{
  "best_global_step": 900,
  "best_metric": 0.3843807876110077,
  "best_model_checkpoint": "./payment-extractor-optimized/checkpoint-900",
  "epoch": 4.391676866585067,
  "eval_steps": 100,
  "global_step": 900,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 2.608897936344147,
      "epoch": 0.04895960832313342,
      "grad_norm": 3.6901261806488037,
      "learning_rate": 3.6000000000000003e-06,
      "loss": 3.7818,
      "mean_token_accuracy": 0.4824275210499763,
      "num_tokens": 38801.0,
      "step": 10
    },
    {
      "entropy": 2.6241641879081725,
      "epoch": 0.09791921664626684,
      "grad_norm": 3.3132846355438232,
      "learning_rate": 7.600000000000001e-06,
      "loss": 3.8589,
      "mean_token_accuracy": 0.4720219738781452,
      "num_tokens": 69778.0,
      "step": 20
    },
    {
      "entropy": 2.4173066079616548,
      "epoch": 0.14687882496940025,
      "grad_norm": 3.6249752044677734,
      "learning_rate": 1.16e-05,
      "loss": 3.3912,
      "mean_token_accuracy": 0.5040820851922035,
      "num_tokens": 96530.0,
      "step": 30
    },
    {
      "entropy": 2.40450359582901,
      "epoch": 0.19583843329253367,
      "grad_norm": 2.8998804092407227,
      "learning_rate": 1.5600000000000003e-05,
      "loss": 2.8447,
      "mean_token_accuracy": 0.5414798364043236,
      "num_tokens": 120256.0,
      "step": 40
    },
    {
      "entropy": 2.284638965129852,
      "epoch": 0.24479804161566707,
      "grad_norm": 3.2823739051818848,
      "learning_rate": 1.9600000000000002e-05,
      "loss": 1.994,
      "mean_token_accuracy": 0.6505917206406593,
      "num_tokens": 142198.0,
      "step": 50
    },
    {
      "entropy": 2.4689642548561097,
      "epoch": 0.2937576499388005,
      "grad_norm": 2.1198339462280273,
      "learning_rate": 1.9995795492789368e-05,
      "loss": 2.407,
      "mean_token_accuracy": 0.6436298653483391,
      "num_tokens": 182368.0,
      "step": 60
    },
    {
      "entropy": 1.9469897836446761,
      "epoch": 0.3427172582619339,
      "grad_norm": 1.6485199928283691,
      "learning_rate": 1.9981265932877486e-05,
      "loss": 1.822,
      "mean_token_accuracy": 0.7381727218627929,
      "num_tokens": 213403.0,
      "step": 70
    },
    {
      "entropy": 1.2083469659090043,
      "epoch": 0.39167686658506734,
      "grad_norm": 1.6277607679367065,
      "learning_rate": 1.995637449278864e-05,
      "loss": 1.1033,
      "mean_token_accuracy": 0.8396818682551384,
      "num_tokens": 239775.0,
      "step": 80
    },
    {
      "entropy": 0.8737979799509048,
      "epoch": 0.44063647490820074,
      "grad_norm": 1.428305745124817,
      "learning_rate": 1.9921147013144782e-05,
      "loss": 0.819,
      "mean_token_accuracy": 0.8934117943048477,
      "num_tokens": 263407.0,
      "step": 90
    },
    {
      "entropy": 0.5950189992785454,
      "epoch": 0.48959608323133413,
      "grad_norm": 1.842503547668457,
      "learning_rate": 1.98756200647502e-05,
      "loss": 0.5144,
      "mean_token_accuracy": 0.9327697649598121,
      "num_tokens": 285288.0,
      "step": 100
    },
    {
      "epoch": 0.48959608323133413,
      "eval_entropy": 1.026704891055238,
      "eval_loss": 1.094792127609253,
      "eval_mean_token_accuracy": 0.8390047088557598,
      "eval_num_tokens": 285288.0,
      "eval_runtime": 21.2516,
      "eval_samples_per_second": 19.199,
      "eval_steps_per_second": 4.8,
      "step": 100
    },
    {
      "entropy": 1.7036595433950423,
      "epoch": 0.5385556915544676,
      "grad_norm": 1.3429725170135498,
      "learning_rate": 1.9819840910626174e-05,
      "loss": 1.8116,
      "mean_token_accuracy": 0.7263100475072861,
      "num_tokens": 325823.0,
      "step": 110
    },
    {
      "entropy": 1.3686422437429429,
      "epoch": 0.587515299877601,
      "grad_norm": 1.5084576606750488,
      "learning_rate": 1.9753867456945653e-05,
      "loss": 1.2567,
      "mean_token_accuracy": 0.8075757145881652,
      "num_tokens": 356947.0,
      "step": 120
    },
    {
      "entropy": 0.9899184465408325,
      "epoch": 0.6364749082007344,
      "grad_norm": 1.1403902769088745,
      "learning_rate": 1.9677768192918973e-05,
      "loss": 0.9506,
      "mean_token_accuracy": 0.8574481099843979,
      "num_tokens": 384860.0,
      "step": 130
    },
    {
      "entropy": 0.6585730135440826,
      "epoch": 0.6854345165238678,
      "grad_norm": 1.179526925086975,
      "learning_rate": 1.9591622119692953e-05,
      "loss": 0.6207,
      "mean_token_accuracy": 0.9152766704559326,
      "num_tokens": 408668.0,
      "step": 140
    },
    {
      "entropy": 0.49890025779604913,
      "epoch": 0.7343941248470012,
      "grad_norm": 2.6344473361968994,
      "learning_rate": 1.9495518668337204e-05,
      "loss": 0.4502,
      "mean_token_accuracy": 0.9399106308817864,
      "num_tokens": 430716.0,
      "step": 150
    },
    {
      "entropy": 1.4263419717550279,
      "epoch": 0.7833537331701347,
      "grad_norm": 1.8926013708114624,
      "learning_rate": 1.9389557607002808e-05,
      "loss": 1.4959,
      "mean_token_accuracy": 0.7735099658370018,
      "num_tokens": 468866.0,
      "step": 160
    },
    {
      "entropy": 1.0953898191452027,
      "epoch": 0.8323133414932681,
      "grad_norm": 0.972065806388855,
      "learning_rate": 1.9273848937349712e-05,
      "loss": 0.9593,
      "mean_token_accuracy": 0.8498499140143394,
      "num_tokens": 499808.0,
      "step": 170
    },
    {
      "entropy": 0.7438366547226906,
      "epoch": 0.8812729498164015,
      "grad_norm": 1.1401448249816895,
      "learning_rate": 1.9148512780350384e-05,
      "loss": 0.7447,
      "mean_token_accuracy": 0.885815504193306,
      "num_tokens": 526616.0,
      "step": 180
    },
    {
      "entropy": 0.5698909372091293,
      "epoch": 0.9302325581395349,
      "grad_norm": 1.2068321704864502,
      "learning_rate": 1.9013679251588304e-05,
      "loss": 0.5473,
      "mean_token_accuracy": 0.9253638312220573,
      "num_tokens": 550407.0,
      "step": 190
    },
    {
      "entropy": 0.44005816280841825,
      "epoch": 0.9791921664626683,
      "grad_norm": 2.358927011489868,
      "learning_rate": 1.8869488326180682e-05,
      "loss": 0.4075,
      "mean_token_accuracy": 0.9496751800179482,
      "num_tokens": 572396.0,
      "step": 200
    },
    {
      "epoch": 0.9791921664626683,
      "eval_entropy": 0.7869388329632142,
      "eval_loss": 0.8091182112693787,
      "eval_mean_token_accuracy": 0.8807467134559855,
      "eval_num_tokens": 572396.0,
      "eval_runtime": 21.2312,
      "eval_samples_per_second": 19.217,
      "eval_steps_per_second": 4.804,
      "step": 200
    },
    {
      "entropy": 1.2279899853306848,
      "epoch": 1.0244798041615668,
      "grad_norm": 2.1428945064544678,
      "learning_rate": 1.8716089693465696e-05,
      "loss": 1.1313,
      "mean_token_accuracy": 0.8170017583950145,
      "num_tokens": 606225.0,
      "step": 210
    },
    {
      "entropy": 1.1651067942380906,
      "epoch": 1.0734394124847002,
      "grad_norm": 1.0429226160049438,
      "learning_rate": 1.855364260160507e-05,
      "loss": 1.0415,
      "mean_token_accuracy": 0.8387726441025734,
      "num_tokens": 639876.0,
      "step": 220
    },
    {
      "entropy": 0.8667519375681877,
      "epoch": 1.1223990208078336,
      "grad_norm": 1.605554461479187,
      "learning_rate": 1.8382315692263324e-05,
      "loss": 0.8179,
      "mean_token_accuracy": 0.8734943136572838,
      "num_tokens": 669254.0,
      "step": 230
    },
    {
      "entropy": 0.5002034060657025,
      "epoch": 1.171358629130967,
      "grad_norm": 1.766509771347046,
      "learning_rate": 1.820228682553533e-05,
      "loss": 0.5005,
      "mean_token_accuracy": 0.9262110635638237,
      "num_tokens": 693566.0,
      "step": 240
    },
    {
      "entropy": 0.4824553743004799,
      "epoch": 1.2203182374541004,
      "grad_norm": 1.3010213375091553,
      "learning_rate": 1.8013742895303883e-05,
      "loss": 0.4382,
      "mean_token_accuracy": 0.9407104596495628,
      "num_tokens": 716324.0,
      "step": 250
    },
    {
      "entropy": 0.886764119565487,
      "epoch": 1.2692778457772338,
      "grad_norm": 2.592423915863037,
      "learning_rate": 1.7816879635219028e-05,
      "loss": 0.8545,
      "mean_token_accuracy": 0.8653567478060722,
      "num_tokens": 749252.0,
      "step": 260
    },
    {
      "entropy": 0.9997715935111046,
      "epoch": 1.3182374541003672,
      "grad_norm": 1.1456091403961182,
      "learning_rate": 1.7611901415500536e-05,
      "loss": 0.9247,
      "mean_token_accuracy": 0.8514459431171417,
      "num_tokens": 782122.0,
      "step": 270
    },
    {
      "entropy": 0.8610424906015396,
      "epoch": 1.3671970624235006,
      "grad_norm": 1.3674919605255127,
      "learning_rate": 1.7399021030774443e-05,
      "loss": 0.7975,
      "mean_token_accuracy": 0.8724059730768203,
      "num_tokens": 812066.0,
      "step": 280
    },
    {
      "entropy": 0.4803216062486172,
      "epoch": 1.416156670746634,
      "grad_norm": 1.0078914165496826,
      "learning_rate": 1.717845947916398e-05,
      "loss": 0.4703,
      "mean_token_accuracy": 0.9251162573695183,
      "num_tokens": 836705.0,
      "step": 290
    },
    {
      "entropy": 0.477826127409935,
      "epoch": 1.4651162790697674,
      "grad_norm": 1.0581692457199097,
      "learning_rate": 1.695044573286413e-05,
      "loss": 0.4365,
      "mean_token_accuracy": 0.9394737258553505,
      "num_tokens": 859600.0,
      "step": 300
    },
    {
      "epoch": 1.4651162790697674,
      "eval_entropy": 0.7299714935760871,
      "eval_loss": 0.6802113056182861,
      "eval_mean_token_accuracy": 0.8902294659147075,
      "eval_num_tokens": 859600.0,
      "eval_runtime": 21.2433,
      "eval_samples_per_second": 19.206,
      "eval_steps_per_second": 4.802,
      "step": 300
    },
    {
      "entropy": 0.8674551732838154,
      "epoch": 1.514075887392901,
      "grad_norm": 2.246056079864502,
      "learning_rate": 1.6715216500438093e-05,
      "loss": 0.7837,
      "mean_token_accuracy": 0.874087019264698,
      "num_tokens": 891961.0,
      "step": 310
    },
    {
      "entropy": 0.9596038445830345,
      "epoch": 1.5630354957160342,
      "grad_norm": 1.619541883468628,
      "learning_rate": 1.647301598108234e-05,
      "loss": 0.8646,
      "mean_token_accuracy": 0.857255433499813,
      "num_tokens": 924600.0,
      "step": 320
    },
    {
      "entropy": 0.7986739322543144,
      "epoch": 1.6119951040391678,
      "grad_norm": 2.2831315994262695,
      "learning_rate": 1.6224095611115385e-05,
      "loss": 0.7373,
      "mean_token_accuracy": 0.8797152638435364,
      "num_tokens": 954438.0,
      "step": 330
    },
    {
      "entropy": 0.47184758856892584,
      "epoch": 1.660954712362301,
      "grad_norm": 2.6723692417144775,
      "learning_rate": 1.596871380295351e-05,
      "loss": 0.4218,
      "mean_token_accuracy": 0.9275540292263031,
      "num_tokens": 979015.0,
      "step": 340
    },
    {
      "entropy": 0.45431587770581244,
      "epoch": 1.7099143206854346,
      "grad_norm": 2.8102011680603027,
      "learning_rate": 1.570713567684432e-05,
      "loss": 0.3684,
      "mean_token_accuracy": 0.9440548986196518,
      "num_tokens": 1001989.0,
      "step": 350
    },
    {
      "entropy": 0.7829712487757206,
      "epoch": 1.758873929008568,
      "grad_norm": 1.7116206884384155,
      "learning_rate": 1.5439632785636707e-05,
      "loss": 0.663,
      "mean_token_accuracy": 0.8908370733261108,
      "num_tokens": 1034477.0,
      "step": 360
    },
    {
      "entropy": 0.8193635582923889,
      "epoch": 1.8078335373317014,
      "grad_norm": 0.9935292601585388,
      "learning_rate": 1.5166482832872923e-05,
      "loss": 0.7506,
      "mean_token_accuracy": 0.8711828157305718,
      "num_tokens": 1067534.0,
      "step": 370
    },
    {
      "entropy": 0.7036949403584003,
      "epoch": 1.8567931456548348,
      "grad_norm": 1.7203794717788696,
      "learning_rate": 1.4887969384495403e-05,
      "loss": 0.6792,
      "mean_token_accuracy": 0.88548723757267,
      "num_tokens": 1097280.0,
      "step": 380
    },
    {
      "entropy": 0.3680045209825039,
      "epoch": 1.9057527539779682,
      "grad_norm": 1.4336254596710205,
      "learning_rate": 1.4604381574467616e-05,
      "loss": 0.3249,
      "mean_token_accuracy": 0.9348379656672478,
      "num_tokens": 1121920.0,
      "step": 390
    },
    {
      "entropy": 0.37416374161839483,
      "epoch": 1.9547123623011016,
      "grad_norm": 1.0702754259109497,
      "learning_rate": 1.4316013804614644e-05,
      "loss": 0.2929,
      "mean_token_accuracy": 0.9463742494583129,
      "num_tokens": 1145002.0,
      "step": 400
    },
    {
      "epoch": 1.9547123623011016,
      "eval_entropy": 0.6033922105151064,
      "eval_loss": 0.49456480145454407,
      "eval_mean_token_accuracy": 0.9079791880121418,
      "eval_num_tokens": 1145002.0,
      "eval_runtime": 21.1773,
      "eval_samples_per_second": 19.266,
      "eval_steps_per_second": 4.816,
      "step": 400
    },
    {
      "entropy": 0.41346504196927353,
      "epoch": 2.0,
      "grad_norm": 0.9998394250869751,
      "learning_rate": 1.4023165438994933e-05,
      "loss": 0.3152,
      "mean_token_accuracy": 0.9379849611101924,
      "num_tokens": 1167618.0,
      "step": 410
    },
    {
      "entropy": 1.0350674405694007,
      "epoch": 2.0489596083231336,
      "grad_norm": 1.5349912643432617,
      "learning_rate": 1.3726140493120639e-05,
      "loss": 0.8548,
      "mean_token_accuracy": 0.8497318252921104,
      "num_tokens": 1208219.0,
      "step": 420
    },
    {
      "entropy": 0.6850787699222565,
      "epoch": 2.097919216646267,
      "grad_norm": 0.9396725296974182,
      "learning_rate": 1.3425247318349137e-05,
      "loss": 0.6238,
      "mean_token_accuracy": 0.8850513309240341,
      "num_tokens": 1239429.0,
      "step": 430
    },
    {
      "entropy": 0.5261640004813671,
      "epoch": 2.1468788249694004,
      "grad_norm": 1.1289340257644653,
      "learning_rate": 1.3120798281773346e-05,
      "loss": 0.4823,
      "mean_token_accuracy": 0.9071076571941376,
      "num_tokens": 1266803.0,
      "step": 440
    },
    {
      "entropy": 0.3201193898916245,
      "epoch": 2.1958384332925336,
      "grad_norm": 0.8056601285934448,
      "learning_rate": 1.2813109441943166e-05,
      "loss": 0.2848,
      "mean_token_accuracy": 0.943960489332676,
      "num_tokens": 1290548.0,
      "step": 450
    },
    {
      "entropy": 0.24066586568951606,
      "epoch": 2.244798041615667,
      "grad_norm": 1.4606080055236816,
      "learning_rate": 1.2502500220754736e-05,
      "loss": 0.198,
      "mean_token_accuracy": 0.9593303337693214,
      "num_tokens": 1312508.0,
      "step": 460
    },
    {
      "entropy": 0.9151711657643318,
      "epoch": 2.2937576499388004,
      "grad_norm": 1.4147366285324097,
      "learning_rate": 1.2189293071848051e-05,
      "loss": 0.7861,
      "mean_token_accuracy": 0.8564901977777482,
      "num_tokens": 1351246.0,
      "step": 470
    },
    {
      "entropy": 0.6647965341806412,
      "epoch": 2.342717258261934,
      "grad_norm": 0.8306525349617004,
      "learning_rate": 1.187381314585725e-05,
      "loss": 0.6222,
      "mean_token_accuracy": 0.8851410359144211,
      "num_tokens": 1382196.0,
      "step": 480
    },
    {
      "entropy": 0.4626566760241985,
      "epoch": 2.391676866585067,
      "grad_norm": 1.2030415534973145,
      "learning_rate": 1.1556387952861036e-05,
      "loss": 0.4374,
      "mean_token_accuracy": 0.9146721437573433,
      "num_tokens": 1409050.0,
      "step": 490
    },
    {
      "entropy": 0.292323163151741,
      "epoch": 2.440636474908201,
      "grad_norm": 0.6598659753799438,
      "learning_rate": 1.1237347022383747e-05,
      "loss": 0.2684,
      "mean_token_accuracy": 0.9467211216688156,
      "num_tokens": 1432781.0,
      "step": 500
    },
    {
      "epoch": 2.440636474908201,
      "eval_entropy": 0.5070152823247162,
      "eval_loss": 0.4397956132888794,
      "eval_mean_token_accuracy": 0.9159132575287539,
      "eval_num_tokens": 1432781.0,
      "eval_runtime": 21.252,
      "eval_samples_per_second": 19.198,
      "eval_steps_per_second": 4.8,
      "step": 500
    },
    {
      "entropy": 0.22589576356112956,
      "epoch": 2.489596083231334,
      "grad_norm": 0.9443103075027466,
      "learning_rate": 1.0917021561299864e-05,
      "loss": 0.191,
      "mean_token_accuracy": 0.9594075292348861,
      "num_tokens": 1454735.0,
      "step": 510
    },
    {
      "entropy": 0.8820951759815217,
      "epoch": 2.5385556915544676,
      "grad_norm": 1.3829594850540161,
      "learning_rate": 1.0595744109997326e-05,
      "loss": 0.7404,
      "mean_token_accuracy": 0.8678841248154641,
      "num_tokens": 1494090.0,
      "step": 520
    },
    {
      "entropy": 0.6331574842333794,
      "epoch": 2.5875152998776008,
      "grad_norm": 1.056781530380249,
      "learning_rate": 1.0273848197156401e-05,
      "loss": 0.5978,
      "mean_token_accuracy": 0.888825386762619,
      "num_tokens": 1525062.0,
      "step": 530
    },
    {
      "entropy": 0.4361429732292891,
      "epoch": 2.6364749082007344,
      "grad_norm": 1.1021668910980225,
      "learning_rate": 9.951667993502599e-06,
      "loss": 0.3999,
      "mean_token_accuracy": 0.9207724243402481,
      "num_tokens": 1551813.0,
      "step": 540
    },
    {
      "entropy": 0.2877490069717169,
      "epoch": 2.685434516523868,
      "grad_norm": 0.7195802927017212,
      "learning_rate": 9.629537964893063e-06,
      "loss": 0.2665,
      "mean_token_accuracy": 0.9467098742723465,
      "num_tokens": 1575433.0,
      "step": 550
    },
    {
      "entropy": 0.2189638450741768,
      "epoch": 2.734394124847001,
      "grad_norm": 1.2461501359939575,
      "learning_rate": 9.307792525096582e-06,
      "loss": 0.1899,
      "mean_token_accuracy": 0.9602582737803459,
      "num_tokens": 1597306.0,
      "step": 560
    },
    {
      "entropy": 0.825089768320322,
      "epoch": 2.783353733170135,
      "grad_norm": 1.1732733249664307,
      "learning_rate": 8.986765688627652e-06,
      "loss": 0.6871,
      "mean_token_accuracy": 0.8730710700154305,
      "num_tokens": 1636454.0,
      "step": 570
    },
    {
      "entropy": 0.6250899910926819,
      "epoch": 2.832313341493268,
      "grad_norm": 0.8510639071464539,
      "learning_rate": 8.666790723995043e-06,
      "loss": 0.597,
      "mean_token_accuracy": 0.8902681171894073,
      "num_tokens": 1667449.0,
      "step": 580
    },
    {
      "entropy": 0.4255227465182543,
      "epoch": 2.8812729498164016,
      "grad_norm": 0.7777973413467407,
      "learning_rate": 8.348199807724806e-06,
      "loss": 0.3826,
      "mean_token_accuracy": 0.921813291311264,
      "num_tokens": 1694019.0,
      "step": 590
    },
    {
      "entropy": 0.280773538723588,
      "epoch": 2.9302325581395348,
      "grad_norm": 0.7043775916099548,
      "learning_rate": 8.0313236795169e-06,
      "loss": 0.2658,
      "mean_token_accuracy": 0.9461780115962029,
      "num_tokens": 1717763.0,
      "step": 600
    },
    {
      "epoch": 2.9302325581395348,
      "eval_entropy": 0.4671570363582349,
      "eval_loss": 0.41600170731544495,
      "eval_mean_token_accuracy": 0.919863361938327,
      "eval_num_tokens": 1717763.0,
      "eval_runtime": 21.0643,
      "eval_samples_per_second": 19.369,
      "eval_steps_per_second": 4.842,
      "step": 600
    },
    {
      "entropy": 0.22502716928720473,
      "epoch": 2.9791921664626684,
      "grad_norm": 1.245797872543335,
      "learning_rate": 7.716491298893443e-06,
      "loss": 0.2014,
      "mean_token_accuracy": 0.9587241023778915,
      "num_tokens": 1739700.0,
      "step": 610
    },
    {
      "entropy": 0.7121120212045876,
      "epoch": 3.0244798041615666,
      "grad_norm": 1.578238606452942,
      "learning_rate": 7.404029503695028e-06,
      "loss": 0.592,
      "mean_token_accuracy": 0.884317248254209,
      "num_tokens": 1773279.0,
      "step": 620
    },
    {
      "entropy": 0.6201752804219722,
      "epoch": 3.0734394124847,
      "grad_norm": 0.9049448370933533,
      "learning_rate": 7.094262670779611e-06,
      "loss": 0.5563,
      "mean_token_accuracy": 0.8980106115341187,
      "num_tokens": 1806614.0,
      "step": 630
    },
    {
      "entropy": 0.5976772040128708,
      "epoch": 3.1223990208078334,
      "grad_norm": 1.2108434438705444,
      "learning_rate": 6.78751237927623e-06,
      "loss": 0.5527,
      "mean_token_accuracy": 0.898520989716053,
      "num_tokens": 1836366.0,
      "step": 640
    },
    {
      "entropy": 0.26630178317427633,
      "epoch": 3.171358629130967,
      "grad_norm": 0.9259095788002014,
      "learning_rate": 6.48409707674317e-06,
      "loss": 0.2551,
      "mean_token_accuracy": 0.9447213500738144,
      "num_tokens": 1860872.0,
      "step": 650
    },
    {
      "entropy": 0.2778078857809305,
      "epoch": 3.2203182374541,
      "grad_norm": 1.1952687501907349,
      "learning_rate": 6.18433174857705e-06,
      "loss": 0.2529,
      "mean_token_accuracy": 0.9497045069932938,
      "num_tokens": 1883874.0,
      "step": 660
    },
    {
      "entropy": 0.5209752138704061,
      "epoch": 3.269277845777234,
      "grad_norm": 1.9098657369613647,
      "learning_rate": 5.8885275910161574e-06,
      "loss": 0.4378,
      "mean_token_accuracy": 0.9142355337738991,
      "num_tokens": 1917087.0,
      "step": 670
    },
    {
      "entropy": 0.64239452034235,
      "epoch": 3.318237454100367,
      "grad_norm": 0.8732991814613342,
      "learning_rate": 5.596991688077409e-06,
      "loss": 0.5759,
      "mean_token_accuracy": 0.8934120118618012,
      "num_tokens": 1949777.0,
      "step": 680
    },
    {
      "entropy": 0.5809963166713714,
      "epoch": 3.3671970624235006,
      "grad_norm": 0.8436909317970276,
      "learning_rate": 5.310026692762316e-06,
      "loss": 0.5335,
      "mean_token_accuracy": 0.8995566830039025,
      "num_tokens": 1978937.0,
      "step": 690
    },
    {
      "entropy": 0.2611778501421213,
      "epoch": 3.416156670746634,
      "grad_norm": 1.1178691387176514,
      "learning_rate": 5.027930512862976e-06,
      "loss": 0.2472,
      "mean_token_accuracy": 0.9472135573625564,
      "num_tokens": 2003325.0,
      "step": 700
    },
    {
      "epoch": 3.416156670746634,
      "eval_entropy": 0.4348171895333365,
      "eval_loss": 0.39616772532463074,
      "eval_mean_token_accuracy": 0.9233323300586027,
      "eval_num_tokens": 2003325.0,
      "eval_runtime": 21.154,
      "eval_samples_per_second": 19.287,
      "eval_steps_per_second": 4.822,
      "step": 700
    },
    {
      "entropy": 0.25852978341281413,
      "epoch": 3.4651162790697674,
      "grad_norm": 1.280714511871338,
      "learning_rate": 4.750996001694215e-06,
      "loss": 0.2374,
      "mean_token_accuracy": 0.9529847070574761,
      "num_tokens": 2026113.0,
      "step": 710
    },
    {
      "entropy": 0.5000423431396485,
      "epoch": 3.514075887392901,
      "grad_norm": 1.6330279111862183,
      "learning_rate": 4.479510654072909e-06,
      "loss": 0.4167,
      "mean_token_accuracy": 0.9191933527588845,
      "num_tokens": 2058740.0,
      "step": 720
    },
    {
      "entropy": 0.5997287526726722,
      "epoch": 3.563035495716034,
      "grad_norm": 0.8560138940811157,
      "learning_rate": 4.213756307860175e-06,
      "loss": 0.5358,
      "mean_token_accuracy": 0.8984866350889206,
      "num_tokens": 2091634.0,
      "step": 730
    },
    {
      "entropy": 0.5770318634808064,
      "epoch": 3.611995104039168,
      "grad_norm": 1.2353603839874268,
      "learning_rate": 3.954008851376252e-06,
      "loss": 0.5377,
      "mean_token_accuracy": 0.9002507776021957,
      "num_tokens": 2121380.0,
      "step": 740
    },
    {
      "entropy": 0.26896645687520504,
      "epoch": 3.660954712362301,
      "grad_norm": 0.5919920802116394,
      "learning_rate": 3.700537936991733e-06,
      "loss": 0.2541,
      "mean_token_accuracy": 0.9460319548845291,
      "num_tokens": 2145937.0,
      "step": 750
    },
    {
      "entropy": 0.27403030432760717,
      "epoch": 3.7099143206854346,
      "grad_norm": 0.9070358276367188,
      "learning_rate": 3.4536067011925945e-06,
      "loss": 0.2555,
      "mean_token_accuracy": 0.9493874981999397,
      "num_tokens": 2168886.0,
      "step": 760
    },
    {
      "entropy": 0.49704274982213975,
      "epoch": 3.758873929008568,
      "grad_norm": 1.7363125085830688,
      "learning_rate": 3.213471491409568e-06,
      "loss": 0.42,
      "mean_token_accuracy": 0.9207542642951012,
      "num_tokens": 2201618.0,
      "step": 770
    },
    {
      "entropy": 0.6181983411312103,
      "epoch": 3.8078335373317014,
      "grad_norm": 0.8889790773391724,
      "learning_rate": 2.9803815998954334e-06,
      "loss": 0.5448,
      "mean_token_accuracy": 0.8981723725795746,
      "num_tokens": 2234716.0,
      "step": 780
    },
    {
      "entropy": 0.5966475777328014,
      "epoch": 3.8567931456548346,
      "grad_norm": 1.2056671380996704,
      "learning_rate": 2.7545790049265506e-06,
      "loss": 0.5489,
      "mean_token_accuracy": 0.8978816866874695,
      "num_tokens": 2264636.0,
      "step": 790
    },
    {
      "entropy": 0.2779835805296898,
      "epoch": 3.905752753977968,
      "grad_norm": 0.9485862851142883,
      "learning_rate": 2.5362981195972627e-06,
      "loss": 0.2623,
      "mean_token_accuracy": 0.9445345297455787,
      "num_tokens": 2289186.0,
      "step": 800
    },
    {
      "epoch": 3.905752753977968,
      "eval_entropy": 0.42678100907919453,
      "eval_loss": 0.3873916268348694,
      "eval_mean_token_accuracy": 0.9257115113968942,
      "eval_num_tokens": 2289186.0,
      "eval_runtime": 21.0498,
      "eval_samples_per_second": 19.383,
      "eval_steps_per_second": 4.846,
      "step": 800
    },
    {
      "entropy": 0.26332913562655447,
      "epoch": 3.954712362301102,
      "grad_norm": 0.8345707058906555,
      "learning_rate": 2.3257655484679376e-06,
      "loss": 0.2408,
      "mean_token_accuracy": 0.9522591263055802,
      "num_tokens": 2312074.0,
      "step": 810
    },
    {
      "entropy": 0.31254331364824967,
      "epoch": 4.0,
      "grad_norm": 1.0298808813095093,
      "learning_rate": 2.123199852319352e-06,
      "loss": 0.2697,
      "mean_token_accuracy": 0.9450424829044858,
      "num_tokens": 2335236.0,
      "step": 820
    },
    {
      "entropy": 0.6749883458018303,
      "epoch": 4.048959608323133,
      "grad_norm": 0.9510061740875244,
      "learning_rate": 1.9288113212575454e-06,
      "loss": 0.5564,
      "mean_token_accuracy": 0.8976946637034416,
      "num_tokens": 2374400.0,
      "step": 830
    },
    {
      "entropy": 0.5979305505752563,
      "epoch": 4.097919216646267,
      "grad_norm": 0.8506686687469482,
      "learning_rate": 1.7428017564047594e-06,
      "loss": 0.562,
      "mean_token_accuracy": 0.8973473310470581,
      "num_tokens": 2405288.0,
      "step": 840
    },
    {
      "entropy": 0.38541266694664955,
      "epoch": 4.1468788249694,
      "grad_norm": 0.8283617496490479,
      "learning_rate": 1.565364260403055e-06,
      "loss": 0.3559,
      "mean_token_accuracy": 0.9307585805654526,
      "num_tokens": 2431701.0,
      "step": 850
    },
    {
      "entropy": 0.26541123948991296,
      "epoch": 4.195838433292534,
      "grad_norm": 0.6910677552223206,
      "learning_rate": 1.3966830369481231e-06,
      "loss": 0.2495,
      "mean_token_accuracy": 0.9491465017199516,
      "num_tokens": 2455249.0,
      "step": 860
    },
    {
      "entropy": 0.19118896164000035,
      "epoch": 4.244798041615667,
      "grad_norm": 0.9370400905609131,
      "learning_rate": 1.2369331995613664e-06,
      "loss": 0.1709,
      "mean_token_accuracy": 0.9622136607766152,
      "num_tokens": 2477038.0,
      "step": 870
    },
    {
      "entropy": 0.7027731344103814,
      "epoch": 4.293757649938801,
      "grad_norm": 1.2064703702926636,
      "learning_rate": 1.0862805897987894e-06,
      "loss": 0.5842,
      "mean_token_accuracy": 0.8925257995724678,
      "num_tokens": 2516496.0,
      "step": 880
    },
    {
      "entropy": 0.5964450292289257,
      "epoch": 4.342717258261934,
      "grad_norm": 1.0601837635040283,
      "learning_rate": 9.448816050854559e-07,
      "loss": 0.5577,
      "mean_token_accuracy": 0.8974330082535744,
      "num_tokens": 2547663.0,
      "step": 890
    },
    {
      "entropy": 0.4454402156174183,
      "epoch": 4.391676866585067,
      "grad_norm": 0.7408022880554199,
      "learning_rate": 8.128830363541574e-07,
      "loss": 0.4126,
      "mean_token_accuracy": 0.9200813502073288,
      "num_tokens": 2575037.0,
      "step": 900
    },
    {
      "epoch": 4.391676866585067,
      "eval_entropy": 0.4241247632924248,
      "eval_loss": 0.3843807876110077,
      "eval_mean_token_accuracy": 0.9258733800813264,
      "eval_num_tokens": 2575037.0,
      "eval_runtime": 21.0482,
      "eval_samples_per_second": 19.384,
      "eval_steps_per_second": 4.846,
      "step": 900
    }
  ],
  "logging_steps": 10,
  "max_steps": 1025,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.525760757970944e+16,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}