5994,
      "step": 2000
    },
    {
      "epoch": 0.5614035087719298,
      "eval_accuracy": 0.29288759924026103,
      "eval_bleu": 0.07406906637840037,
      "eval_loss": 4.5052409172058105,
      "eval_perplexity": 90.49014205518364,
      "eval_runtime": 19.5875,
      "eval_samples_per_second": 58.251,
      "eval_steps_per_second": 3.676,
      "step": 2000
    },
    {
      "epoch": 0.5894736842105263,
      "grad_norm": 0.9353361129760742,
      "learning_rate": 9.800985713394473e-05,
      "loss": 4.5176,
      "step": 2100
    },
    {
      "epoch": 0.6175438596491228,
      "grad_norm": 0.9589385390281677,
      "learning_rate": 9.73859878969368e-05,
      "loss": 4.4355,
      "step": 2200
    },
    {
      "epoch": 0.6456140350877193,
      "grad_norm": 1.0331965684890747,
      "learning_rate": 9.676211865992888e-05,
      "loss": 4.3647,
      "step": 2300
    },
    {
      "epoch": 0.6736842105263158,
      "grad_norm": 0.9204744100570679,
      "learning_rate": 9.613824942292095e-05,
      "loss": 4.2935,
      "step": 2400
    },
    {
      "epoch": 0.7017543859649122,
      "grad_norm": 0.9139348268508911,
      "learning_rate": 9.551438018591303e-05,
      "loss": 4.2338,
      "step": 2500
    },
    {
      "epoch": 0.7017543859649122,
      "eval_accuracy": 0.32728917628977,
      "eval_bleu": 0.09369678160356684,
      "eval_loss": 4.137840747833252,
      "eval_perplexity": 62.66736062058415,
      "eval_runtime": 20.1595,
      "eval_samples_per_second": 56.598,
      "eval_steps_per_second": 3.572,
      "step": 2500
    },
    {
      "epoch": 0.7298245614035088,
      "grad_norm": 0.9106621146202087,
      "learning_rate": 9.489051094890511e-05,
      "loss": 4.1694,
      "step": 2600
    },
    {
      "epoch": 0.7578947368421053,
      "grad_norm": 0.9036598801612854,
      "learning_rate": 9.426664171189718e-05,
      "loss": 4.1386,
      "step": 2700
    },
    {
      "epoch": 0.7859649122807018,
      "grad_norm": 0.8367689847946167,
      "learning_rate": 9.364277247488927e-05,
      "loss": 4.0731,
      "step": 2800
    },
    {
      "epoch": 0.8140350877192982,
      "grad_norm": 0.9330971837043762,
      "learning_rate": 9.301890323788135e-05,
      "loss": 4.0342,
      "step": 2900
    },
    {
      "epoch": 0.8421052631578947,
      "grad_norm": 0.847179114818573,
      "learning_rate": 9.239503400087343e-05,
      "loss": 3.9975,
      "step": 3000
    },
    {
      "epoch": 0.8421052631578947,
      "eval_accuracy": 0.3465208187155545,
      "eval_bleu": 0.10305263495616372,
      "eval_loss": 3.928612470626831,
      "eval_perplexity": 50.83639172374704,
      "eval_runtime": 20.1917,
      "eval_samples_per_second": 56.508,
      "eval_steps_per_second": 3.566,
      "step": 3000
    },
    {
      "epoch": 0.8701754385964913,
      "grad_norm": 0.8944078087806702,
      "learning_rate": 9.17711647638655e-05,
      "loss": 3.9805,
      "step": 3100
    },
    {
      "epoch": 0.8982456140350877,
      "grad_norm": 0.8461546301841736,
      "learning_rate": 9.114729552685758e-05,
      "loss": 3.9468,
      "step": 3200
    },
    {
      "epoch": 0.9263157894736842,
      "grad_norm": 0.8034185767173767,
      "learning_rate": 9.052342628984965e-05,
      "loss": 3.9049,
      "step": 3300
    },
    {
      "epoch": 0.9543859649122807,
      "grad_norm": 0.8450652360916138,
      "learning_rate": 8.989955705284174e-05,
      "loss": 3.8727,
      "step": 3400
    },
    {
      "epoch": 0.9824561403508771,
      "grad_norm": 0.7972449064254761,
      "learning_rate": 8.92756878158338e-05,
      "loss": 3.8648,
      "step": 3500
    },
    {
      "epoch": 0.9824561403508771,
      "eval_accuracy": 0.35825016727450926,
      "eval_bleu": 0.11656022916295831,
      "eval_loss": 3.7925570011138916,
      "eval_perplexity": 44.36970874398872,
      "eval_runtime": 20.1635,
      "eval_samples_per_second": 56.587,
      "eval_steps_per_second": 3.571,
      "step": 3500
    },
    {
      "epoch": 1.0103859649122806,
      "grad_norm": 0.766300618648529,
      "learning_rate": 8.865181857882589e-05,
      "loss": 3.8193,
      "step": 3600
    },
    {
      "epoch": 1.0384561403508772,
      "grad_norm": 0.8034301996231079,
      "learning_rate": 8.802794934181796e-05,
      "loss": 3.7801,
      "step": 3700
    },
    {
      "epoch": 1.0665263157894738,
      "grad_norm": 0.8166210651397705,
      "learning_rate": 8.740408010481004e-05,
      "loss": 3.7505,
      "step": 3800
    },
    {
      "epoch": 1.0945964912280701,
      "grad_norm": 0.7962071895599365,
      "learning_rate": 8.678021086780212e-05,
      "loss": 3.7398,
      "step": 3900
    },
    {
      "epoch": 1.1226666666666667,
      "grad_norm": 0.7868255972862244,
      "learning_rate": 8.615634163079419e-05,
      "loss": 3.7164,
      "step": 4000
    },
    {
      "epoch": 1.1226666666666667,
      "eval_accuracy": 0.3667368320049895,
      "eval_bleu": 0.1225879386890452,
      "eval_loss": 3.6986546516418457,
      "eval_perplexity": 40.39292523319135,
      "eval_runtime": 20.1969,
      "eval_samples_per_second": 56.494,
      "eval_steps_per_second": 3.565,
      "step": 4000
    },
    {
      "epoch": 1.150736842105263,
      "grad_norm": 0.7974035739898682,
      "learning_rate": 8.553247239378627e-05,
      "loss": 3.7099,
      "step": 4100
    },
    {
      "epoch": 1.1788070175438596,
      "grad_norm": 0.7970170974731445,
      "learning_rate": 8.490860315677834e-05,
      "loss": 3.6871,
      "step": 4200
    },
    {
      "epoch": 1.2068771929824562,
      "grad_norm": 0.7822019457817078,
      "learning_rate": 8.428473391977042e-05,
      "loss": 3.6628,
      "step": 4300
    },
    {
      "epoch": 1.2349473684210526,
      "grad_norm": 0.7597511410713196,
      "learning_rate": 8.36608646827625e-05,
      "loss": 3.6631,
      "step": 4400
    },
    {
      "epoch": 1.2630175438596491,
      "grad_norm": 0.7736372947692871,
      "learning_rate": 8.303699544575458e-05,
      "loss": 3.6639,
      "step": 4500
    },
    {
      "epoch": 1.2630175438596491,
      "eval_accuracy": 0.37342095861787133,
      "eval_bleu": 0.12824268044271067,
      "eval_loss": 3.622089147567749,
      "eval_perplexity": 37.415653048434656,
      "eval_runtime": 20.1594,
      "eval_samples_per_second": 56.599,
      "eval_steps_per_second": 3.572,
      "step": 4500
    },
    {
      "epoch": 1.2910877192982455,
      "grad_norm": 0.7472370266914368,
      "learning_rate": 8.241312620874664e-05,
      "loss": 3.6404,
      "step": 4600
    },
    {
      "epoch": 1.319157894736842,
      "grad_norm": 0.771800696849823,
      "learning_rate": 8.178925697173873e-05,
      "loss": 3.6235,
      "step": 4700
    },
    {
      "epoch": 1.3472280701754387,
      "grad_norm": 0.7624268531799316,
      "learning_rate": 8.11653877347308e-05,
      "loss": 3.6123,
      "step": 4800
    },
    {
      "epoch": 1.375298245614035,
      "grad_norm": 0.7671138644218445,
      "learning_rate": 8.054151849772288e-05,
      "loss": 3.592,
      "step": 4900
    },
    {
      "epoch": 1.4033684210526316,
      "grad_norm": 0.7753613591194153,
      "learning_rate": 7.991764926071496e-05,
      "loss": 3.582,
      "step": 5000
    },
    {
      "epoch": 1.4033684210526316,
      "eval_accuracy": 0.3795687787375893,
      "eval_bleu": 0.1276585909940598,
      "eval_loss": 3.5575244426727295,
      "eval_perplexity": 35.07625629269647,
      "eval_runtime": 20.1451,
      "eval_samples_per_second": 56.639,
      "eval_steps_per_second": 3.574,
      "step": 5000
    },
    {
      "epoch": 1.431438596491228,
      "grad_norm": 0.7279810309410095,
      "learning_rate": 7.929378002370703e-05,
      "loss": 3.5846,
      "step": 5100
    },
    {
      "epoch": 1.4595087719298245,
      "grad_norm": 0.728085458278656,
      "learning_rate": 7.866991078669911e-05,
      "loss": 3.5677,
      "step": 5200
    },
    {
      "epoch": 1.4875789473684211,
      "grad_norm": 0.7683869004249573,
      "learning_rate": 7.804604154969118e-05,
      "loss": 3.5452,
      "step": 5300
    },
    {
      "epoch": 1.5156491228070177,
      "grad_norm": 0.7563459277153015,
      "learning_rate": 7.742217231268327e-05,
      "loss": 3.5557,
      "step": 5400
    },
    {
      "epoch": 1.543719298245614,
      "grad_norm": 0.7477959394454956,
      "learning_rate": 7.679830307567533e-05,
      "loss": 3.5315,
      "step": 5500
    },
    {
      "epoch": 1.543719298245614,
      "eval_accuracy": 0.38395261312340273,
      "eval_bleu": 0.13124123866011556,
      "eval_loss": 3.5063869953155518,
      "eval_perplexity": 33.32763708661619,
      "eval_runtime": 20.2862,
      "eval_samples_per_second": 56.245,
      "eval_steps_per_second": 3.549,
      "step": 5500
    },
    {
      "epoch": 1.5717894736842104,
      "grad_norm": 0.732016921043396,
      "learning_rate": 7.617443383866742e-05,
      "loss": 3.5185,
      "step": 5600
    },
    {
      "epoch": 1.599859649122807,
      "grad_norm": 0.7203788757324219,
      "learning_rate": 7.555056460165949e-05,
      "loss": 3.5216,
      "step": 5700
    },
    {
      "epoch": 1.6279298245614036,
      "grad_norm": 0.7561742663383484,
      "learning_rate": 7.492669536465158e-05,
      "loss": 3.5057,
      "step": 5800
    },
    {
      "epoch": 1.6560000000000001,
      "grad_norm": 0.7377160787582397,
      "learning_rate": 7.430282612764365e-05,
      "loss": 3.5058,
      "step": 5900
    },
    {
      "epoch": 1.6840701754385965,
      "grad_norm": 0.8002681136131287,
      "learning_rate": 7.367895689063573e-05,
      "loss": 3.5025,
      "step": 6000
    },
    {
      "epoch": 1.6840701754385965,
      "eval_accuracy": 0.3880588703466202,
      "eval_bleu": 0.13662311841214886,
      "eval_loss": 3.4594311714172363,
      "eval_perplexity": 31.798883255478383,
      "eval_runtime": 20.2931,
      "eval_samples_per_second": 56.226,
      "eval_steps_per_second": 3.548,
      "step": 6000
    },
    {
      "epoch": 1.7121403508771929,
      "grad_norm": 0.7451775074005127,
      "learning_rate": 7.30550876536278e-05,
      "loss": 3.4714,
      "step": 6100
    },
    {
      "epoch": 1.7402105263157894,
      "grad_norm": 0.7121933102607727,
      "learning_rate": 7.243121841661989e-05,
      "loss": 3.474,
      "step": 6200
    },
    {
      "epoch": 1.768280701754386,
      "grad_norm": 0.7507015466690063,
      "learning_rate": 7.180734917961195e-05,
      "loss": 3.4577,
      "step": 6300
    },
    {
      "epoch": 1.7963508771929826,
      "grad_norm": 0.7061077356338501,
      "learning_rate": 7.118347994260404e-05,
      "loss": 3.463,
      "step": 6400
    },
    {
      "epoch": 1.824421052631579,
      "grad_norm": 0.7301498055458069,
      "learning_rate": 7.055961070559612e-05,
      "loss": 3.4462,
      "step": 6500
    },
    {
      "epoch": 1.824421052631579,
      "eval_accuracy": 0.3919063982392698,
      "eval_bleu": 0.13096345408557636,
      "eval_loss": 3.420844554901123,
      "eval_perplexity": 30.5952434755694,
      "eval_runtime": 20.2154,
      "eval_samples_per_second": 56.442,
      "eval_steps_per_second": 3.562,
      "step": 6500
    },
    {
      "epoch": 1.8524912280701753,
      "grad_norm": 0.713631272315979,
      "learning_rate": 6.993574146858819e-05,
      "loss": 3.4245,
      "step": 6600
    },
    {
      "epoch": 1.8805614035087719,
      "grad_norm": 0.7331680059432983,
      "learning_rate": 6.931187223158027e-05,
      "loss": 3.4289,
      "step": 6700
    },
    {
      "epoch": 1.9086315789473685,
      "grad_norm": 0.7402526140213013,
      "learning_rate": 6.868800299457234e-05,
      "loss": 3.434,
      "step": 6800
    },
    {
      "epoch": 1.936701754385965,
      "grad_norm": 0.7350389361381531,
      "learning_rate": 6.806413375756442e-05,
      "loss": 3.413,
      "step": 6900
    },
    {
      "epoch": 1.9647719298245614,
      "grad_norm": 0.7257172465324402,
      "learning_rate": 6.744026452055649e-05,
      "loss": 3.4167,
      "step": 7000
    },
    {
      "epoch": 1.9647719298245614,
      "eval_accuracy": 0.39561171067207085,
      "eval_bleu": 0.1355256129038967,
      "eval_loss": 3.3863015174865723,
      "eval_perplexity": 29.556435911658944,
      "eval_runtime": 20.2056,
      "eval_samples_per_second": 56.469,
      "eval_steps_per_second": 3.563,
      "step": 7000
    },
    {
      "epoch": 1.9928421052631577,
      "grad_norm": 0.7187970876693726,
      "learning_rate": 6.681639528354857e-05,
      "loss": 3.4086,
      "step": 7100
    },
    {
      "epoch": 2.020771929824561,
      "grad_norm": 0.7202402353286743,
      "learning_rate": 6.619252604654064e-05,
      "loss": 3.3372,
      "step": 7200
    },
    {
      "epoch": 2.048842105263158,
      "grad_norm": 0.7403327822685242,
      "learning_rate": 6.556865680953273e-05,
      "loss": 3.3079,
      "step": 7300
    },
    {
      "epoch": 2.0769122807017544,
      "grad_norm": 0.7321439385414124,
      "learning_rate": 6.49447875725248e-05,
      "loss": 3.3128,
      "step": 7400
    },
    {
      "epoch": 2.104982456140351,
      "grad_norm": 0.7240020632743835,
      "learning_rate": 6.432091833551688e-05,
      "loss": 3.2967,
      "step": 7500
    },
    {
      "epoch": 2.104982456140351,
      "eval_accuracy": 0.39890408423952856,
      "eval_bleu": 0.1316612565684532,
      "eval_loss": 3.3547873497009277,
      "eval_perplexity": 28.639513342041703,
      "eval_runtime": 20.2147,
      "eval_samples_per_second": 56.444,
      "eval_steps_per_second": 3.562,
      "step": 7500
    },
    {
      "epoch": 2.1330526315789475,
      "grad_norm": 0.7483673095703125,
      "learning_rate": 6.369704909850896e-05,
      "loss": 3.2857,
      "step": 7600
    },
    {
      "epoch": 2.1611228070175437,
      "grad_norm": 0.7482850551605225,
      "learning_rate": 6.307317986150103e-05,
      "loss": 3.3038,
      "step": 7700
    },
    {
      "epoch": 2.1891929824561402,
      "grad_norm": 0.7295832633972168,
      "learning_rate": 6.244931062449311e-05,
      "loss": 3.284,
      "step": 7800
    },
    {
      "epoch": 2.217263157894737,
      "grad_norm": 0.7610414028167725,
      "learning_rate": 6.182544138748518e-05,
      "loss": 3.2862,
      "step": 7900
    },
    {
      "epoch": 2.2453333333333334,
      "grad_norm": 0.7399555444717407,
      "learning_rate": 6.120157215047726e-05,
      "loss": 3.2909,
      "step": 8000
    },
    {
      "epoch": 2.2453333333333334,
      "eval_accuracy": 0.4014682461149906,
      "eval_bleu": 0.13811168690281525,
      "eval_loss": 3.329040050506592,
      "eval_perplexity": 27.911535174787044,
      "eval_runtime": 20.1568,
      "eval_samples_per_second": 56.606,
      "eval_steps_per_second": 3.572,
      "step": 8000
    },
    {
      "epoch": 2.27340350877193,
      "grad_norm": 0.7306208610534668,
      "learning_rate": 6.057770291346934e-05,
      "loss": 3.2853,
      "step": 8100
    },
    {
      "epoch": 2.301473684210526,
      "grad_norm": 0.7453253269195557,
      "learning_rate": 5.9953833676461415e-05,
      "loss": 3.2744,
      "step": 8200
    },
    {
      "epoch": 2.3295438596491227,
      "grad_norm": 0.7369946837425232,
      "learning_rate": 5.932996443945349e-05,
      "loss": 3.2604,
      "step": 8300
    },
    {
      "epoch": 2.3576140350877193,
      "grad_norm": 0.7455064654350281,
      "learning_rate": 5.870609520244557e-05,
      "loss": 3.2622,
      "step": 8400
    },
    {
      "epoch": 2.385684210526316,
      "grad_norm": 0.7593878507614136,
      "learning_rate": 5.808222596543764e-05,
      "loss": 3.2593,
      "step": 8500
    },
    {
      "epoch": 2.385684210526316,
      "eval_accuracy": 0.40385506702546087,
      "eval_bleu": 0.14220030200690656,
      "eval_loss": 3.3044025897979736,
      "eval_perplexity": 27.232267893585384,
      "eval_runtime": 20.1746,
      "eval_samples_per_second": 56.556,
      "eval_steps_per_second": 3.569,
      "step": 8500
    },
    {
      "epoch": 2.4137543859649124,
      "grad_norm": 0.7376691102981567,
      "learning_rate": 5.745835672842972e-05,
      "loss": 3.2582,
      "step": 8600
    },
    {
      "epoch": 2.441824561403509,
      "grad_norm": 0.7452288269996643,
      "learning_rate": 5.683448749142181e-05,
      "loss": 3.2491,
      "step": 8700
    },
    {
      "epoch": 2.469894736842105,
      "grad_norm": 0.749320924282074,
      "learning_rate": 5.6210618254413884e-05,
      "loss": 3.24,
      "step": 8800
    },
    {
      "epoch": 2.4979649122807017,
      "grad_norm": 0.7482838034629822,
      "learning_rate": 5.558674901740596e-05,
      "loss": 3.2374,
      "step": 8900
    },
    {
      "epoch": 2.5260350877192983,
      "grad_norm": 0.7401648163795471,
      "learning_rate": 5.4962879780398035e-05,
      "loss": 3.2408,
      "step": 9000
    },
    {
      "epoch": 2.5260350877192983,
      "eval_accuracy": 0.4061330845419506,
      "eval_bleu": 0.14119139857697727,
      "eval_loss": 3.282592296600342,
      "eval_perplexity": 26.644754356923094,
      "eval_runtime": 20.1938,
      "eval_samples_per_second": 56.503,
      "eval_steps_per_second": 3.565,
      "step": 9000
    },
    {
      "epoch": 2.554105263157895,
      "grad_norm": 0.7236805558204651,
      "learning_rate": 5.433901054339011e-05,
      "loss": 3.2434,
      "step": 9100
    },
    {
      "epoch": 2.582175438596491,
      "grad_norm": 0.7386437058448792,
      "learning_rate": 5.371514130638219e-05,
      "loss": 3.2445,
      "step": 9200
    },
    {
      "epoch": 2.6102456140350876,
      "grad_norm": 0.7296267747879028,
      "learning_rate": 5.309127206937426e-05,
      "loss": 3.2257,
      "step": 9300
    },
    {
      "epoch": 2.638315789473684,
      "grad_norm": 0.7237492203712463,
      "learning_rate": 5.246740283236634e-05,
      "loss": 3.2336,
      "step": 9400
    },
    {
      "epoch": 2.6663859649122807,
      "grad_norm": 0.7474483847618103,
      "learning_rate": 5.1843533595358414e-05,
      "loss": 3.2278,
      "step": 9500
    },
    {
      "epoch": 2.6663859649122807,
      "eval_accuracy": 0.40903479395464354,
      "eval_bleu": 0.14356756322892422,
      "eval_loss": 3.2591938972473145,
      "eval_perplexity": 26.028547000163957,
      "eval_runtime": 20.1915,
      "eval_samples_per_second": 56.509,
      "eval_steps_per_second": 3.566,
      "step": 9500
    },
    {
      "epoch": 2.6944561403508773,
      "grad_norm": 0.7365037798881531,
      "learning_rate": 5.12196643583505e-05,
      "loss": 3.2235,
      "step": 9600
    },
    {
      "epoch": 2.722526315789474,
      "grad_norm": 0.7330195307731628,
      "learning_rate": 5.059579512134257e-05,
      "loss": 3.2179,
      "step": 9700
    },
    {
      "epoch": 2.75059649122807,
      "grad_norm": 0.7449477314949036,
      "learning_rate": 4.997192588433465e-05,
      "loss": 3.2178,
      "step": 9800
    },
    {
      "epoch": 2.7786666666666666,
      "grad_norm": 0.7243569493293762,
      "learning_rate": 4.9348056647326725e-05,
      "loss": 3.2126,
      "step": 9900
    },
    {
      "epoch": 2.806736842105263,
      "grad_norm": 0.7603092789649963,
      "learning_rate": 4.87241874103188e-05,
      "loss": 3.2172,
      "step": 10000
    },
    {
      "epoch": 2.806736842105263,
      "eval_accuracy": 0.4104792232637077,
      "eval_bleu": 0.1411554473599086,
      "eval_loss": 3.2415478229522705,
      "eval_perplexity": 25.573274029997943,
      "eval_runtime": 19.9283,
      "eval_samples_per_second": 57.255,
      "eval_steps_per_second": 3.613,
      "step": 10000
    },
    {
      "epoch": 2.8348070175438598,
      "grad_norm": 0.7319552302360535,
      "learning_rate": 4.8100318173310876e-05,
      "loss": 3.2138,
      "step": 10100
    },
    {
      "epoch": 2.862877192982456,
      "grad_norm": 0.7321156859397888,
      "learning_rate": 4.747644893630295e-05,
      "loss": 3.2114,
      "step": 10200
    },
    {
      "epoch": 2.8909473684210525,
      "grad_norm": 0.7428148984909058,
      "learning_rate": 4.685257969929503e-05,
      "loss": 3.1985,
      "step": 10300
    },
    {
      "epoch": 2.919017543859649,
      "grad_norm": 0.7498809099197388,
      "learning_rate": 4.6228710462287104e-05,
      "loss": 3.1861,
      "step": 10400
    },
    {
      "epoch": 2.9470877192982456,
      "grad_norm": 0.7468050718307495,
      "learning_rate": 4.560484122527918e-05,
      "loss": 3.2145,
      "step": 10500
    },
    {
      "epoch": 2.9470877192982456,
      "eval_accuracy": 0.41254477431006226,
      "eval_bleu": 0.14022835224132563,
      "eval_loss": 3.222651243209839,
      "eval_perplexity": 25.094563855273414,
      "eval_runtime": 20.1985,
      "eval_samples_per_second": 56.489,
      "eval_steps_per_second": 3.565,
      "step": 10500
    },
    {
      "epoch": 2.9751578947368422,
      "grad_norm": 0.7379462122917175,
      "learning_rate": 4.498097198827126e-05,
      "loss": 3.1799,
      "step": 10600
    },
    {
      "epoch": 3.0030877192982457,
      "grad_norm": 0.7374199032783508,
      "learning_rate": 4.435710275126334e-05,
      "loss": 3.1681,
      "step": 10700
    },
    {
      "epoch": 3.0311578947368423,
      "grad_norm": 0.7500795722007751,
      "learning_rate": 4.3733233514255414e-05,
      "loss": 3.096,
      "step": 10800
    },
    {
      "epoch": 3.0592280701754384,
      "grad_norm": 0.7644433975219727,
      "learning_rate": 4.3109364277247496e-05,
      "loss": 3.0842,
      "step": 10900
    },
    {
      "epoch": 3.087298245614035,
      "grad_norm": 0.7695332169532776,
      "learning_rate": 4.248549504023957e-05,
      "loss": 3.0749,
      "step": 11000
    },
    {
      "epoch": 3.087298245614035,
      "eval_accuracy": 0.41432503771708207,
      "eval_bleu": 0.1413248900872156,
      "eval_loss": 3.209906578063965,
      "eval_perplexity": 24.776771422507743,
      "eval_runtime": 20.1941,
      "eval_samples_per_second": 56.502,
      "eval_steps_per_second": 3.565,
      "step": 11000
    },
    {
      "epoch": 3.1153684210526316,
      "grad_norm": 0.7983810901641846,
      "learning_rate": 4.186162580323165e-05,
      "loss": 3.0886,
      "step": 11100
    },
    {
      "epoch": 3.143438596491228,
      "grad_norm": 0.7757524847984314,
      "learning_rate": 4.1237756566223724e-05,
      "loss": 3.0873,
      "step": 11200
    },
    {
      "epoch": 3.1715087719298247,
      "grad_norm": 0.8084014654159546,
      "learning_rate": 4.06138873292158e-05,
      "loss": 3.0882,
      "step": 11300
    },
    {
      "epoch": 3.199578947368421,
      "grad_norm": 0.7784711718559265,
      "learning_rate": 3.9990018092207875e-05,
      "loss": 3.0807,
      "step": 11400
    },
    {
      "epoch": 3.2276491228070174,
      "grad_norm": 0.7808659076690674,
      "learning_rate": 3.936614885519995e-05,
      "loss": 3.0777,
      "step": 11500
    },
    {
      "epoch": 3.2276491228070174,
      "eval_accuracy": 0.4159836469355567,
      "eval_bleu": 0.14196972146127593,
      "eval_loss": 3.1977927684783936,
      "eval_perplexity": 24.478440938871064,
      "eval_runtime": 20.2549,
      "eval_samples_per_second": 56.332,
      "eval_steps_per_second": 3.555,
      "step": 11500
    },
    {
      "epoch": 3.255719298245614,
      "grad_norm": 0.7880111336708069,
      "learning_rate": 3.874227961819203e-05,
      "loss": 3.0907,
      "step": 11600
    },
    {
      "epoch": 3.2837894736842106,
      "grad_norm": 0.7764289975166321,
      "learning_rate": 3.81184103811841e-05,
      "loss": 3.0829,
      "step": 11700
    },
    {
      "epoch": 3.311859649122807,
      "grad_norm": 0.7774072885513306,
      "learning_rate": 3.749454114417618e-05,
      "loss": 3.084,
      "step": 11800
    },
    {
      "epoch": 3.3399298245614037,
      "grad_norm": 0.7633748650550842,
      "learning_rate": 3.687067190716826e-05,
      "loss": 3.0832,
      "step": 11900
    },
    {
      "epoch": 3.368,
      "grad_norm": 0.7821294069290161,
      "learning_rate": 3.624680267016034e-05,
      "loss": 3.0743,
      "step": 12000
    },
    {
      "epoch": 3.368,
      "eval_accuracy": 0.41744863751592426,
      "eval_bleu": 0.1437616042304935,
      "eval_loss": 3.185511589050293,
      "eval_perplexity": 24.1796552890801,
      "eval_runtime": 20.1722,
      "eval_samples_per_second": 56.563,
      "eval_steps_per_second": 3.569,
      "step": 12000
    },
    {
      "epoch": 3.3960701754385965,
      "grad_norm": 0.7699295878410339,
      "learning_rate": 3.562293343315241e-05,
      "loss": 3.0603,
      "step": 12100
    },
    {
      "epoch": 3.424140350877193,
      "grad_norm": 0.7893390655517578,
      "learning_rate": 3.499906419614449e-05,
      "loss": 3.064,
      "step": 12200
    },
    {
      "epoch": 3.4522105263157896,
      "grad_norm": 0.7758604884147644,
      "learning_rate": 3.4375194959136565e-05,
      "loss": 3.0791,
      "step": 12300
    },
    {
      "epoch": 3.4802807017543858,
      "grad_norm": 0.7830843925476074,
      "learning_rate": 3.375132572212865e-05,
      "loss": 3.0691,
      "step": 12400
    },
    {
      "epoch": 3.5083508771929823,
      "grad_norm": 0.7860715389251709,
      "learning_rate": 3.312745648512072e-05,
      "loss": 3.0679,
      "step": 12500
    },
    {
      "epoch": 3.5083508771929823,
      "eval_accuracy": 0.4182539539753076,
      "eval_bleu": 0.13973142885022957,
      "eval_loss": 3.1735119819641113,
      "eval_perplexity": 23.891242805090002,
      "eval_runtime": 20.2985,
      "eval_samples_per_second": 56.211,
      "eval_steps_per_second": 3.547,
      "step": 12500
    },
    {
      "epoch": 3.536421052631579,
      "grad_norm": 0.7781902551651001,
      "learning_rate": 3.25035872481128e-05,
      "loss": 3.0577,
      "step": 12600
    },
    {
      "epoch": 3.5644912280701755,
      "grad_norm": 0.7784900665283203,
      "learning_rate": 3.1879718011104875e-05,
      "loss": 3.0775,
      "step": 12700
    },
    {
      "epoch": 3.592561403508772,
      "grad_norm": 0.8009820580482483,
      "learning_rate": 3.125584877409695e-05,
      "loss": 3.0532,
      "step": 12800
    },
    {
      "epoch": 3.6206315789473686,
      "grad_norm": 0.7675787210464478,
      "learning_rate": 3.0631979537089026e-05,
      "loss": 3.0597,
      "step": 12900
    },
    {
      "epoch": 3.648701754385965,
      "grad_norm": 0.7848743796348572,
      "learning_rate": 3.0008110300081106e-05,
      "loss": 3.0635,
      "step": 13000
    },
    {
      "epoch": 3.648701754385965,
      "eval_accuracy": 0.4199536857363891,
      "eval_bleu": 0.14234657853553717,
      "eval_loss": 3.1599371433258057,
      "eval_perplexity": 23.56911440636153,
      "eval_runtime": 20.1704,
      "eval_samples_per_second": 56.568,
      "eval_steps_per_second": 3.57,
      "step": 13000
    },
    {
      "epoch": 3.6767719298245614,
      "grad_norm": 0.7748121023178101,
      "learning_rate": 2.938424106307318e-05,
      "loss": 3.0489,
      "step": 13100
    },
    {
      "epoch": 3.704842105263158,
      "grad_norm": 0.7777696251869202,
      "learning_rate": 2.8760371826065257e-05,
      "loss": 3.0603,
      "step": 13200
    },
    {
      "epoch": 3.7329122807017545,
      "grad_norm": 0.7850595116615295,
      "learning_rate": 2.8136502589057333e-05,
      "loss": 3.0361,
      "step": 13300
    },
    {
      "epoch": 3.7609824561403506,
      "grad_norm": 0.7980267405509949,
      "learning_rate": 2.7512633352049412e-05,
      "loss": 3.0565,
      "step": 13400
    },
    {
      "epoch": 3.7890526315789472,
      "grad_norm": 0.8132106065750122,
      "learning_rate": 2.6888764115041488e-05,
      "loss": 3.0262,
      "step": 13500
    },
    {
      "epoch": 3.7890526315789472,
      "eval_accuracy": 0.4210717048635117,
      "eval_bleu": 0.14320922241168227,
      "eval_loss": 3.148859977722168,
      "eval_perplexity": 23.309476106050013,
      "eval_runtime": 20.2248,
      "eval_samples_per_second": 56.416,
      "eval_steps_per_second": 3.56,
      "step": 13500
    },
    {
      "epoch": 3.817122807017544,
      "grad_norm": 0.8002068400382996,
      "learning_rate": 2.6264894878033564e-05,
      "loss": 3.0504,
      "step": 13600
    },
    {
      "epoch": 3.8451929824561404,
      "grad_norm": 0.7945306301116943,
      "learning_rate": 2.564102564102564e-05,
      "loss": 3.0412,
      "step": 13700
    },
    {
      "epoch": 3.873263157894737,
      "grad_norm": 0.7735077738761902,
      "learning_rate": 2.5017156404017715e-05,
      "loss": 3.0193,
      "step": 13800
    },
    {
      "epoch": 3.9013333333333335,
      "grad_norm": 0.7823268175125122,
      "learning_rate": 2.4393287167009795e-05,
      "loss": 3.0408,
      "step": 13900
    },
    {
      "epoch": 3.9294035087719297,
      "grad_norm": 0.7943819165229797,
      "learning_rate": 2.376941793000187e-05,
      "loss": 3.0382,
      "step": 14000
    },
    {
      "epoch": 3.9294035087719297,
      "eval_accuracy": 0.4223199453755559,
      "eval_bleu": 0.1460635167412373,
      "eval_loss": 3.139704465866089,
      "eval_perplexity": 23.097039886291473,
      "eval_runtime": 20.2485,
      "eval_samples_per_second": 56.35,
      "eval_steps_per_second": 3.556,
      "step": 14000
    },
    {
      "epoch": 3.9574736842105263,
      "grad_norm": 0.7909451723098755,
      "learning_rate": 2.314554869299395e-05,
      "loss": 3.032,
      "step": 14100
    },
    {
      "epoch": 3.985543859649123,
      "grad_norm": 0.7830550670623779,
      "learning_rate": 2.2521679455986026e-05,
      "loss": 3.0383,
      "step": 14200
    },
    {
      "epoch": 4.013473684210527,
      "grad_norm": 0.7939472794532776,
      "learning_rate": 2.1897810218978105e-05,
      "loss": 2.9744,
      "step": 14300
    },
    {
      "epoch": 4.041543859649122,
      "grad_norm": 0.8031191229820251,
      "learning_rate": 2.127394098197018e-05,
      "loss": 2.9426,
      "step": 14400
    },
    {
      "epoch": 4.069614035087719,
      "grad_norm": 0.8071146607398987,
      "learning_rate": 2.0650071744962256e-05,
      "loss": 2.9525,
      "step": 14500
    },
    {
      "epoch": 4.069614035087719,
      "eval_accuracy": 0.4232880385660912,
      "eval_bleu": 0.1457336973656826,
      "eval_loss": 3.1334879398345947,
      "eval_perplexity": 22.95390190803781,
      "eval_runtime": 20.1672,
      "eval_samples_per_second": 56.577,
      "eval_steps_per_second": 3.57,
      "step": 14500
    },
    {
      "epoch": 4.097684210526316,
      "grad_norm": 0.8192425966262817,
      "learning_rate": 2.0026202507954332e-05,
      "loss": 2.9449,
      "step": 14600
    },
    {
      "epoch": 4.125754385964912,
      "grad_norm": 0.8179590702056885,
      "learning_rate": 1.940233327094641e-05,
      "loss": 2.9534,
      "step": 14700
    },
    {
      "epoch": 4.153824561403509,
      "grad_norm": 0.8264966607093811,
      "learning_rate": 1.8778464033938487e-05,
      "loss": 2.957,
      "step": 14800
    },
    {
      "epoch": 4.181894736842105,
      "grad_norm": 0.8411971926689148,
      "learning_rate": 1.8154594796930563e-05,
      "loss": 2.9503,
      "step": 14900
    },
    {
      "epoch": 4.209964912280702,
      "grad_norm": 0.818305253982544,
      "learning_rate": 1.7530725559922642e-05,
      "loss": 2.9621,
      "step": 15000
    },
    {
      "epoch": 4.209964912280702,
      "eval_accuracy": 0.4239245812568591,
      "eval_bleu": 0.14536464175084168,
      "eval_loss": 3.1270124912261963,
      "eval_perplexity": 22.805745303809648,
      "eval_runtime": 12.3884,
      "eval_samples_per_second": 92.102,
      "eval_steps_per_second": 5.812,
      "step": 15000
    },
    {
      "epoch": 4.2380350877192985,
      "grad_norm": 0.8167079091072083,
      "learning_rate": 1.6906856322914718e-05,
      "loss": 2.944,
      "step": 15100
    },
    {
      "epoch": 4.266105263157895,
      "grad_norm": 0.837181031703949,
      "learning_rate": 1.6282987085906794e-05,
      "loss": 2.9407,
      "step": 15200
    },
    {
      "epoch": 4.294175438596492,
      "grad_norm": 0.8356810808181763,
      "learning_rate": 1.5659117848898873e-05,
      "loss": 2.9364,
      "step": 15300
    },
    {
      "epoch": 4.322245614035087,
      "grad_norm": 0.8358649015426636,
      "learning_rate": 1.5035248611890947e-05,
      "loss": 2.9307,
      "step": 15400
    },
    {
      "epoch": 4.350315789473684,
      "grad_norm": 0.842452347278595,
      "learning_rate": 1.4411379374883025e-05,
      "loss": 2.9422,
      "step": 15500
    },
    {
      "epoch": 4.350315789473684,
      "eval_accuracy": 0.42498777032717266,
      "eval_bleu": 0.1467649753048137,
      "eval_loss": 3.1211211681365967,
      "eval_perplexity": 22.671784281214766,
      "eval_runtime": 20.2024,
      "eval_samples_per_second": 56.479,
      "eval_steps_per_second": 3.564,
      "step": 15500
    },
    {
      "epoch": 4.3783859649122805,
      "grad_norm": 0.8464500904083252,
      "learning_rate": 1.37875101378751e-05,
      "loss": 2.9435,
      "step": 15600
    },
    {
      "epoch": 4.406456140350877,
      "grad_norm": 0.8314975500106812,
      "learning_rate": 1.316364090086718e-05,
      "loss": 2.9405,
      "step": 15700
    },
    {
      "epoch": 4.434526315789474,
      "grad_norm": 0.8358827829360962,
      "learning_rate": 1.2539771663859257e-05,
      "loss": 2.9438,
      "step": 15800
    },
    {
      "epoch": 4.46259649122807,
      "grad_norm": 0.8394728899002075,
      "learning_rate": 1.1915902426851333e-05,
      "loss": 2.9233,
      "step": 15900
    },
    {
      "epoch": 4.490666666666667,
      "grad_norm": 0.839545488357544,
      "learning_rate": 1.1292033189843409e-05,
      "loss": 2.9224,
      "step": 16000
    },
    {
      "epoch": 4.490666666666667,
      "eval_accuracy": 0.4256748594765614,
      "eval_bleu": 0.1454115201862865,
      "eval_loss": 3.1149473190307617,
      "eval_perplexity": 22.53224330181115,
      "eval_runtime": 20.2018,
      "eval_samples_per_second": 56.48,
      "eval_steps_per_second": 3.564,
      "step": 16000
    },
    {
      "epoch": 4.518736842105263,
      "grad_norm": 0.847416341304779,
      "learning_rate": 1.0668163952835486e-05,
      "loss": 2.9372,
      "step": 16100
    },
    {
      "epoch": 4.54680701754386,
      "grad_norm": 0.8180661201477051,
      "learning_rate": 1.0044294715827562e-05,
      "loss": 2.9422,
      "step": 16200
    },
    {
      "epoch": 4.574877192982456,
      "grad_norm": 0.8318551182746887,
      "learning_rate": 9.42042547881964e-06,
      "loss": 2.9373,
      "step": 16300
    },
    {
      "epoch": 4.602947368421052,
      "grad_norm": 0.8445524573326111,
      "learning_rate": 8.796556241811717e-06,
      "loss": 2.9212,
      "step": 16400
    },
    {
      "epoch": 4.631017543859649,
      "grad_norm": 0.8220964670181274,
      "learning_rate": 8.172687004803793e-06,
      "loss": 2.9475,
      "step": 16500
    },
    {
      "epoch": 4.631017543859649,
      "eval_accuracy": 0.4264193488416722,
      "eval_bleu": 0.149677834207122,
      "eval_loss": 3.1084439754486084,
      "eval_perplexity": 22.38618383417341,
      "eval_runtime": 20.2407,
      "eval_samples_per_second": 56.372,
      "eval_steps_per_second": 3.557,
      "step": 16500
    },
    {
      "epoch": 4.659087719298245,
      "grad_norm": 0.8326128721237183,
      "learning_rate": 7.548817767795871e-06,
      "loss": 2.9335,
      "step": 16600
    },
    {
      "epoch": 4.687157894736842,
      "grad_norm": 0.822635293006897,
      "learning_rate": 6.924948530787947e-06,
      "loss": 2.9389,
      "step": 16700
    },
    {
      "epoch": 4.7152280701754385,
      "grad_norm": 0.8398991227149963,
      "learning_rate": 6.301079293780024e-06,
      "loss": 2.9219,
      "step": 16800
    },
    {
      "epoch": 4.743298245614035,
      "grad_norm": 0.8308337330818176,
      "learning_rate": 5.677210056772101e-06,
      "loss": 2.9374,
      "step": 16900
    },
    {
      "epoch": 4.771368421052632,
      "grad_norm": 0.8378835320472717,
      "learning_rate": 5.053340819764177e-06,
      "loss": 2.9318,
      "step": 17000
    },
    {
      "epoch": 4.771368421052632,
      "eval_accuracy": 0.4270327601022238,
      "eval_bleu": 0.14678838838741357,
      "eval_loss": 3.104135036468506,
      "eval_perplexity": 22.289930657494395,
      "eval_runtime": 10.9973,
      "eval_samples_per_second": 103.753,
      "eval_steps_per_second": 6.547,
      "step": 17000
    },
    {
      "epoch": 4.799438596491228,
      "grad_norm": 0.8561663031578064,
      "learning_rate": 4.429471582756255e-06,
      "loss": 2.9288,
      "step": 17100
    },
    {
      "epoch": 4.827508771929825,
      "grad_norm": 0.8296898007392883,
      "learning_rate": 3.805602345748331e-06,
      "loss": 2.9205,
      "step": 17200
    },
    {
      "epoch": 4.855578947368421,
      "grad_norm": 0.8513033390045166,
      "learning_rate": 3.1817331087404082e-06,
      "loss": 2.9328,
      "step": 17300
    },
    {
      "epoch": 4.883649122807018,
      "grad_norm": 0.8511661887168884,
      "learning_rate": 2.557863871732485e-06,
      "loss": 2.9392,
      "step": 17400
    },
    {
      "epoch": 4.911719298245614,
      "grad_norm": 0.8221142292022705,
      "learning_rate": 1.933994634724562e-06,
      "loss": 2.9268,
      "step": 17500
    },
    {
      "epoch": 4.911719298245614,
      "eval_accuracy": 0.4273857285929322,
      "eval_bleu": 0.1460515137764717,
      "eval_loss": 3.101016044616699,
      "eval_perplexity": 22.220516852214956,
      "eval_runtime": 10.9797,
      "eval_samples_per_second": 103.919,
      "eval_steps_per_second": 6.558,
      "step": 17500
    },
    {
      "epoch": 4.93978947368421,
      "grad_norm": 0.8456584811210632,
      "learning_rate": 1.3101253977166387e-06,
      "loss": 2.9326,
      "step": 17600
    },
    {
      "epoch": 4.967859649122807,
      "grad_norm": 0.8384252786636353,
      "learning_rate": 6.862561607087155e-07,
      "loss": 2.9198,
      "step": 17700
    },
    {
      "epoch": 4.995929824561403,
      "grad_norm": 0.8392898440361023,
      "learning_rate": 6.238692370079231e-08,
      "loss": 2.9187,
      "step": 17800
    },
    {
      "epoch": 4.998736842105263,
      "step": 17810,
      "total_flos": 1.0584067483285586e+18,
      "train_loss": 3.584141966119902,
      "train_runtime": 28526.973,
      "train_samples_per_second": 19.98,
      "train_steps_per_second": 0.624
    }
  ],
  "logging_steps": 100,
  "max_steps": 17810,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 2,
        "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": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.0584067483285586e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}