{
  "best_metric": 0.03436192870140076,
  "best_model_checkpoint": "saves/psy-course/Llama-3.1-8B-Instruct/train/fold4/checkpoint-1300",
  "epoch": 4.999522946283752,
  "eval_steps": 50,
  "global_step": 3275,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.015265718919950386,
      "grad_norm": 4.221195220947266,
      "learning_rate": 3.0487804878048782e-06,
      "loss": 1.6278,
      "step": 10
    },
    {
      "epoch": 0.030531437839900772,
      "grad_norm": 5.125793933868408,
      "learning_rate": 6.0975609756097564e-06,
      "loss": 1.5554,
      "step": 20
    },
    {
      "epoch": 0.04579715675985116,
      "grad_norm": 6.733878135681152,
      "learning_rate": 9.146341463414634e-06,
      "loss": 1.3738,
      "step": 30
    },
    {
      "epoch": 0.061062875679801544,
      "grad_norm": 2.5405797958374023,
      "learning_rate": 1.2195121951219513e-05,
      "loss": 1.0174,
      "step": 40
    },
    {
      "epoch": 0.07632859459975193,
      "grad_norm": 1.4538228511810303,
      "learning_rate": 1.524390243902439e-05,
      "loss": 0.5681,
      "step": 50
    },
    {
      "epoch": 0.07632859459975193,
      "eval_loss": 0.3813890814781189,
      "eval_runtime": 158.8708,
      "eval_samples_per_second": 7.333,
      "eval_steps_per_second": 7.333,
      "step": 50
    },
    {
      "epoch": 0.09159431351970232,
      "grad_norm": 1.5504209995269775,
      "learning_rate": 1.8292682926829268e-05,
      "loss": 0.3362,
      "step": 60
    },
    {
      "epoch": 0.10686003243965271,
      "grad_norm": 1.2897658348083496,
      "learning_rate": 2.134146341463415e-05,
      "loss": 0.246,
      "step": 70
    },
    {
      "epoch": 0.12212575135960309,
      "grad_norm": 1.0303564071655273,
      "learning_rate": 2.4390243902439026e-05,
      "loss": 0.1841,
      "step": 80
    },
    {
      "epoch": 0.13739147027955348,
      "grad_norm": 0.9430064558982849,
      "learning_rate": 2.7439024390243906e-05,
      "loss": 0.115,
      "step": 90
    },
    {
      "epoch": 0.15265718919950386,
      "grad_norm": 0.9477012157440186,
      "learning_rate": 3.048780487804878e-05,
      "loss": 0.122,
      "step": 100
    },
    {
      "epoch": 0.15265718919950386,
      "eval_loss": 0.07960689812898636,
      "eval_runtime": 158.7007,
      "eval_samples_per_second": 7.341,
      "eval_steps_per_second": 7.341,
      "step": 100
    },
    {
      "epoch": 0.16792290811945426,
      "grad_norm": 0.690284788608551,
      "learning_rate": 3.353658536585366e-05,
      "loss": 0.0802,
      "step": 110
    },
    {
      "epoch": 0.18318862703940464,
      "grad_norm": 0.651168704032898,
      "learning_rate": 3.6585365853658535e-05,
      "loss": 0.0705,
      "step": 120
    },
    {
      "epoch": 0.19845434595935502,
      "grad_norm": 0.766960084438324,
      "learning_rate": 3.9634146341463416e-05,
      "loss": 0.08,
      "step": 130
    },
    {
      "epoch": 0.21372006487930542,
      "grad_norm": 0.49084338545799255,
      "learning_rate": 4.26829268292683e-05,
      "loss": 0.0655,
      "step": 140
    },
    {
      "epoch": 0.2289857837992558,
      "grad_norm": 0.6232007145881653,
      "learning_rate": 4.573170731707318e-05,
      "loss": 0.0541,
      "step": 150
    },
    {
      "epoch": 0.2289857837992558,
      "eval_loss": 0.059219248592853546,
      "eval_runtime": 158.6529,
      "eval_samples_per_second": 7.343,
      "eval_steps_per_second": 7.343,
      "step": 150
    },
    {
      "epoch": 0.24425150271920618,
      "grad_norm": 0.6613311767578125,
      "learning_rate": 4.878048780487805e-05,
      "loss": 0.066,
      "step": 160
    },
    {
      "epoch": 0.2595172216391566,
      "grad_norm": 1.2490918636322021,
      "learning_rate": 5.182926829268293e-05,
      "loss": 0.0751,
      "step": 170
    },
    {
      "epoch": 0.27478294055910696,
      "grad_norm": 0.5701612234115601,
      "learning_rate": 5.487804878048781e-05,
      "loss": 0.0772,
      "step": 180
    },
    {
      "epoch": 0.29004865947905734,
      "grad_norm": 0.8585615754127502,
      "learning_rate": 5.792682926829268e-05,
      "loss": 0.0687,
      "step": 190
    },
    {
      "epoch": 0.3053143783990077,
      "grad_norm": 0.5563382506370544,
      "learning_rate": 6.097560975609756e-05,
      "loss": 0.0502,
      "step": 200
    },
    {
      "epoch": 0.3053143783990077,
      "eval_loss": 0.05221959203481674,
      "eval_runtime": 158.6341,
      "eval_samples_per_second": 7.344,
      "eval_steps_per_second": 7.344,
      "step": 200
    },
    {
      "epoch": 0.3205800973189581,
      "grad_norm": 0.6384761929512024,
      "learning_rate": 6.402439024390244e-05,
      "loss": 0.0572,
      "step": 210
    },
    {
      "epoch": 0.3358458162389085,
      "grad_norm": 0.5302607417106628,
      "learning_rate": 6.707317073170732e-05,
      "loss": 0.0493,
      "step": 220
    },
    {
      "epoch": 0.3511115351588589,
      "grad_norm": 0.7148872017860413,
      "learning_rate": 7.012195121951219e-05,
      "loss": 0.0421,
      "step": 230
    },
    {
      "epoch": 0.3663772540788093,
      "grad_norm": 0.6102215051651001,
      "learning_rate": 7.317073170731707e-05,
      "loss": 0.0603,
      "step": 240
    },
    {
      "epoch": 0.38164297299875966,
      "grad_norm": 0.552837610244751,
      "learning_rate": 7.621951219512195e-05,
      "loss": 0.0666,
      "step": 250
    },
    {
      "epoch": 0.38164297299875966,
      "eval_loss": 0.05392138659954071,
      "eval_runtime": 158.5738,
      "eval_samples_per_second": 7.347,
      "eval_steps_per_second": 7.347,
      "step": 250
    },
    {
      "epoch": 0.39690869191871003,
      "grad_norm": 0.6186545491218567,
      "learning_rate": 7.926829268292683e-05,
      "loss": 0.0653,
      "step": 260
    },
    {
      "epoch": 0.4121744108386604,
      "grad_norm": 0.5888819098472595,
      "learning_rate": 8.231707317073171e-05,
      "loss": 0.0563,
      "step": 270
    },
    {
      "epoch": 0.42744012975861084,
      "grad_norm": 0.36493584513664246,
      "learning_rate": 8.53658536585366e-05,
      "loss": 0.0576,
      "step": 280
    },
    {
      "epoch": 0.4427058486785612,
      "grad_norm": 0.506234347820282,
      "learning_rate": 8.841463414634147e-05,
      "loss": 0.0602,
      "step": 290
    },
    {
      "epoch": 0.4579715675985116,
      "grad_norm": 0.5001620650291443,
      "learning_rate": 9.146341463414635e-05,
      "loss": 0.046,
      "step": 300
    },
    {
      "epoch": 0.4579715675985116,
      "eval_loss": 0.049273185431957245,
      "eval_runtime": 158.5042,
      "eval_samples_per_second": 7.35,
      "eval_steps_per_second": 7.35,
      "step": 300
    },
    {
      "epoch": 0.473237286518462,
      "grad_norm": 0.56379634141922,
      "learning_rate": 9.451219512195122e-05,
      "loss": 0.0649,
      "step": 310
    },
    {
      "epoch": 0.48850300543841235,
      "grad_norm": 0.4352799355983734,
      "learning_rate": 9.75609756097561e-05,
      "loss": 0.0436,
      "step": 320
    },
    {
      "epoch": 0.5037687243583627,
      "grad_norm": 0.6286426186561584,
      "learning_rate": 9.999988635788465e-05,
      "loss": 0.0507,
      "step": 330
    },
    {
      "epoch": 0.5190344432783132,
      "grad_norm": 0.7499465346336365,
      "learning_rate": 9.999590893808788e-05,
      "loss": 0.0449,
      "step": 340
    },
    {
      "epoch": 0.5343001621982635,
      "grad_norm": 0.6524621248245239,
      "learning_rate": 9.998624992909386e-05,
      "loss": 0.0458,
      "step": 350
    },
    {
      "epoch": 0.5343001621982635,
      "eval_loss": 0.05271424353122711,
      "eval_runtime": 158.5134,
      "eval_samples_per_second": 7.35,
      "eval_steps_per_second": 7.35,
      "step": 350
    },
    {
      "epoch": 0.5495658811182139,
      "grad_norm": 0.3585219979286194,
      "learning_rate": 9.997091042856284e-05,
      "loss": 0.0488,
      "step": 360
    },
    {
      "epoch": 0.5648316000381643,
      "grad_norm": 0.2826538681983948,
      "learning_rate": 9.994989217969224e-05,
      "loss": 0.0646,
      "step": 370
    },
    {
      "epoch": 0.5800973189581147,
      "grad_norm": 1.2497502565383911,
      "learning_rate": 9.992319757101863e-05,
      "loss": 0.0588,
      "step": 380
    },
    {
      "epoch": 0.5953630378780651,
      "grad_norm": 0.4816202223300934,
      "learning_rate": 9.98908296361462e-05,
      "loss": 0.0586,
      "step": 390
    },
    {
      "epoch": 0.6106287567980154,
      "grad_norm": 0.349490225315094,
      "learning_rate": 9.98527920534021e-05,
      "loss": 0.0448,
      "step": 400
    },
    {
      "epoch": 0.6106287567980154,
      "eval_loss": 0.048783060163259506,
      "eval_runtime": 157.7617,
      "eval_samples_per_second": 7.385,
      "eval_steps_per_second": 7.385,
      "step": 400
    },
    {
      "epoch": 0.6258944757179659,
      "grad_norm": 0.6848151683807373,
      "learning_rate": 9.980908914541844e-05,
      "loss": 0.054,
      "step": 410
    },
    {
      "epoch": 0.6411601946379162,
      "grad_norm": 0.5100162625312805,
      "learning_rate": 9.975972587864095e-05,
      "loss": 0.0506,
      "step": 420
    },
    {
      "epoch": 0.6564259135578666,
      "grad_norm": 0.41106978058815,
      "learning_rate": 9.970470786276467e-05,
      "loss": 0.0369,
      "step": 430
    },
    {
      "epoch": 0.671691632477817,
      "grad_norm": 0.29269176721572876,
      "learning_rate": 9.964404135009648e-05,
      "loss": 0.0527,
      "step": 440
    },
    {
      "epoch": 0.6869573513977674,
      "grad_norm": 0.23321561515331268,
      "learning_rate": 9.957773323484454e-05,
      "loss": 0.0567,
      "step": 450
    },
    {
      "epoch": 0.6869573513977674,
      "eval_loss": 0.04618264362215996,
      "eval_runtime": 157.632,
      "eval_samples_per_second": 7.391,
      "eval_steps_per_second": 7.391,
      "step": 450
    },
    {
      "epoch": 0.7022230703177178,
      "grad_norm": 0.22461284697055817,
      "learning_rate": 9.950579105233483e-05,
      "loss": 0.04,
      "step": 460
    },
    {
      "epoch": 0.7174887892376681,
      "grad_norm": 0.41304975748062134,
      "learning_rate": 9.94282229781548e-05,
      "loss": 0.0383,
      "step": 470
    },
    {
      "epoch": 0.7327545081576186,
      "grad_norm": 0.8002622723579407,
      "learning_rate": 9.934503782722438e-05,
      "loss": 0.0495,
      "step": 480
    },
    {
      "epoch": 0.748020227077569,
      "grad_norm": 0.27543845772743225,
      "learning_rate": 9.925624505279411e-05,
      "loss": 0.0379,
      "step": 490
    },
    {
      "epoch": 0.7632859459975193,
      "grad_norm": 0.19786208868026733,
      "learning_rate": 9.916185474537098e-05,
      "loss": 0.0358,
      "step": 500
    },
    {
      "epoch": 0.7632859459975193,
      "eval_loss": 0.04097205400466919,
      "eval_runtime": 157.5197,
      "eval_samples_per_second": 7.396,
      "eval_steps_per_second": 7.396,
      "step": 500
    },
    {
      "epoch": 0.7785516649174697,
      "grad_norm": 0.7004953622817993,
      "learning_rate": 9.906187763157168e-05,
      "loss": 0.0508,
      "step": 510
    },
    {
      "epoch": 0.7938173838374201,
      "grad_norm": 0.28915488719940186,
      "learning_rate": 9.895632507290362e-05,
      "loss": 0.05,
      "step": 520
    },
    {
      "epoch": 0.8090831027573705,
      "grad_norm": 0.38373371958732605,
      "learning_rate": 9.884520906447379e-05,
      "loss": 0.0376,
      "step": 530
    },
    {
      "epoch": 0.8243488216773208,
      "grad_norm": 0.5942443609237671,
      "learning_rate": 9.872854223362562e-05,
      "loss": 0.0482,
      "step": 540
    },
    {
      "epoch": 0.8396145405972713,
      "grad_norm": 0.44209200143814087,
      "learning_rate": 9.860633783850406e-05,
      "loss": 0.0445,
      "step": 550
    },
    {
      "epoch": 0.8396145405972713,
      "eval_loss": 0.04072929546236992,
      "eval_runtime": 157.4739,
      "eval_samples_per_second": 7.398,
      "eval_steps_per_second": 7.398,
      "step": 550
    },
    {
      "epoch": 0.8548802595172217,
      "grad_norm": 0.14016468822956085,
      "learning_rate": 9.847860976654879e-05,
      "loss": 0.0501,
      "step": 560
    },
    {
      "epoch": 0.870145978437172,
      "grad_norm": 0.32843533158302307,
      "learning_rate": 9.834537253291616e-05,
      "loss": 0.0401,
      "step": 570
    },
    {
      "epoch": 0.8854116973571224,
      "grad_norm": 0.7663369178771973,
      "learning_rate": 9.820664127882957e-05,
      "loss": 0.0582,
      "step": 580
    },
    {
      "epoch": 0.9006774162770728,
      "grad_norm": 0.40339088439941406,
      "learning_rate": 9.806243176985888e-05,
      "loss": 0.0494,
      "step": 590
    },
    {
      "epoch": 0.9159431351970232,
      "grad_norm": 0.4325726628303528,
      "learning_rate": 9.791276039412875e-05,
      "loss": 0.0462,
      "step": 600
    },
    {
      "epoch": 0.9159431351970232,
      "eval_loss": 0.040742386132478714,
      "eval_runtime": 157.4061,
      "eval_samples_per_second": 7.401,
      "eval_steps_per_second": 7.401,
      "step": 600
    },
    {
      "epoch": 0.9312088541169735,
      "grad_norm": 0.39752867817878723,
      "learning_rate": 9.775764416045628e-05,
      "loss": 0.0595,
      "step": 610
    },
    {
      "epoch": 0.946474573036924,
      "grad_norm": 0.37356245517730713,
      "learning_rate": 9.759710069641814e-05,
      "loss": 0.0463,
      "step": 620
    },
    {
      "epoch": 0.9617402919568744,
      "grad_norm": 0.3723120391368866,
      "learning_rate": 9.743114824634734e-05,
      "loss": 0.033,
      "step": 630
    },
    {
      "epoch": 0.9770060108768247,
      "grad_norm": 0.4107719659805298,
      "learning_rate": 9.725980566925989e-05,
      "loss": 0.0497,
      "step": 640
    },
    {
      "epoch": 0.9922717297967751,
      "grad_norm": 0.31700581312179565,
      "learning_rate": 9.708309243671165e-05,
      "loss": 0.0363,
      "step": 650
    },
    {
      "epoch": 0.9922717297967751,
      "eval_loss": 0.040968675166368484,
      "eval_runtime": 157.3806,
      "eval_samples_per_second": 7.402,
      "eval_steps_per_second": 7.402,
      "step": 650
    },
    {
      "epoch": 1.0075374487167255,
      "grad_norm": 0.18884813785552979,
      "learning_rate": 9.690102863058563e-05,
      "loss": 0.0343,
      "step": 660
    },
    {
      "epoch": 1.0228031676366758,
      "grad_norm": 0.3541209399700165,
      "learning_rate": 9.67136349408098e-05,
      "loss": 0.0333,
      "step": 670
    },
    {
      "epoch": 1.0380688865566263,
      "grad_norm": 0.17939338088035583,
      "learning_rate": 9.652093266300583e-05,
      "loss": 0.0401,
      "step": 680
    },
    {
      "epoch": 1.0533346054765766,
      "grad_norm": 0.20926208794116974,
      "learning_rate": 9.632294369606916e-05,
      "loss": 0.0271,
      "step": 690
    },
    {
      "epoch": 1.068600324396527,
      "grad_norm": 0.2087617814540863,
      "learning_rate": 9.61196905396802e-05,
      "loss": 0.0343,
      "step": 700
    },
    {
      "epoch": 1.068600324396527,
      "eval_loss": 0.03695390745997429,
      "eval_runtime": 157.4136,
      "eval_samples_per_second": 7.401,
      "eval_steps_per_second": 7.401,
      "step": 700
    },
    {
      "epoch": 1.0838660433164775,
      "grad_norm": 0.46884167194366455,
      "learning_rate": 9.591119629174764e-05,
      "loss": 0.0442,
      "step": 710
    },
    {
      "epoch": 1.0991317622364278,
      "grad_norm": 0.2890946865081787,
      "learning_rate": 9.569748464578343e-05,
      "loss": 0.0288,
      "step": 720
    },
    {
      "epoch": 1.1143974811563782,
      "grad_norm": 0.20250333845615387,
      "learning_rate": 9.54785798882103e-05,
      "loss": 0.0273,
      "step": 730
    },
    {
      "epoch": 1.1296632000763287,
      "grad_norm": 0.26456066966056824,
      "learning_rate": 9.525450689560181e-05,
      "loss": 0.0378,
      "step": 740
    },
    {
      "epoch": 1.144928918996279,
      "grad_norm": 0.3689366281032562,
      "learning_rate": 9.502529113185532e-05,
      "loss": 0.0413,
      "step": 750
    },
    {
      "epoch": 1.144928918996279,
      "eval_loss": 0.03775772452354431,
      "eval_runtime": 157.6202,
      "eval_samples_per_second": 7.391,
      "eval_steps_per_second": 7.391,
      "step": 750
    },
    {
      "epoch": 1.1601946379162293,
      "grad_norm": 0.21495144069194794,
      "learning_rate": 9.479095864529828e-05,
      "loss": 0.0308,
      "step": 760
    },
    {
      "epoch": 1.1754603568361797,
      "grad_norm": 0.17624302208423615,
      "learning_rate": 9.455153606572806e-05,
      "loss": 0.0316,
      "step": 770
    },
    {
      "epoch": 1.1907260757561302,
      "grad_norm": 0.26850008964538574,
      "learning_rate": 9.430705060138569e-05,
      "loss": 0.0348,
      "step": 780
    },
    {
      "epoch": 1.2059917946760805,
      "grad_norm": 0.633297860622406,
      "learning_rate": 9.405753003586395e-05,
      "loss": 0.0277,
      "step": 790
    },
    {
      "epoch": 1.2212575135960309,
      "grad_norm": 0.2546226978302002,
      "learning_rate": 9.38030027249499e-05,
      "loss": 0.0322,
      "step": 800
    },
    {
      "epoch": 1.2212575135960309,
      "eval_loss": 0.03979082778096199,
      "eval_runtime": 157.5322,
      "eval_samples_per_second": 7.395,
      "eval_steps_per_second": 7.395,
      "step": 800
    },
    {
      "epoch": 1.2365232325159814,
      "grad_norm": 0.3185233175754547,
      "learning_rate": 9.354349759340263e-05,
      "loss": 0.0372,
      "step": 810
    },
    {
      "epoch": 1.2517889514359317,
      "grad_norm": 0.32911208271980286,
      "learning_rate": 9.327904413166615e-05,
      "loss": 0.0338,
      "step": 820
    },
    {
      "epoch": 1.267054670355882,
      "grad_norm": 0.5364266037940979,
      "learning_rate": 9.300967239251798e-05,
      "loss": 0.0405,
      "step": 830
    },
    {
      "epoch": 1.2823203892758324,
      "grad_norm": 0.1836143583059311,
      "learning_rate": 9.27354129876541e-05,
      "loss": 0.0348,
      "step": 840
    },
    {
      "epoch": 1.297586108195783,
      "grad_norm": 0.18749667704105377,
      "learning_rate": 9.245629708421008e-05,
      "loss": 0.0342,
      "step": 850
    },
    {
      "epoch": 1.297586108195783,
      "eval_loss": 0.03846587613224983,
      "eval_runtime": 157.7572,
      "eval_samples_per_second": 7.385,
      "eval_steps_per_second": 7.385,
      "step": 850
    },
    {
      "epoch": 1.3128518271157332,
      "grad_norm": 0.473058819770813,
      "learning_rate": 9.217235640121926e-05,
      "loss": 0.0379,
      "step": 860
    },
    {
      "epoch": 1.3281175460356835,
      "grad_norm": 0.5135864019393921,
      "learning_rate": 9.188362320600812e-05,
      "loss": 0.0425,
      "step": 870
    },
    {
      "epoch": 1.343383264955634,
      "grad_norm": 0.46343910694122314,
      "learning_rate": 9.159013031052943e-05,
      "loss": 0.0324,
      "step": 880
    },
    {
      "epoch": 1.3586489838755844,
      "grad_norm": 0.3623945116996765,
      "learning_rate": 9.129191106763346e-05,
      "loss": 0.0342,
      "step": 890
    },
    {
      "epoch": 1.3739147027955347,
      "grad_norm": 0.3033890724182129,
      "learning_rate": 9.098899936727771e-05,
      "loss": 0.0337,
      "step": 900
    },
    {
      "epoch": 1.3739147027955347,
      "eval_loss": 0.043599650263786316,
      "eval_runtime": 157.7841,
      "eval_samples_per_second": 7.384,
      "eval_steps_per_second": 7.384,
      "step": 900
    },
    {
      "epoch": 1.3891804217154853,
      "grad_norm": 0.4166663587093353,
      "learning_rate": 9.068142963267558e-05,
      "loss": 0.029,
      "step": 910
    },
    {
      "epoch": 1.4044461406354356,
      "grad_norm": 0.273224413394928,
      "learning_rate": 9.036923681638463e-05,
      "loss": 0.0337,
      "step": 920
    },
    {
      "epoch": 1.419711859555386,
      "grad_norm": 0.2528637647628784,
      "learning_rate": 9.00524563963343e-05,
      "loss": 0.0269,
      "step": 930
    },
    {
      "epoch": 1.4349775784753362,
      "grad_norm": 0.3691759407520294,
      "learning_rate": 8.973112437179436e-05,
      "loss": 0.0372,
      "step": 940
    },
    {
      "epoch": 1.4502432973952868,
      "grad_norm": 0.30305129289627075,
      "learning_rate": 8.940527725928383e-05,
      "loss": 0.0295,
      "step": 950
    },
    {
      "epoch": 1.4502432973952868,
      "eval_loss": 0.037314947694540024,
      "eval_runtime": 157.8345,
      "eval_samples_per_second": 7.381,
      "eval_steps_per_second": 7.381,
      "step": 950
    },
    {
      "epoch": 1.4655090163152371,
      "grad_norm": 0.26650476455688477,
      "learning_rate": 8.90749520884212e-05,
      "loss": 0.0324,
      "step": 960
    },
    {
      "epoch": 1.4807747352351874,
      "grad_norm": 0.34022459387779236,
      "learning_rate": 8.874018639771637e-05,
      "loss": 0.0357,
      "step": 970
    },
    {
      "epoch": 1.4960404541551378,
      "grad_norm": 0.22096742689609528,
      "learning_rate": 8.840101823030471e-05,
      "loss": 0.0276,
      "step": 980
    },
    {
      "epoch": 1.5113061730750883,
      "grad_norm": 0.3219353258609772,
      "learning_rate": 8.805748612962382e-05,
      "loss": 0.0381,
      "step": 990
    },
    {
      "epoch": 1.5265718919950386,
      "grad_norm": 0.12673026323318481,
      "learning_rate": 8.77096291350334e-05,
      "loss": 0.0267,
      "step": 1000
    },
    {
      "epoch": 1.5265718919950386,
      "eval_loss": 0.03864908590912819,
      "eval_runtime": 158.4803,
      "eval_samples_per_second": 7.351,
      "eval_steps_per_second": 7.351,
      "step": 1000
    },
    {
      "epoch": 1.541837610914989,
      "grad_norm": 0.2603205442428589,
      "learning_rate": 8.735748677737874e-05,
      "loss": 0.0346,
      "step": 1010
    },
    {
      "epoch": 1.5571033298349395,
      "grad_norm": 0.3782278001308441,
      "learning_rate": 8.700109907449845e-05,
      "loss": 0.0296,
      "step": 1020
    },
    {
      "epoch": 1.5723690487548898,
      "grad_norm": 0.18346253037452698,
      "learning_rate": 8.66405065266768e-05,
      "loss": 0.0311,
      "step": 1030
    },
    {
      "epoch": 1.5876347676748401,
      "grad_norm": 0.20292475819587708,
      "learning_rate": 8.627575011204115e-05,
      "loss": 0.0281,
      "step": 1040
    },
    {
      "epoch": 1.6029004865947907,
      "grad_norm": 0.17125393450260162,
      "learning_rate": 8.590687128190516e-05,
      "loss": 0.0287,
      "step": 1050
    },
    {
      "epoch": 1.6029004865947907,
      "eval_loss": 0.03797018155455589,
      "eval_runtime": 174.0303,
      "eval_samples_per_second": 6.694,
      "eval_steps_per_second": 6.694,
      "step": 1050
    },
    {
      "epoch": 1.618166205514741,
      "grad_norm": 0.10430210083723068,
      "learning_rate": 8.553391195605833e-05,
      "loss": 0.0273,
      "step": 1060
    },
    {
      "epoch": 1.6334319244346913,
      "grad_norm": 0.5334872007369995,
      "learning_rate": 8.515691451800205e-05,
      "loss": 0.0469,
      "step": 1070
    },
    {
      "epoch": 1.6486976433546419,
      "grad_norm": 0.2953313887119293,
      "learning_rate": 8.477592181013316e-05,
      "loss": 0.0291,
      "step": 1080
    },
    {
      "epoch": 1.663963362274592,
      "grad_norm": 0.19991376996040344,
      "learning_rate": 8.439097712887531e-05,
      "loss": 0.0297,
      "step": 1090
    },
    {
      "epoch": 1.6792290811945425,
      "grad_norm": 0.37268176674842834,
      "learning_rate": 8.400212421975865e-05,
      "loss": 0.0504,
      "step": 1100
    },
    {
      "epoch": 1.6792290811945425,
      "eval_loss": 0.03881072625517845,
      "eval_runtime": 175.3493,
      "eval_samples_per_second": 6.644,
      "eval_steps_per_second": 6.644,
      "step": 1100
    },
    {
      "epoch": 1.694494800114493,
      "grad_norm": 0.2996631860733032,
      "learning_rate": 8.360940727244859e-05,
      "loss": 0.0418,
      "step": 1110
    },
    {
      "epoch": 1.7097605190344431,
      "grad_norm": 0.2608354687690735,
      "learning_rate": 8.321287091572403e-05,
      "loss": 0.0285,
      "step": 1120
    },
    {
      "epoch": 1.7250262379543937,
      "grad_norm": 0.45164602994918823,
      "learning_rate": 8.281256021240566e-05,
      "loss": 0.0299,
      "step": 1130
    },
    {
      "epoch": 1.740291956874344,
      "grad_norm": 0.21513421833515167,
      "learning_rate": 8.240852065423506e-05,
      "loss": 0.0384,
      "step": 1140
    },
    {
      "epoch": 1.7555576757942943,
      "grad_norm": 0.2836594879627228,
      "learning_rate": 8.20007981567048e-05,
      "loss": 0.0317,
      "step": 1150
    },
    {
      "epoch": 1.7555576757942943,
      "eval_loss": 0.03912714123725891,
      "eval_runtime": 176.2466,
      "eval_samples_per_second": 6.61,
      "eval_steps_per_second": 6.61,
      "step": 1150
    },
    {
      "epoch": 1.7708233947142449,
      "grad_norm": 0.25921815633773804,
      "learning_rate": 8.158943905384082e-05,
      "loss": 0.0384,
      "step": 1160
    },
    {
      "epoch": 1.7860891136341952,
      "grad_norm": 0.18497350811958313,
      "learning_rate": 8.117449009293668e-05,
      "loss": 0.0305,
      "step": 1170
    },
    {
      "epoch": 1.8013548325541455,
      "grad_norm": 0.16409793496131897,
      "learning_rate": 8.075599842924139e-05,
      "loss": 0.0306,
      "step": 1180
    },
    {
      "epoch": 1.816620551474096,
      "grad_norm": 0.16807886958122253,
      "learning_rate": 8.033401162060049e-05,
      "loss": 0.0361,
      "step": 1190
    },
    {
      "epoch": 1.8318862703940464,
      "grad_norm": 0.4135243594646454,
      "learning_rate": 7.990857762205157e-05,
      "loss": 0.0448,
      "step": 1200
    },
    {
      "epoch": 1.8318862703940464,
      "eval_loss": 0.03655826300382614,
      "eval_runtime": 176.7043,
      "eval_samples_per_second": 6.593,
      "eval_steps_per_second": 6.593,
      "step": 1200
    },
    {
      "epoch": 1.8471519893139967,
      "grad_norm": 0.28721022605895996,
      "learning_rate": 7.947974478037468e-05,
      "loss": 0.0378,
      "step": 1210
    },
    {
      "epoch": 1.8624177082339473,
      "grad_norm": 0.5347437262535095,
      "learning_rate": 7.904756182859797e-05,
      "loss": 0.0309,
      "step": 1220
    },
    {
      "epoch": 1.8776834271538976,
      "grad_norm": 0.10005893558263779,
      "learning_rate": 7.861207788045984e-05,
      "loss": 0.0318,
      "step": 1230
    },
    {
      "epoch": 1.892949146073848,
      "grad_norm": 0.355914443731308,
      "learning_rate": 7.817334242482738e-05,
      "loss": 0.0322,
      "step": 1240
    },
    {
      "epoch": 1.9082148649937984,
      "grad_norm": 0.1554553210735321,
      "learning_rate": 7.773140532007262e-05,
      "loss": 0.0278,
      "step": 1250
    },
    {
      "epoch": 1.9082148649937984,
      "eval_loss": 0.036195144057273865,
      "eval_runtime": 177.2966,
      "eval_samples_per_second": 6.571,
      "eval_steps_per_second": 6.571,
      "step": 1250
    },
    {
      "epoch": 1.9234805839137485,
      "grad_norm": 0.13769276440143585,
      "learning_rate": 7.728631678840638e-05,
      "loss": 0.0294,
      "step": 1260
    },
    {
      "epoch": 1.938746302833699,
      "grad_norm": 0.21283772587776184,
      "learning_rate": 7.683812741017112e-05,
      "loss": 0.0319,
      "step": 1270
    },
    {
      "epoch": 1.9540120217536494,
      "grad_norm": 0.5776203274726868,
      "learning_rate": 7.638688811809274e-05,
      "loss": 0.0388,
      "step": 1280
    },
    {
      "epoch": 1.9692777406735997,
      "grad_norm": 0.21127784252166748,
      "learning_rate": 7.593265019149275e-05,
      "loss": 0.0267,
      "step": 1290
    },
    {
      "epoch": 1.9845434595935503,
      "grad_norm": 0.4357820153236389,
      "learning_rate": 7.547546525046073e-05,
      "loss": 0.0347,
      "step": 1300
    },
    {
      "epoch": 1.9845434595935503,
      "eval_loss": 0.03436192870140076,
      "eval_runtime": 175.7034,
      "eval_samples_per_second": 6.63,
      "eval_steps_per_second": 6.63,
      "step": 1300
    },
    {
      "epoch": 1.9998091785135006,
      "grad_norm": 0.10159356147050858,
      "learning_rate": 7.501538524998812e-05,
      "loss": 0.0353,
      "step": 1310
    },
    {
      "epoch": 2.015074897433451,
      "grad_norm": 0.18945099413394928,
      "learning_rate": 7.455246247406406e-05,
      "loss": 0.0161,
      "step": 1320
    },
    {
      "epoch": 2.0303406163534015,
      "grad_norm": 0.3977908790111542,
      "learning_rate": 7.408674952973382e-05,
      "loss": 0.0256,
      "step": 1330
    },
    {
      "epoch": 2.0456063352733516,
      "grad_norm": 0.2554134428501129,
      "learning_rate": 7.361829934112036e-05,
      "loss": 0.0235,
      "step": 1340
    },
    {
      "epoch": 2.060872054193302,
      "grad_norm": 0.2563072144985199,
      "learning_rate": 7.314716514341006e-05,
      "loss": 0.0201,
      "step": 1350
    },
    {
      "epoch": 2.060872054193302,
      "eval_loss": 0.035528797656297684,
      "eval_runtime": 177.4606,
      "eval_samples_per_second": 6.565,
      "eval_steps_per_second": 6.565,
      "step": 1350
    },
    {
      "epoch": 2.0761377731132526,
      "grad_norm": 0.1881485879421234,
      "learning_rate": 7.267340047680305e-05,
      "loss": 0.0164,
      "step": 1360
    },
    {
      "epoch": 2.0914034920332027,
      "grad_norm": 0.29883432388305664,
      "learning_rate": 7.21970591804287e-05,
      "loss": 0.0182,
      "step": 1370
    },
    {
      "epoch": 2.1066692109531533,
      "grad_norm": 0.36703965067863464,
      "learning_rate": 7.171819538622747e-05,
      "loss": 0.0254,
      "step": 1380
    },
    {
      "epoch": 2.121934929873104,
      "grad_norm": 0.1899896115064621,
      "learning_rate": 7.123686351279914e-05,
      "loss": 0.0231,
      "step": 1390
    },
    {
      "epoch": 2.137200648793054,
      "grad_norm": 0.1549081802368164,
      "learning_rate": 7.07531182592187e-05,
      "loss": 0.0238,
      "step": 1400
    },
    {
      "epoch": 2.137200648793054,
      "eval_loss": 0.03574272617697716,
      "eval_runtime": 177.5378,
      "eval_samples_per_second": 6.562,
      "eval_steps_per_second": 6.562,
      "step": 1400
    },
    {
      "epoch": 2.1524663677130045,
      "grad_norm": 0.4319554567337036,
      "learning_rate": 7.026701459882026e-05,
      "loss": 0.0315,
      "step": 1410
    },
    {
      "epoch": 2.167732086632955,
      "grad_norm": 0.19883885979652405,
      "learning_rate": 6.977860777294988e-05,
      "loss": 0.0227,
      "step": 1420
    },
    {
      "epoch": 2.182997805552905,
      "grad_norm": 0.18116509914398193,
      "learning_rate": 6.92879532846878e-05,
      "loss": 0.0175,
      "step": 1430
    },
    {
      "epoch": 2.1982635244728557,
      "grad_norm": 0.43378233909606934,
      "learning_rate": 6.879510689254104e-05,
      "loss": 0.0187,
      "step": 1440
    },
    {
      "epoch": 2.213529243392806,
      "grad_norm": 0.3628329038619995,
      "learning_rate": 6.830012460410697e-05,
      "loss": 0.0299,
      "step": 1450
    },
    {
      "epoch": 2.213529243392806,
      "eval_loss": 0.037096813321113586,
      "eval_runtime": 176.5754,
      "eval_samples_per_second": 6.598,
      "eval_steps_per_second": 6.598,
      "step": 1450
    },
    {
      "epoch": 2.2287949623127563,
      "grad_norm": 0.28517764806747437,
      "learning_rate": 6.780306266970851e-05,
      "loss": 0.017,
      "step": 1460
    },
    {
      "epoch": 2.244060681232707,
      "grad_norm": 0.3066461682319641,
      "learning_rate": 6.73039775760018e-05,
      "loss": 0.0209,
      "step": 1470
    },
    {
      "epoch": 2.2593264001526574,
      "grad_norm": 0.22259260714054108,
      "learning_rate": 6.680292603955702e-05,
      "loss": 0.0201,
      "step": 1480
    },
    {
      "epoch": 2.2745921190726075,
      "grad_norm": 0.6137614846229553,
      "learning_rate": 6.629996500041299e-05,
      "loss": 0.022,
      "step": 1490
    },
    {
      "epoch": 2.289857837992558,
      "grad_norm": 0.11755680292844772,
      "learning_rate": 6.579515161560649e-05,
      "loss": 0.0155,
      "step": 1500
    },
    {
      "epoch": 2.289857837992558,
      "eval_loss": 0.038406338542699814,
      "eval_runtime": 175.5413,
      "eval_samples_per_second": 6.637,
      "eval_steps_per_second": 6.637,
      "step": 1500
    },
    {
      "epoch": 2.305123556912508,
      "grad_norm": 0.2847287058830261,
      "learning_rate": 6.528854325267692e-05,
      "loss": 0.0268,
      "step": 1510
    },
    {
      "epoch": 2.3203892758324587,
      "grad_norm": 0.412030965089798,
      "learning_rate": 6.478019748314686e-05,
      "loss": 0.0213,
      "step": 1520
    },
    {
      "epoch": 2.3356549947524092,
      "grad_norm": 0.31995999813079834,
      "learning_rate": 6.42701720759797e-05,
      "loss": 0.0238,
      "step": 1530
    },
    {
      "epoch": 2.3509207136723593,
      "grad_norm": 0.2422753870487213,
      "learning_rate": 6.375852499101467e-05,
      "loss": 0.019,
      "step": 1540
    },
    {
      "epoch": 2.36618643259231,
      "grad_norm": 0.10162404924631119,
      "learning_rate": 6.324531437238019e-05,
      "loss": 0.0157,
      "step": 1550
    },
    {
      "epoch": 2.36618643259231,
      "eval_loss": 0.03914285823702812,
      "eval_runtime": 174.6144,
      "eval_samples_per_second": 6.672,
      "eval_steps_per_second": 6.672,
      "step": 1550
    },
    {
      "epoch": 2.3814521515122604,
      "grad_norm": 0.27880334854125977,
      "learning_rate": 6.273059854188636e-05,
      "loss": 0.025,
      "step": 1560
    },
    {
      "epoch": 2.3967178704322105,
      "grad_norm": 0.4933589994907379,
      "learning_rate": 6.221443599239721e-05,
      "loss": 0.0242,
      "step": 1570
    },
    {
      "epoch": 2.411983589352161,
      "grad_norm": 0.5834177136421204,
      "learning_rate": 6.169688538118342e-05,
      "loss": 0.028,
      "step": 1580
    },
    {
      "epoch": 2.4272493082721116,
      "grad_norm": 0.4436355233192444,
      "learning_rate": 6.117800552325655e-05,
      "loss": 0.0232,
      "step": 1590
    },
    {
      "epoch": 2.4425150271920617,
      "grad_norm": 0.19233737885951996,
      "learning_rate": 6.06578553846852e-05,
      "loss": 0.0222,
      "step": 1600
    },
    {
      "epoch": 2.4425150271920617,
      "eval_loss": 0.03701253607869148,
      "eval_runtime": 175.2711,
      "eval_samples_per_second": 6.647,
      "eval_steps_per_second": 6.647,
      "step": 1600
    },
    {
      "epoch": 2.4577807461120122,
      "grad_norm": 0.22640623152256012,
      "learning_rate": 6.013649407589401e-05,
      "loss": 0.0228,
      "step": 1610
    },
    {
      "epoch": 2.473046465031963,
      "grad_norm": 0.19025041162967682,
      "learning_rate": 5.961398084494634e-05,
      "loss": 0.0246,
      "step": 1620
    },
    {
      "epoch": 2.488312183951913,
      "grad_norm": 0.5301237106323242,
      "learning_rate": 5.909037507081121e-05,
      "loss": 0.0283,
      "step": 1630
    },
    {
      "epoch": 2.5035779028718634,
      "grad_norm": 0.13809001445770264,
      "learning_rate": 5.8565736256615434e-05,
      "loss": 0.029,
      "step": 1640
    },
    {
      "epoch": 2.518843621791814,
      "grad_norm": 0.17507204413414001,
      "learning_rate": 5.8040124022881625e-05,
      "loss": 0.0245,
      "step": 1650
    },
    {
      "epoch": 2.518843621791814,
      "eval_loss": 0.035995036363601685,
      "eval_runtime": 170.1729,
      "eval_samples_per_second": 6.846,
      "eval_steps_per_second": 6.846,
      "step": 1650
    },
    {
      "epoch": 2.534109340711764,
      "grad_norm": 0.5437776446342468,
      "learning_rate": 5.751359810075284e-05,
      "loss": 0.0173,
      "step": 1660
    },
    {
      "epoch": 2.5493750596317146,
      "grad_norm": 0.21639682352542877,
      "learning_rate": 5.6986218325204676e-05,
      "loss": 0.0125,
      "step": 1670
    },
    {
      "epoch": 2.5646407785516647,
      "grad_norm": 0.36389076709747314,
      "learning_rate": 5.645804462824556e-05,
      "loss": 0.0158,
      "step": 1680
    },
    {
      "epoch": 2.5799064974716153,
      "grad_norm": 0.1342177391052246,
      "learning_rate": 5.5929137032106005e-05,
      "loss": 0.0247,
      "step": 1690
    },
    {
      "epoch": 2.595172216391566,
      "grad_norm": 0.2882712483406067,
      "learning_rate": 5.53995556424176e-05,
      "loss": 0.0206,
      "step": 1700
    },
    {
      "epoch": 2.595172216391566,
      "eval_loss": 0.03758452460169792,
      "eval_runtime": 158.0975,
      "eval_samples_per_second": 7.369,
      "eval_steps_per_second": 7.369,
      "step": 1700
    },
    {
      "epoch": 2.610437935311516,
      "grad_norm": 0.31510451436042786,
      "learning_rate": 5.4869360641382615e-05,
      "loss": 0.0175,
      "step": 1710
    },
    {
      "epoch": 2.6257036542314665,
      "grad_norm": 0.22791199386119843,
      "learning_rate": 5.433861228093471e-05,
      "loss": 0.0225,
      "step": 1720
    },
    {
      "epoch": 2.640969373151417,
      "grad_norm": 0.16334158182144165,
      "learning_rate": 5.380737087589197e-05,
      "loss": 0.0179,
      "step": 1730
    },
    {
      "epoch": 2.656235092071367,
      "grad_norm": 0.4029809534549713,
      "learning_rate": 5.327569679710256e-05,
      "loss": 0.032,
      "step": 1740
    },
    {
      "epoch": 2.6715008109913176,
      "grad_norm": 0.17828147113323212,
      "learning_rate": 5.274365046458416e-05,
      "loss": 0.0198,
      "step": 1750
    },
    {
      "epoch": 2.6715008109913176,
      "eval_loss": 0.036341313272714615,
      "eval_runtime": 157.9022,
      "eval_samples_per_second": 7.378,
      "eval_steps_per_second": 7.378,
      "step": 1750
    },
    {
      "epoch": 2.686766529911268,
      "grad_norm": 0.3760712742805481,
      "learning_rate": 5.2211292340657804e-05,
      "loss": 0.0256,
      "step": 1760
    },
    {
      "epoch": 2.7020322488312183,
      "grad_norm": 0.18362978100776672,
      "learning_rate": 5.167868292307678e-05,
      "loss": 0.0274,
      "step": 1770
    },
    {
      "epoch": 2.717297967751169,
      "grad_norm": 0.18960805237293243,
      "learning_rate": 5.114588273815173e-05,
      "loss": 0.0184,
      "step": 1780
    },
    {
      "epoch": 2.732563686671119,
      "grad_norm": 0.3714582622051239,
      "learning_rate": 5.061295233387223e-05,
      "loss": 0.0272,
      "step": 1790
    },
    {
      "epoch": 2.7478294055910695,
      "grad_norm": 0.3221634328365326,
      "learning_rate": 5.007995227302617e-05,
      "loss": 0.0209,
      "step": 1800
    },
    {
      "epoch": 2.7478294055910695,
      "eval_loss": 0.037039320915937424,
      "eval_runtime": 158.1041,
      "eval_samples_per_second": 7.369,
      "eval_steps_per_second": 7.369,
      "step": 1800
    },
    {
      "epoch": 2.76309512451102,
      "grad_norm": 0.13622955977916718,
      "learning_rate": 4.954694312631729e-05,
      "loss": 0.022,
      "step": 1810
    },
    {
      "epoch": 2.7783608434309706,
      "grad_norm": 0.2798239290714264,
      "learning_rate": 4.901398546548181e-05,
      "loss": 0.0205,
      "step": 1820
    },
    {
      "epoch": 2.7936265623509207,
      "grad_norm": 0.34951063990592957,
      "learning_rate": 4.848113985640513e-05,
      "loss": 0.0213,
      "step": 1830
    },
    {
      "epoch": 2.808892281270871,
      "grad_norm": 0.19509588181972504,
      "learning_rate": 4.794846685223884e-05,
      "loss": 0.0231,
      "step": 1840
    },
    {
      "epoch": 2.8241580001908213,
      "grad_norm": 0.4014818072319031,
      "learning_rate": 4.741602698651966e-05,
      "loss": 0.026,
      "step": 1850
    },
    {
      "epoch": 2.8241580001908213,
      "eval_loss": 0.036198943853378296,
      "eval_runtime": 157.9325,
      "eval_samples_per_second": 7.377,
      "eval_steps_per_second": 7.377,
      "step": 1850
    },
    {
      "epoch": 2.839423719110772,
      "grad_norm": 0.5051618814468384,
      "learning_rate": 4.6883880766290086e-05,
      "loss": 0.0276,
      "step": 1860
    },
    {
      "epoch": 2.8546894380307224,
      "grad_norm": 0.2065005749464035,
      "learning_rate": 4.635208866522251e-05,
      "loss": 0.0188,
      "step": 1870
    },
    {
      "epoch": 2.8699551569506725,
      "grad_norm": 0.27732136845588684,
      "learning_rate": 4.5820711116746785e-05,
      "loss": 0.0193,
      "step": 1880
    },
    {
      "epoch": 2.885220875870623,
      "grad_norm": 0.35616669058799744,
      "learning_rate": 4.528980850718255e-05,
      "loss": 0.0237,
      "step": 1890
    },
    {
      "epoch": 2.9004865947905736,
      "grad_norm": 0.30379197001457214,
      "learning_rate": 4.475944116887695e-05,
      "loss": 0.0197,
      "step": 1900
    },
    {
      "epoch": 2.9004865947905736,
      "eval_loss": 0.03581969812512398,
      "eval_runtime": 158.3121,
      "eval_samples_per_second": 7.359,
      "eval_steps_per_second": 7.359,
      "step": 1900
    },
    {
      "epoch": 2.9157523137105237,
      "grad_norm": 0.3760513365268707,
      "learning_rate": 4.4229669373348226e-05,
      "loss": 0.0278,
      "step": 1910
    },
    {
      "epoch": 2.9310180326304742,
      "grad_norm": 0.14965511858463287,
      "learning_rate": 4.3700553324436575e-05,
      "loss": 0.0216,
      "step": 1920
    },
    {
      "epoch": 2.9462837515504248,
      "grad_norm": 0.3847552537918091,
      "learning_rate": 4.317215315146238e-05,
      "loss": 0.0277,
      "step": 1930
    },
    {
      "epoch": 2.961549470470375,
      "grad_norm": 0.21934816241264343,
      "learning_rate": 4.264452890239315e-05,
      "loss": 0.0242,
      "step": 1940
    },
    {
      "epoch": 2.9768151893903254,
      "grad_norm": 0.347125768661499,
      "learning_rate": 4.211774053701952e-05,
      "loss": 0.0291,
      "step": 1950
    },
    {
      "epoch": 2.9768151893903254,
      "eval_loss": 0.03554380312561989,
      "eval_runtime": 158.3407,
      "eval_samples_per_second": 7.358,
      "eval_steps_per_second": 7.358,
      "step": 1950
    },
    {
      "epoch": 2.9920809083102755,
      "grad_norm": 0.09311053901910782,
      "learning_rate": 4.159184792014145e-05,
      "loss": 0.0155,
      "step": 1960
    },
    {
      "epoch": 3.007346627230226,
      "grad_norm": 0.094869464635849,
      "learning_rate": 4.1066910814765016e-05,
      "loss": 0.0118,
      "step": 1970
    },
    {
      "epoch": 3.0226123461501766,
      "grad_norm": 0.2167072743177414,
      "learning_rate": 4.0542988875310995e-05,
      "loss": 0.0148,
      "step": 1980
    },
    {
      "epoch": 3.0378780650701267,
      "grad_norm": 0.14555399119853973,
      "learning_rate": 4.002014164083552e-05,
      "loss": 0.0136,
      "step": 1990
    },
    {
      "epoch": 3.0531437839900772,
      "grad_norm": 0.1664235144853592,
      "learning_rate": 3.9498428528264204e-05,
      "loss": 0.0091,
      "step": 2000
    },
    {
      "epoch": 3.0531437839900772,
      "eval_loss": 0.04159301891922951,
      "eval_runtime": 158.3202,
      "eval_samples_per_second": 7.359,
      "eval_steps_per_second": 7.359,
      "step": 2000
    },
    {
      "epoch": 3.068409502910028,
      "grad_norm": 0.2766287326812744,
      "learning_rate": 3.89779088256397e-05,
      "loss": 0.0092,
      "step": 2010
    },
    {
      "epoch": 3.083675221829978,
      "grad_norm": 0.38887709379196167,
      "learning_rate": 3.845864168538437e-05,
      "loss": 0.0097,
      "step": 2020
    },
    {
      "epoch": 3.0989409407499284,
      "grad_norm": 0.35072392225265503,
      "learning_rate": 3.794068611757794e-05,
      "loss": 0.0117,
      "step": 2030
    },
    {
      "epoch": 3.114206659669879,
      "grad_norm": 0.2347632646560669,
      "learning_rate": 3.7424100983251695e-05,
      "loss": 0.0157,
      "step": 2040
    },
    {
      "epoch": 3.129472378589829,
      "grad_norm": 0.6192054748535156,
      "learning_rate": 3.6908944987699345e-05,
      "loss": 0.0132,
      "step": 2050
    },
    {
      "epoch": 3.129472378589829,
      "eval_loss": 0.042059410363435745,
      "eval_runtime": 158.4528,
      "eval_samples_per_second": 7.352,
      "eval_steps_per_second": 7.352,
      "step": 2050
    },
    {
      "epoch": 3.1447380975097796,
      "grad_norm": 0.07461170107126236,
      "learning_rate": 3.639527667380571e-05,
      "loss": 0.0087,
      "step": 2060
    },
    {
      "epoch": 3.16000381642973,
      "grad_norm": 0.24892112612724304,
      "learning_rate": 3.5883154415393885e-05,
      "loss": 0.0086,
      "step": 2070
    },
    {
      "epoch": 3.1752695353496803,
      "grad_norm": 0.25514039397239685,
      "learning_rate": 3.537263641059152e-05,
      "loss": 0.0111,
      "step": 2080
    },
    {
      "epoch": 3.190535254269631,
      "grad_norm": 0.3669221103191376,
      "learning_rate": 3.486378067521718e-05,
      "loss": 0.01,
      "step": 2090
    },
    {
      "epoch": 3.2058009731895813,
      "grad_norm": 0.24360391497612,
      "learning_rate": 3.435664503618732e-05,
      "loss": 0.0115,
      "step": 2100
    },
    {
      "epoch": 3.2058009731895813,
      "eval_loss": 0.04429946094751358,
      "eval_runtime": 158.3749,
      "eval_samples_per_second": 7.356,
      "eval_steps_per_second": 7.356,
      "step": 2100
    },
    {
      "epoch": 3.2210666921095314,
      "grad_norm": 0.14438766241073608,
      "learning_rate": 3.3851287124944756e-05,
      "loss": 0.0108,
      "step": 2110
    },
    {
      "epoch": 3.236332411029482,
      "grad_norm": 0.21432369947433472,
      "learning_rate": 3.334776437090944e-05,
      "loss": 0.0086,
      "step": 2120
    },
    {
      "epoch": 3.251598129949432,
      "grad_norm": 0.030950259417295456,
      "learning_rate": 3.2846133994952046e-05,
      "loss": 0.0101,
      "step": 2130
    },
    {
      "epoch": 3.2668638488693826,
      "grad_norm": 0.3596804440021515,
      "learning_rate": 3.234645300289136e-05,
      "loss": 0.0061,
      "step": 2140
    },
    {
      "epoch": 3.282129567789333,
      "grad_norm": 0.9810014963150024,
      "learning_rate": 3.1848778179016075e-05,
      "loss": 0.0131,
      "step": 2150
    },
    {
      "epoch": 3.282129567789333,
      "eval_loss": 0.04585850238800049,
      "eval_runtime": 158.4392,
      "eval_samples_per_second": 7.353,
      "eval_steps_per_second": 7.353,
      "step": 2150
    },
    {
      "epoch": 3.2973952867092833,
      "grad_norm": 0.48682817816734314,
      "learning_rate": 3.135316607963176e-05,
      "loss": 0.0127,
      "step": 2160
    },
    {
      "epoch": 3.312661005629234,
      "grad_norm": 0.23697037994861603,
      "learning_rate": 3.085967302663375e-05,
      "loss": 0.0108,
      "step": 2170
    },
    {
      "epoch": 3.3279267245491844,
      "grad_norm": 0.5131348371505737,
      "learning_rate": 3.0368355101106615e-05,
      "loss": 0.0168,
      "step": 2180
    },
    {
      "epoch": 3.3431924434691345,
      "grad_norm": 0.4184378981590271,
      "learning_rate": 2.987926813695116e-05,
      "loss": 0.0095,
      "step": 2190
    },
    {
      "epoch": 3.358458162389085,
      "grad_norm": 0.14508870244026184,
      "learning_rate": 2.939246771453924e-05,
      "loss": 0.0132,
      "step": 2200
    },
    {
      "epoch": 3.358458162389085,
      "eval_loss": 0.04094408452510834,
      "eval_runtime": 158.4285,
      "eval_samples_per_second": 7.353,
      "eval_steps_per_second": 7.353,
      "step": 2200
    },
    {
      "epoch": 3.3737238813090356,
      "grad_norm": 0.41974306106567383,
      "learning_rate": 2.890800915439772e-05,
      "loss": 0.0136,
      "step": 2210
    },
    {
      "epoch": 3.3889896002289857,
      "grad_norm": 0.24298487603664398,
      "learning_rate": 2.842594751092159e-05,
      "loss": 0.0094,
      "step": 2220
    },
    {
      "epoch": 3.404255319148936,
      "grad_norm": 0.2780051827430725,
      "learning_rate": 2.794633756611776e-05,
      "loss": 0.0111,
      "step": 2230
    },
    {
      "epoch": 3.4195210380688867,
      "grad_norm": 0.4046504497528076,
      "learning_rate": 2.7469233823379347e-05,
      "loss": 0.0106,
      "step": 2240
    },
    {
      "epoch": 3.434786756988837,
      "grad_norm": 0.5508965849876404,
      "learning_rate": 2.6994690501292032e-05,
      "loss": 0.0077,
      "step": 2250
    },
    {
      "epoch": 3.434786756988837,
      "eval_loss": 0.04445834830403328,
      "eval_runtime": 158.3347,
      "eval_samples_per_second": 7.358,
      "eval_steps_per_second": 7.358,
      "step": 2250
    },
    {
      "epoch": 3.4500524759087874,
      "grad_norm": 0.1673806607723236,
      "learning_rate": 2.652276152747246e-05,
      "loss": 0.0073,
      "step": 2260
    },
    {
      "epoch": 3.465318194828738,
      "grad_norm": 0.11956673860549927,
      "learning_rate": 2.6053500532439968e-05,
      "loss": 0.0058,
      "step": 2270
    },
    {
      "epoch": 3.480583913748688,
      "grad_norm": 0.08597150444984436,
      "learning_rate": 2.5586960843521824e-05,
      "loss": 0.0135,
      "step": 2280
    },
    {
      "epoch": 3.4958496326686386,
      "grad_norm": 0.456600159406662,
      "learning_rate": 2.5123195478793217e-05,
      "loss": 0.0137,
      "step": 2290
    },
    {
      "epoch": 3.5111153515885887,
      "grad_norm": 0.10607762634754181,
      "learning_rate": 2.466225714105202e-05,
      "loss": 0.0156,
      "step": 2300
    },
    {
      "epoch": 3.5111153515885887,
      "eval_loss": 0.04438282549381256,
      "eval_runtime": 158.4663,
      "eval_samples_per_second": 7.352,
      "eval_steps_per_second": 7.352,
      "step": 2300
    },
    {
      "epoch": 3.526381070508539,
      "grad_norm": 0.2712307572364807,
      "learning_rate": 2.420419821182982e-05,
      "loss": 0.014,
      "step": 2310
    },
    {
      "epoch": 3.5416467894284898,
      "grad_norm": 0.34561389684677124,
      "learning_rate": 2.3749070745438996e-05,
      "loss": 0.0097,
      "step": 2320
    },
    {
      "epoch": 3.5569125083484403,
      "grad_norm": 0.1999119520187378,
      "learning_rate": 2.3296926463057396e-05,
      "loss": 0.0092,
      "step": 2330
    },
    {
      "epoch": 3.5721782272683904,
      "grad_norm": 0.24684706330299377,
      "learning_rate": 2.284781674685058e-05,
      "loss": 0.0076,
      "step": 2340
    },
    {
      "epoch": 3.587443946188341,
      "grad_norm": 0.13131554424762726,
      "learning_rate": 2.2401792634132708e-05,
      "loss": 0.0125,
      "step": 2350
    },
    {
      "epoch": 3.587443946188341,
      "eval_loss": 0.048045314848423004,
      "eval_runtime": 158.5084,
      "eval_samples_per_second": 7.35,
      "eval_steps_per_second": 7.35,
      "step": 2350
    },
    {
      "epoch": 3.602709665108291,
      "grad_norm": 0.5896148085594177,
      "learning_rate": 2.195890481156666e-05,
      "loss": 0.0094,
      "step": 2360
    },
    {
      "epoch": 3.6179753840282416,
      "grad_norm": 0.13258546590805054,
      "learning_rate": 2.151920360940387e-05,
      "loss": 0.0122,
      "step": 2370
    },
    {
      "epoch": 3.633241102948192,
      "grad_norm": 0.40482622385025024,
      "learning_rate": 2.1082738995764785e-05,
      "loss": 0.0075,
      "step": 2380
    },
    {
      "epoch": 3.6485068218681422,
      "grad_norm": 0.34933677315711975,
      "learning_rate": 2.0649560570960465e-05,
      "loss": 0.0063,
      "step": 2390
    },
    {
      "epoch": 3.663772540788093,
      "grad_norm": 0.37669917941093445,
      "learning_rate": 2.0219717561855855e-05,
      "loss": 0.0089,
      "step": 2400
    },
    {
      "epoch": 3.663772540788093,
      "eval_loss": 0.049949660897254944,
      "eval_runtime": 158.5713,
      "eval_samples_per_second": 7.347,
      "eval_steps_per_second": 7.347,
      "step": 2400
    },
    {
      "epoch": 3.679038259708043,
      "grad_norm": 0.26693466305732727,
      "learning_rate": 1.9793258816275728e-05,
      "loss": 0.0168,
      "step": 2410
    },
    {
      "epoch": 3.6943039786279934,
      "grad_norm": 0.3945235311985016,
      "learning_rate": 1.9370232797453402e-05,
      "loss": 0.0106,
      "step": 2420
    },
    {
      "epoch": 3.709569697547944,
      "grad_norm": 0.529176652431488,
      "learning_rate": 1.8950687578523502e-05,
      "loss": 0.01,
      "step": 2430
    },
    {
      "epoch": 3.7248354164678945,
      "grad_norm": 0.22456000745296478,
      "learning_rate": 1.853467083705869e-05,
      "loss": 0.0067,
      "step": 2440
    },
    {
      "epoch": 3.7401011353878446,
      "grad_norm": 0.406535267829895,
      "learning_rate": 1.8122229849651716e-05,
      "loss": 0.0125,
      "step": 2450
    },
    {
      "epoch": 3.7401011353878446,
      "eval_loss": 0.04671522229909897,
      "eval_runtime": 158.5667,
      "eval_samples_per_second": 7.347,
      "eval_steps_per_second": 7.347,
      "step": 2450
    },
    {
      "epoch": 3.755366854307795,
      "grad_norm": 0.24689412117004395,
      "learning_rate": 1.7713411486542707e-05,
      "loss": 0.014,
      "step": 2460
    },
    {
      "epoch": 3.7706325732277453,
      "grad_norm": 0.27206507325172424,
      "learning_rate": 1.7308262206292897e-05,
      "loss": 0.0076,
      "step": 2470
    },
    {
      "epoch": 3.785898292147696,
      "grad_norm": 0.6516662240028381,
      "learning_rate": 1.6906828050504907e-05,
      "loss": 0.015,
      "step": 2480
    },
    {
      "epoch": 3.8011640110676463,
      "grad_norm": 0.7136443853378296,
      "learning_rate": 1.650915463859068e-05,
      "loss": 0.0096,
      "step": 2490
    },
    {
      "epoch": 3.8164297299875964,
      "grad_norm": 0.08043626695871353,
      "learning_rate": 1.6115287162587055e-05,
      "loss": 0.0115,
      "step": 2500
    },
    {
      "epoch": 3.8164297299875964,
      "eval_loss": 0.04467522352933884,
      "eval_runtime": 158.5106,
      "eval_samples_per_second": 7.35,
      "eval_steps_per_second": 7.35,
      "step": 2500
    },
    {
      "epoch": 3.831695448907547,
      "grad_norm": 0.37196823954582214,
      "learning_rate": 1.57252703820203e-05,
      "loss": 0.0064,
      "step": 2510
    },
    {
      "epoch": 3.8469611678274975,
      "grad_norm": 0.5090697407722473,
      "learning_rate": 1.5339148618819393e-05,
      "loss": 0.0159,
      "step": 2520
    },
    {
      "epoch": 3.8622268867474476,
      "grad_norm": 0.317979633808136,
      "learning_rate": 1.4956965752279395e-05,
      "loss": 0.0135,
      "step": 2530
    },
    {
      "epoch": 3.877492605667398,
      "grad_norm": 0.30272015929222107,
      "learning_rate": 1.457876521407484e-05,
      "loss": 0.0103,
      "step": 2540
    },
    {
      "epoch": 3.8927583245873487,
      "grad_norm": 0.33453094959259033,
      "learning_rate": 1.4204589983324173e-05,
      "loss": 0.0062,
      "step": 2550
    },
    {
      "epoch": 3.8927583245873487,
      "eval_loss": 0.044938329607248306,
      "eval_runtime": 158.251,
      "eval_samples_per_second": 7.362,
      "eval_steps_per_second": 7.362,
      "step": 2550
    },
    {
      "epoch": 3.908024043507299,
      "grad_norm": 0.22449414432048798,
      "learning_rate": 1.383448258170557e-05,
      "loss": 0.0095,
      "step": 2560
    },
    {
      "epoch": 3.9232897624272494,
      "grad_norm": 0.35782113671302795,
      "learning_rate": 1.3468485068624653e-05,
      "loss": 0.0069,
      "step": 2570
    },
    {
      "epoch": 3.9385554813471995,
      "grad_norm": 0.019237499684095383,
      "learning_rate": 1.310663903643492e-05,
      "loss": 0.0119,
      "step": 2580
    },
    {
      "epoch": 3.95382120026715,
      "grad_norm": 0.19343972206115723,
      "learning_rate": 1.2748985605711028e-05,
      "loss": 0.0109,
      "step": 2590
    },
    {
      "epoch": 3.9690869191871005,
      "grad_norm": 0.3041229844093323,
      "learning_rate": 1.2395565420575932e-05,
      "loss": 0.0112,
      "step": 2600
    },
    {
      "epoch": 3.9690869191871005,
      "eval_loss": 0.046159546822309494,
      "eval_runtime": 158.163,
      "eval_samples_per_second": 7.366,
      "eval_steps_per_second": 7.366,
      "step": 2600
    },
    {
      "epoch": 3.984352638107051,
      "grad_norm": 0.18217749893665314,
      "learning_rate": 1.2046418644081903e-05,
      "loss": 0.0194,
      "step": 2610
    },
    {
      "epoch": 3.999618357027001,
      "grad_norm": 0.40362730622291565,
      "learning_rate": 1.1701584953646505e-05,
      "loss": 0.0084,
      "step": 2620
    },
    {
      "epoch": 4.014884075946951,
      "grad_norm": 0.20328281819820404,
      "learning_rate": 1.1361103536543466e-05,
      "loss": 0.006,
      "step": 2630
    },
    {
      "epoch": 4.030149794866902,
      "grad_norm": 0.03458496183156967,
      "learning_rate": 1.1025013085449527e-05,
      "loss": 0.0033,
      "step": 2640
    },
    {
      "epoch": 4.045415513786852,
      "grad_norm": 0.24609656631946564,
      "learning_rate": 1.0693351794047224e-05,
      "loss": 0.005,
      "step": 2650
    },
    {
      "epoch": 4.045415513786852,
      "eval_loss": 0.04647543653845787,
      "eval_runtime": 158.0997,
      "eval_samples_per_second": 7.369,
      "eval_steps_per_second": 7.369,
      "step": 2650
    },
    {
      "epoch": 4.060681232706803,
      "grad_norm": 0.12650443613529205,
      "learning_rate": 1.036615735268468e-05,
      "loss": 0.0043,
      "step": 2660
    },
    {
      "epoch": 4.0759469516267535,
      "grad_norm": 0.06227023899555206,
      "learning_rate": 1.0043466944092272e-05,
      "loss": 0.0058,
      "step": 2670
    },
    {
      "epoch": 4.091212670546703,
      "grad_norm": 0.09046266973018646,
      "learning_rate": 9.72531723915726e-06,
      "loss": 0.0081,
      "step": 2680
    },
    {
      "epoch": 4.106478389466654,
      "grad_norm": 0.09273139387369156,
      "learning_rate": 9.411744392756405e-06,
      "loss": 0.0055,
      "step": 2690
    },
    {
      "epoch": 4.121744108386604,
      "grad_norm": 0.07838311046361923,
      "learning_rate": 9.102784039647339e-06,
      "loss": 0.0065,
      "step": 2700
    },
    {
      "epoch": 4.121744108386604,
      "eval_loss": 0.050155650824308395,
      "eval_runtime": 158.2759,
      "eval_samples_per_second": 7.361,
      "eval_steps_per_second": 7.361,
      "step": 2700
    },
    {
      "epoch": 4.137009827306555,
      "grad_norm": 1.3356140851974487,
      "learning_rate": 8.79847129041893e-06,
      "loss": 0.007,
      "step": 2710
    },
    {
      "epoch": 4.152275546226505,
      "grad_norm": 0.06380099803209305,
      "learning_rate": 8.498840727501316e-06,
      "loss": 0.0032,
      "step": 2720
    },
    {
      "epoch": 4.167541265146456,
      "grad_norm": 0.05524059012532234,
      "learning_rate": 8.203926401235957e-06,
      "loss": 0.0028,
      "step": 2730
    },
    {
      "epoch": 4.1828069840664055,
      "grad_norm": 0.048491813242435455,
      "learning_rate": 7.913761826006017e-06,
      "loss": 0.0029,
      "step": 2740
    },
    {
      "epoch": 4.198072702986356,
      "grad_norm": 0.07295604050159454,
      "learning_rate": 7.628379976427868e-06,
      "loss": 0.0021,
      "step": 2750
    },
    {
      "epoch": 4.198072702986356,
      "eval_loss": 0.054289042949676514,
      "eval_runtime": 158.2341,
      "eval_samples_per_second": 7.363,
      "eval_steps_per_second": 7.363,
      "step": 2750
    },
    {
      "epoch": 4.213338421906307,
      "grad_norm": 0.10423220694065094,
      "learning_rate": 7.347813283603705e-06,
      "loss": 0.008,
      "step": 2760
    },
    {
      "epoch": 4.228604140826257,
      "grad_norm": 0.07456283271312714,
      "learning_rate": 7.072093631436161e-06,
      "loss": 0.0041,
      "step": 2770
    },
    {
      "epoch": 4.243869859746208,
      "grad_norm": 0.01388398464769125,
      "learning_rate": 6.801252353004867e-06,
      "loss": 0.0035,
      "step": 2780
    },
    {
      "epoch": 4.259135578666158,
      "grad_norm": 0.2674631178379059,
      "learning_rate": 6.535320227005826e-06,
      "loss": 0.0037,
      "step": 2790
    },
    {
      "epoch": 4.274401297586108,
      "grad_norm": 0.028004569932818413,
      "learning_rate": 6.274327474253611e-06,
      "loss": 0.0033,
      "step": 2800
    },
    {
      "epoch": 4.274401297586108,
      "eval_loss": 0.055618129670619965,
      "eval_runtime": 158.2941,
      "eval_samples_per_second": 7.36,
      "eval_steps_per_second": 7.36,
      "step": 2800
    },
    {
      "epoch": 4.289667016506058,
      "grad_norm": 0.02966715581715107,
      "learning_rate": 6.018303754247112e-06,
      "loss": 0.004,
      "step": 2810
    },
    {
      "epoch": 4.304932735426009,
      "grad_norm": 0.07398536056280136,
      "learning_rate": 5.767278161798911e-06,
      "loss": 0.0045,
      "step": 2820
    },
    {
      "epoch": 4.3201984543459595,
      "grad_norm": 0.06184925138950348,
      "learning_rate": 5.521279223729026e-06,
      "loss": 0.0018,
      "step": 2830
    },
    {
      "epoch": 4.33546417326591,
      "grad_norm": 0.10655587166547775,
      "learning_rate": 5.280334895622968e-06,
      "loss": 0.0036,
      "step": 2840
    },
    {
      "epoch": 4.350729892185861,
      "grad_norm": 0.2469986528158188,
      "learning_rate": 5.044472558654961e-06,
      "loss": 0.0068,
      "step": 2850
    },
    {
      "epoch": 4.350729892185861,
      "eval_loss": 0.057221900671720505,
      "eval_runtime": 158.3313,
      "eval_samples_per_second": 7.358,
      "eval_steps_per_second": 7.358,
      "step": 2850
    },
    {
      "epoch": 4.36599561110581,
      "grad_norm": 0.07252271473407745,
      "learning_rate": 4.813719016476203e-06,
      "loss": 0.0024,
      "step": 2860
    },
    {
      "epoch": 4.381261330025761,
      "grad_norm": 0.3494785726070404,
      "learning_rate": 4.588100492168973e-06,
      "loss": 0.0017,
      "step": 2870
    },
    {
      "epoch": 4.396527048945711,
      "grad_norm": 0.10995281487703323,
      "learning_rate": 4.367642625266511e-06,
      "loss": 0.0029,
      "step": 2880
    },
    {
      "epoch": 4.411792767865662,
      "grad_norm": 0.01004809606820345,
      "learning_rate": 4.1523704688394176e-06,
      "loss": 0.0023,
      "step": 2890
    },
    {
      "epoch": 4.427058486785612,
      "grad_norm": 0.11159130930900574,
      "learning_rate": 3.9423084866484884e-06,
      "loss": 0.0015,
      "step": 2900
    },
    {
      "epoch": 4.427058486785612,
      "eval_loss": 0.05985542759299278,
      "eval_runtime": 158.3929,
      "eval_samples_per_second": 7.355,
      "eval_steps_per_second": 7.355,
      "step": 2900
    },
    {
      "epoch": 4.442324205705562,
      "grad_norm": 0.35261520743370056,
      "learning_rate": 3.737480550364736e-06,
      "loss": 0.0071,
      "step": 2910
    },
    {
      "epoch": 4.457589924625513,
      "grad_norm": 0.26947641372680664,
      "learning_rate": 3.5379099368564817e-06,
      "loss": 0.0041,
      "step": 2920
    },
    {
      "epoch": 4.472855643545463,
      "grad_norm": 0.24151833355426788,
      "learning_rate": 3.3436193255442396e-06,
      "loss": 0.0026,
      "step": 2930
    },
    {
      "epoch": 4.488121362465414,
      "grad_norm": 0.007592054083943367,
      "learning_rate": 3.1546307958233214e-06,
      "loss": 0.0031,
      "step": 2940
    },
    {
      "epoch": 4.503387081385364,
      "grad_norm": 0.16517309844493866,
      "learning_rate": 2.9709658245547834e-06,
      "loss": 0.0036,
      "step": 2950
    },
    {
      "epoch": 4.503387081385364,
      "eval_loss": 0.060201384127140045,
      "eval_runtime": 158.4197,
      "eval_samples_per_second": 7.354,
      "eval_steps_per_second": 7.354,
      "step": 2950
    },
    {
      "epoch": 4.518652800305315,
      "grad_norm": 0.08729831874370575,
      "learning_rate": 2.792645283624712e-06,
      "loss": 0.0027,
      "step": 2960
    },
    {
      "epoch": 4.5339185192252645,
      "grad_norm": 0.11076062172651291,
      "learning_rate": 2.6196894375723645e-06,
      "loss": 0.0014,
      "step": 2970
    },
    {
      "epoch": 4.549184238145215,
      "grad_norm": 0.009186147712171078,
      "learning_rate": 2.452117941287246e-06,
      "loss": 0.0064,
      "step": 2980
    },
    {
      "epoch": 4.5644499570651655,
      "grad_norm": 0.14165641367435455,
      "learning_rate": 2.2899498377755566e-06,
      "loss": 0.0012,
      "step": 2990
    },
    {
      "epoch": 4.579715675985116,
      "grad_norm": 0.09586022794246674,
      "learning_rate": 2.1332035559960663e-06,
      "loss": 0.0027,
      "step": 3000
    },
    {
      "epoch": 4.579715675985116,
      "eval_loss": 0.06145218014717102,
      "eval_runtime": 158.4209,
      "eval_samples_per_second": 7.354,
      "eval_steps_per_second": 7.354,
      "step": 3000
    },
    {
      "epoch": 4.594981394905067,
      "grad_norm": 0.12076330184936523,
      "learning_rate": 1.9818969087658735e-06,
      "loss": 0.0043,
      "step": 3010
    },
    {
      "epoch": 4.610247113825016,
      "grad_norm": 0.002956979675218463,
      "learning_rate": 1.8360470907361093e-06,
      "loss": 0.0011,
      "step": 3020
    },
    {
      "epoch": 4.625512832744967,
      "grad_norm": 0.022843187674880028,
      "learning_rate": 1.6956706764379438e-06,
      "loss": 0.0056,
      "step": 3030
    },
    {
      "epoch": 4.640778551664917,
      "grad_norm": 0.12420812249183655,
      "learning_rate": 1.5607836183989921e-06,
      "loss": 0.0028,
      "step": 3040
    },
    {
      "epoch": 4.656044270584868,
      "grad_norm": 0.23208186030387878,
      "learning_rate": 1.4314012453305215e-06,
      "loss": 0.0013,
      "step": 3050
    },
    {
      "epoch": 4.656044270584868,
      "eval_loss": 0.061488546431064606,
      "eval_runtime": 158.5492,
      "eval_samples_per_second": 7.348,
      "eval_steps_per_second": 7.348,
      "step": 3050
    },
    {
      "epoch": 4.6713099895048185,
      "grad_norm": 0.36894094944000244,
      "learning_rate": 1.3075382603854157e-06,
      "loss": 0.0028,
      "step": 3060
    },
    {
      "epoch": 4.686575708424769,
      "grad_norm": 0.011954453773796558,
      "learning_rate": 1.1892087394873353e-06,
      "loss": 0.0027,
      "step": 3070
    },
    {
      "epoch": 4.701841427344719,
      "grad_norm": 0.025959955528378487,
      "learning_rate": 1.076426129731084e-06,
      "loss": 0.0038,
      "step": 3080
    },
    {
      "epoch": 4.717107146264669,
      "grad_norm": 0.13131509721279144,
      "learning_rate": 9.692032478545e-07,
      "loss": 0.005,
      "step": 3090
    },
    {
      "epoch": 4.73237286518462,
      "grad_norm": 0.2080388367176056,
      "learning_rate": 8.675522787819023e-07,
      "loss": 0.0056,
      "step": 3100
    },
    {
      "epoch": 4.73237286518462,
      "eval_loss": 0.061799220740795135,
      "eval_runtime": 158.5326,
      "eval_samples_per_second": 7.349,
      "eval_steps_per_second": 7.349,
      "step": 3100
    },
    {
      "epoch": 4.74763858410457,
      "grad_norm": 0.04518610239028931,
      "learning_rate": 7.714847742394337e-07,
      "loss": 0.0051,
      "step": 3110
    },
    {
      "epoch": 4.762904303024521,
      "grad_norm": 0.613032877445221,
      "learning_rate": 6.810116514422593e-07,
      "loss": 0.0089,
      "step": 3120
    },
    {
      "epoch": 4.7781700219444705,
      "grad_norm": 0.09960233420133591,
      "learning_rate": 5.961431918539817e-07,
      "loss": 0.0031,
      "step": 3130
    },
    {
      "epoch": 4.793435740864421,
      "grad_norm": 0.15351538360118866,
      "learning_rate": 5.16889040018187e-07,
      "loss": 0.0058,
      "step": 3140
    },
    {
      "epoch": 4.808701459784372,
      "grad_norm": 0.039991118013858795,
      "learning_rate": 4.432582024624543e-07,
      "loss": 0.0028,
      "step": 3150
    },
    {
      "epoch": 4.808701459784372,
      "eval_loss": 0.06179770454764366,
      "eval_runtime": 158.4692,
      "eval_samples_per_second": 7.352,
      "eval_steps_per_second": 7.352,
      "step": 3150
    },
    {
      "epoch": 4.823967178704322,
      "grad_norm": 0.03549914062023163,
      "learning_rate": 3.7525904667484737e-07,
      "loss": 0.0028,
      "step": 3160
    },
    {
      "epoch": 4.839232897624273,
      "grad_norm": 0.3593372702598572,
      "learning_rate": 3.128993001530245e-07,
      "loss": 0.0037,
      "step": 3170
    },
    {
      "epoch": 4.854498616544223,
      "grad_norm": 0.10235150158405304,
      "learning_rate": 2.5618604952605816e-07,
      "loss": 0.003,
      "step": 3180
    },
    {
      "epoch": 4.869764335464174,
      "grad_norm": 0.029046185314655304,
      "learning_rate": 2.0512573974912908e-07,
      "loss": 0.0034,
      "step": 3190
    },
    {
      "epoch": 4.885030054384123,
      "grad_norm": 0.3789989948272705,
      "learning_rate": 1.597241733711008e-07,
      "loss": 0.0044,
      "step": 3200
    },
    {
      "epoch": 4.885030054384123,
      "eval_loss": 0.06198294460773468,
      "eval_runtime": 158.4955,
      "eval_samples_per_second": 7.35,
      "eval_steps_per_second": 7.35,
      "step": 3200
    },
    {
      "epoch": 4.900295773304074,
      "grad_norm": 0.03306398168206215,
      "learning_rate": 1.1998650987510295e-07,
      "loss": 0.0077,
      "step": 3210
    },
    {
      "epoch": 4.9155614922240245,
      "grad_norm": 0.06163898482918739,
      "learning_rate": 8.591726509222242e-08,
      "loss": 0.0097,
      "step": 3220
    },
    {
      "epoch": 4.930827211143975,
      "grad_norm": 0.7225600481033325,
      "learning_rate": 5.752031068830266e-08,
      "loss": 0.0048,
      "step": 3230
    },
    {
      "epoch": 4.946092930063926,
      "grad_norm": 0.07761527597904205,
      "learning_rate": 3.4798873723984605e-08,
      "loss": 0.0013,
      "step": 3240
    },
    {
      "epoch": 4.961358648983875,
      "grad_norm": 0.5555828213691711,
      "learning_rate": 1.7755536287944464e-08,
      "loss": 0.0061,
      "step": 3250
    },
    {
      "epoch": 4.961358648983875,
      "eval_loss": 0.06215130165219307,
      "eval_runtime": 158.5647,
      "eval_samples_per_second": 7.347,
      "eval_steps_per_second": 7.347,
      "step": 3250
    },
    {
      "epoch": 4.976624367903826,
      "grad_norm": 0.009183814749121666,
      "learning_rate": 6.3922352034895276e-09,
      "loss": 0.0082,
      "step": 3260
    },
    {
      "epoch": 4.991890086823776,
      "grad_norm": 0.18164366483688354,
      "learning_rate": 7.102618084620094e-10,
      "loss": 0.005,
      "step": 3270
    },
    {
      "epoch": 4.999522946283752,
      "step": 3275,
      "total_flos": 8.22077121848279e+17,
      "train_loss": 0.04503766617872788,
      "train_runtime": 36493.551,
      "train_samples_per_second": 1.436,
      "train_steps_per_second": 0.09
    }
  ],
  "logging_steps": 10,
  "max_steps": 3275,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 8.22077121848279e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}