{
  "best_metric": 0.2531912922859192,
  "best_model_checkpoint": "./ryan_model314/checkpoint-600",
  "epoch": 4.0,
  "eval_steps": 100,
  "global_step": 2500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "grad_norm": 1.1103402376174927,
      "learning_rate": 0.00019920000000000002,
      "loss": 0.5731,
      "step": 10
    },
    {
      "epoch": 0.03,
      "grad_norm": 0.8193413019180298,
      "learning_rate": 0.0001984,
      "loss": 0.4217,
      "step": 20
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.3536686897277832,
      "learning_rate": 0.0001976,
      "loss": 0.3709,
      "step": 30
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.9998810887336731,
      "learning_rate": 0.0001968,
      "loss": 0.3398,
      "step": 40
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.5689244270324707,
      "learning_rate": 0.000196,
      "loss": 0.3346,
      "step": 50
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.1778826713562012,
      "learning_rate": 0.0001952,
      "loss": 0.3406,
      "step": 60
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.3193926811218262,
      "learning_rate": 0.0001944,
      "loss": 0.2755,
      "step": 70
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.1302804946899414,
      "learning_rate": 0.00019360000000000002,
      "loss": 0.3944,
      "step": 80
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.8255844712257385,
      "learning_rate": 0.0001928,
      "loss": 0.3473,
      "step": 90
    },
    {
      "epoch": 0.16,
      "grad_norm": 1.0871790647506714,
      "learning_rate": 0.000192,
      "loss": 0.3042,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_loss": 0.3673088252544403,
      "eval_na_accuracy": 0.928,
      "eval_ordinal_accuracy": 0.4671280276816609,
      "eval_runtime": 110.8646,
      "eval_samples_per_second": 9.02,
      "eval_steps_per_second": 1.128,
      "step": 100
    },
    {
      "epoch": 0.18,
      "grad_norm": 1.5816177129745483,
      "learning_rate": 0.0001912,
      "loss": 0.4058,
      "step": 110
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.8431822061538696,
      "learning_rate": 0.0001904,
      "loss": 0.2781,
      "step": 120
    },
    {
      "epoch": 0.21,
      "grad_norm": 1.0826754570007324,
      "learning_rate": 0.0001896,
      "loss": 0.2587,
      "step": 130
    },
    {
      "epoch": 0.22,
      "grad_norm": 3.7366294860839844,
      "learning_rate": 0.0001888,
      "loss": 0.3432,
      "step": 140
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.5233088731765747,
      "learning_rate": 0.000188,
      "loss": 0.3484,
      "step": 150
    },
    {
      "epoch": 0.26,
      "grad_norm": 1.8766111135482788,
      "learning_rate": 0.00018720000000000002,
      "loss": 0.3597,
      "step": 160
    },
    {
      "epoch": 0.27,
      "grad_norm": 1.0037935972213745,
      "learning_rate": 0.00018640000000000003,
      "loss": 0.288,
      "step": 170
    },
    {
      "epoch": 0.29,
      "grad_norm": 1.3281046152114868,
      "learning_rate": 0.0001856,
      "loss": 0.3207,
      "step": 180
    },
    {
      "epoch": 0.3,
      "grad_norm": 1.4793013334274292,
      "learning_rate": 0.00018480000000000002,
      "loss": 0.3372,
      "step": 190
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.8796727657318115,
      "learning_rate": 0.00018400000000000003,
      "loss": 0.2904,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_loss": 0.29769936203956604,
      "eval_na_accuracy": 0.933,
      "eval_ordinal_accuracy": 0.5790080738177624,
      "eval_runtime": 39.6284,
      "eval_samples_per_second": 25.234,
      "eval_steps_per_second": 3.154,
      "step": 200
    },
    {
      "epoch": 0.34,
      "grad_norm": 1.3502057790756226,
      "learning_rate": 0.0001832,
      "loss": 0.3519,
      "step": 210
    },
    {
      "epoch": 0.35,
      "grad_norm": 1.5546174049377441,
      "learning_rate": 0.00018240000000000002,
      "loss": 0.3243,
      "step": 220
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.7677227854728699,
      "learning_rate": 0.00018160000000000002,
      "loss": 0.2914,
      "step": 230
    },
    {
      "epoch": 0.38,
      "grad_norm": 1.1754639148712158,
      "learning_rate": 0.0001808,
      "loss": 0.3539,
      "step": 240
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8472470641136169,
      "learning_rate": 0.00018,
      "loss": 0.2395,
      "step": 250
    },
    {
      "epoch": 0.42,
      "grad_norm": 1.1917964220046997,
      "learning_rate": 0.00017920000000000002,
      "loss": 0.2295,
      "step": 260
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.7398644685745239,
      "learning_rate": 0.0001784,
      "loss": 0.2398,
      "step": 270
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.5953208804130554,
      "learning_rate": 0.0001776,
      "loss": 0.2786,
      "step": 280
    },
    {
      "epoch": 0.46,
      "grad_norm": 2.0648913383483887,
      "learning_rate": 0.00017680000000000001,
      "loss": 0.3661,
      "step": 290
    },
    {
      "epoch": 0.48,
      "grad_norm": 1.4048805236816406,
      "learning_rate": 0.00017600000000000002,
      "loss": 0.2648,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_loss": 0.2830840051174164,
      "eval_na_accuracy": 0.944,
      "eval_ordinal_accuracy": 0.5940023068050749,
      "eval_runtime": 39.7255,
      "eval_samples_per_second": 25.173,
      "eval_steps_per_second": 3.147,
      "step": 300
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8102580904960632,
      "learning_rate": 0.0001752,
      "loss": 0.2359,
      "step": 310
    },
    {
      "epoch": 0.51,
      "grad_norm": 2.0220913887023926,
      "learning_rate": 0.0001744,
      "loss": 0.2557,
      "step": 320
    },
    {
      "epoch": 0.53,
      "grad_norm": 1.2111886739730835,
      "learning_rate": 0.00017360000000000002,
      "loss": 0.3025,
      "step": 330
    },
    {
      "epoch": 0.54,
      "grad_norm": 1.788378119468689,
      "learning_rate": 0.0001728,
      "loss": 0.3067,
      "step": 340
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.7332974076271057,
      "learning_rate": 0.000172,
      "loss": 0.2612,
      "step": 350
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.5220205783843994,
      "learning_rate": 0.00017120000000000001,
      "loss": 0.2924,
      "step": 360
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8991191387176514,
      "learning_rate": 0.0001704,
      "loss": 0.2379,
      "step": 370
    },
    {
      "epoch": 0.61,
      "grad_norm": 1.6633837223052979,
      "learning_rate": 0.0001696,
      "loss": 0.2792,
      "step": 380
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.9553330540657043,
      "learning_rate": 0.0001688,
      "loss": 0.2512,
      "step": 390
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.4544942080974579,
      "learning_rate": 0.000168,
      "loss": 0.3036,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_loss": 0.27759096026420593,
      "eval_na_accuracy": 0.949,
      "eval_ordinal_accuracy": 0.5870818915801614,
      "eval_runtime": 39.7611,
      "eval_samples_per_second": 25.15,
      "eval_steps_per_second": 3.144,
      "step": 400
    },
    {
      "epoch": 0.66,
      "grad_norm": 1.2751814126968384,
      "learning_rate": 0.0001672,
      "loss": 0.3042,
      "step": 410
    },
    {
      "epoch": 0.67,
      "grad_norm": 1.791074514389038,
      "learning_rate": 0.0001664,
      "loss": 0.3341,
      "step": 420
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.9887642860412598,
      "learning_rate": 0.0001656,
      "loss": 0.2868,
      "step": 430
    },
    {
      "epoch": 0.7,
      "grad_norm": 1.3511923551559448,
      "learning_rate": 0.0001648,
      "loss": 0.3763,
      "step": 440
    },
    {
      "epoch": 0.72,
      "grad_norm": 1.7992609739303589,
      "learning_rate": 0.000164,
      "loss": 0.2264,
      "step": 450
    },
    {
      "epoch": 0.74,
      "grad_norm": 1.0241813659667969,
      "learning_rate": 0.0001632,
      "loss": 0.3018,
      "step": 460
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.628193736076355,
      "learning_rate": 0.00016240000000000002,
      "loss": 0.3323,
      "step": 470
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8471026420593262,
      "learning_rate": 0.00016160000000000002,
      "loss": 0.2005,
      "step": 480
    },
    {
      "epoch": 0.78,
      "grad_norm": 1.1799852848052979,
      "learning_rate": 0.0001608,
      "loss": 0.2984,
      "step": 490
    },
    {
      "epoch": 0.8,
      "grad_norm": 1.595058560371399,
      "learning_rate": 0.00016,
      "loss": 0.2656,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_loss": 0.2846027612686157,
      "eval_na_accuracy": 0.931,
      "eval_ordinal_accuracy": 0.6101499423298731,
      "eval_runtime": 39.9154,
      "eval_samples_per_second": 25.053,
      "eval_steps_per_second": 3.132,
      "step": 500
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.9246352910995483,
      "learning_rate": 0.00015920000000000002,
      "loss": 0.3591,
      "step": 510
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.9456105828285217,
      "learning_rate": 0.00015840000000000003,
      "loss": 0.3569,
      "step": 520
    },
    {
      "epoch": 0.85,
      "grad_norm": 1.111274003982544,
      "learning_rate": 0.0001576,
      "loss": 0.3243,
      "step": 530
    },
    {
      "epoch": 0.86,
      "grad_norm": NaN,
      "learning_rate": 0.00015688,
      "loss": 0.2911,
      "step": 540
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8232502341270447,
      "learning_rate": 0.00015616000000000002,
      "loss": 0.3236,
      "step": 550
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.6359846591949463,
      "learning_rate": 0.00015536,
      "loss": 0.3211,
      "step": 560
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.545005738735199,
      "learning_rate": 0.00015456,
      "loss": 0.205,
      "step": 570
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.6029797196388245,
      "learning_rate": 0.00015376000000000002,
      "loss": 0.1928,
      "step": 580
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.7442355155944824,
      "learning_rate": 0.00015296000000000003,
      "loss": 0.3273,
      "step": 590
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.6751519441604614,
      "learning_rate": 0.00015216,
      "loss": 0.2954,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_loss": 0.2531912922859192,
      "eval_na_accuracy": 0.947,
      "eval_ordinal_accuracy": 0.5951557093425606,
      "eval_runtime": 39.4037,
      "eval_samples_per_second": 25.378,
      "eval_steps_per_second": 3.172,
      "step": 600
    },
    {
      "epoch": 0.98,
      "grad_norm": 1.6347012519836426,
      "learning_rate": 0.00015136000000000001,
      "loss": 0.2256,
      "step": 610
    },
    {
      "epoch": 0.99,
      "grad_norm": 6.180319309234619,
      "learning_rate": 0.00015056000000000002,
      "loss": 0.2003,
      "step": 620
    },
    {
      "epoch": 1.01,
      "grad_norm": 0.8919633626937866,
      "learning_rate": 0.00014976,
      "loss": 0.198,
      "step": 630
    },
    {
      "epoch": 1.02,
      "grad_norm": 0.9197341203689575,
      "learning_rate": 0.00014896,
      "loss": 0.1685,
      "step": 640
    },
    {
      "epoch": 1.04,
      "grad_norm": 0.37014976143836975,
      "learning_rate": 0.00014816000000000002,
      "loss": 0.1729,
      "step": 650
    },
    {
      "epoch": 1.06,
      "grad_norm": 0.8919755220413208,
      "learning_rate": 0.00014736,
      "loss": 0.1993,
      "step": 660
    },
    {
      "epoch": 1.07,
      "grad_norm": 0.7291600704193115,
      "learning_rate": 0.00014656,
      "loss": 0.1893,
      "step": 670
    },
    {
      "epoch": 1.09,
      "grad_norm": 2.347400665283203,
      "learning_rate": 0.00014576000000000001,
      "loss": 0.1799,
      "step": 680
    },
    {
      "epoch": 1.1,
      "grad_norm": 0.3188568949699402,
      "learning_rate": 0.00014496,
      "loss": 0.167,
      "step": 690
    },
    {
      "epoch": 1.12,
      "grad_norm": 1.342278242111206,
      "learning_rate": 0.00014416,
      "loss": 0.1991,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_loss": 0.26034072041511536,
      "eval_na_accuracy": 0.942,
      "eval_ordinal_accuracy": 0.6078431372549019,
      "eval_runtime": 39.5088,
      "eval_samples_per_second": 25.311,
      "eval_steps_per_second": 3.164,
      "step": 700
    },
    {
      "epoch": 1.14,
      "grad_norm": 0.48655757308006287,
      "learning_rate": 0.00014336,
      "loss": 0.1885,
      "step": 710
    },
    {
      "epoch": 1.15,
      "grad_norm": 0.556333065032959,
      "learning_rate": 0.00014256000000000002,
      "loss": 0.1449,
      "step": 720
    },
    {
      "epoch": 1.17,
      "grad_norm": 0.4880894422531128,
      "learning_rate": 0.00014176,
      "loss": 0.1164,
      "step": 730
    },
    {
      "epoch": 1.18,
      "grad_norm": 0.599926233291626,
      "learning_rate": 0.00014096,
      "loss": 0.2113,
      "step": 740
    },
    {
      "epoch": 1.2,
      "grad_norm": 0.6070149540901184,
      "learning_rate": 0.00014016,
      "loss": 0.1534,
      "step": 750
    },
    {
      "epoch": 1.22,
      "grad_norm": 0.7789746522903442,
      "learning_rate": 0.00013936,
      "loss": 0.1655,
      "step": 760
    },
    {
      "epoch": 1.23,
      "grad_norm": 0.5523375868797302,
      "learning_rate": 0.00013856,
      "loss": 0.298,
      "step": 770
    },
    {
      "epoch": 1.25,
      "grad_norm": 2.4257819652557373,
      "learning_rate": 0.00013776,
      "loss": 0.2101,
      "step": 780
    },
    {
      "epoch": 1.26,
      "grad_norm": 0.5729731321334839,
      "learning_rate": 0.00013696,
      "loss": 0.133,
      "step": 790
    },
    {
      "epoch": 1.28,
      "grad_norm": 0.4050444960594177,
      "learning_rate": 0.00013616,
      "loss": 0.1678,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_loss": 0.2904650568962097,
      "eval_na_accuracy": 0.942,
      "eval_ordinal_accuracy": 0.6332179930795848,
      "eval_runtime": 39.7139,
      "eval_samples_per_second": 25.18,
      "eval_steps_per_second": 3.148,
      "step": 800
    },
    {
      "epoch": 1.3,
      "grad_norm": 4.782747745513916,
      "learning_rate": 0.00013536,
      "loss": 0.207,
      "step": 810
    },
    {
      "epoch": 1.31,
      "grad_norm": 2.577669143676758,
      "learning_rate": 0.00013455999999999999,
      "loss": 0.1818,
      "step": 820
    },
    {
      "epoch": 1.33,
      "grad_norm": 2.8163273334503174,
      "learning_rate": 0.00013376,
      "loss": 0.1761,
      "step": 830
    },
    {
      "epoch": 1.34,
      "grad_norm": 2.213799238204956,
      "learning_rate": 0.00013296,
      "loss": 0.2966,
      "step": 840
    },
    {
      "epoch": 1.36,
      "grad_norm": 0.8946444988250732,
      "learning_rate": 0.00013216,
      "loss": 0.1569,
      "step": 850
    },
    {
      "epoch": 1.38,
      "grad_norm": 0.6494708061218262,
      "learning_rate": 0.00013136000000000002,
      "loss": 0.1746,
      "step": 860
    },
    {
      "epoch": 1.39,
      "grad_norm": 1.0058079957962036,
      "learning_rate": 0.00013056000000000002,
      "loss": 0.1204,
      "step": 870
    },
    {
      "epoch": 1.41,
      "grad_norm": 1.1752161979675293,
      "learning_rate": 0.00012976,
      "loss": 0.2082,
      "step": 880
    },
    {
      "epoch": 1.42,
      "grad_norm": 0.5655858516693115,
      "learning_rate": 0.00012896,
      "loss": 0.1971,
      "step": 890
    },
    {
      "epoch": 1.44,
      "grad_norm": 2.5486743450164795,
      "learning_rate": 0.00012816000000000002,
      "loss": 0.2514,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_loss": 0.25656750798225403,
      "eval_na_accuracy": 0.94,
      "eval_ordinal_accuracy": 0.6089965397923875,
      "eval_runtime": 39.7194,
      "eval_samples_per_second": 25.177,
      "eval_steps_per_second": 3.147,
      "step": 900
    },
    {
      "epoch": 1.46,
      "grad_norm": 0.878511369228363,
      "learning_rate": 0.00012736,
      "loss": 0.162,
      "step": 910
    },
    {
      "epoch": 1.47,
      "grad_norm": 1.1985282897949219,
      "learning_rate": 0.00012656,
      "loss": 0.2268,
      "step": 920
    },
    {
      "epoch": 1.49,
      "grad_norm": 0.521425187587738,
      "learning_rate": 0.00012576000000000002,
      "loss": 0.1556,
      "step": 930
    },
    {
      "epoch": 1.5,
      "grad_norm": 0.9773241877555847,
      "learning_rate": 0.00012496000000000002,
      "loss": 0.1457,
      "step": 940
    },
    {
      "epoch": 1.52,
      "grad_norm": 1.6476322412490845,
      "learning_rate": 0.00012416,
      "loss": 0.1913,
      "step": 950
    },
    {
      "epoch": 1.54,
      "grad_norm": 1.7127236127853394,
      "learning_rate": 0.00012336,
      "loss": 0.1961,
      "step": 960
    },
    {
      "epoch": 1.55,
      "grad_norm": 4.41243314743042,
      "learning_rate": 0.00012256000000000002,
      "loss": 0.2061,
      "step": 970
    },
    {
      "epoch": 1.57,
      "grad_norm": 1.5907992124557495,
      "learning_rate": 0.00012176000000000001,
      "loss": 0.1299,
      "step": 980
    },
    {
      "epoch": 1.58,
      "grad_norm": 0.5711427927017212,
      "learning_rate": 0.00012096000000000001,
      "loss": 0.1755,
      "step": 990
    },
    {
      "epoch": 1.6,
      "grad_norm": 2.925363302230835,
      "learning_rate": 0.00012016,
      "loss": 0.2328,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_loss": 0.2884255647659302,
      "eval_na_accuracy": 0.94,
      "eval_ordinal_accuracy": 0.5617070357554786,
      "eval_runtime": 39.4898,
      "eval_samples_per_second": 25.323,
      "eval_steps_per_second": 3.165,
      "step": 1000
    },
    {
      "epoch": 1.62,
      "grad_norm": 1.1306260824203491,
      "learning_rate": 0.00011936000000000001,
      "loss": 0.1595,
      "step": 1010
    },
    {
      "epoch": 1.63,
      "grad_norm": 1.8953267335891724,
      "learning_rate": 0.00011856,
      "loss": 0.2489,
      "step": 1020
    },
    {
      "epoch": 1.65,
      "grad_norm": 0.7074128985404968,
      "learning_rate": 0.00011776,
      "loss": 0.2485,
      "step": 1030
    },
    {
      "epoch": 1.66,
      "grad_norm": 0.7052355408668518,
      "learning_rate": 0.00011696,
      "loss": 0.2075,
      "step": 1040
    },
    {
      "epoch": 1.68,
      "grad_norm": 0.7830259203910828,
      "learning_rate": 0.00011616,
      "loss": 0.2346,
      "step": 1050
    },
    {
      "epoch": 1.7,
      "grad_norm": 0.5882430672645569,
      "learning_rate": 0.00011536000000000001,
      "loss": 0.2136,
      "step": 1060
    },
    {
      "epoch": 1.71,
      "grad_norm": 1.0235962867736816,
      "learning_rate": 0.00011456,
      "loss": 0.1753,
      "step": 1070
    },
    {
      "epoch": 1.73,
      "grad_norm": 0.9401603937149048,
      "learning_rate": 0.00011376,
      "loss": 0.173,
      "step": 1080
    },
    {
      "epoch": 1.74,
      "grad_norm": 1.0735399723052979,
      "learning_rate": 0.00011296,
      "loss": 0.1993,
      "step": 1090
    },
    {
      "epoch": 1.76,
      "grad_norm": 0.6592912673950195,
      "learning_rate": 0.00011216,
      "loss": 0.1826,
      "step": 1100
    },
    {
      "epoch": 1.76,
      "eval_loss": 0.2869604229927063,
      "eval_na_accuracy": 0.943,
      "eval_ordinal_accuracy": 0.6043829296424452,
      "eval_runtime": 39.3069,
      "eval_samples_per_second": 25.441,
      "eval_steps_per_second": 3.18,
      "step": 1100
    },
    {
      "epoch": 1.78,
      "grad_norm": 2.1649601459503174,
      "learning_rate": 0.00011135999999999999,
      "loss": 0.221,
      "step": 1110
    },
    {
      "epoch": 1.79,
      "grad_norm": 1.7881801128387451,
      "learning_rate": 0.00011056,
      "loss": 0.1765,
      "step": 1120
    },
    {
      "epoch": 1.81,
      "grad_norm": 1.3527191877365112,
      "learning_rate": 0.00010975999999999999,
      "loss": 0.1325,
      "step": 1130
    },
    {
      "epoch": 1.82,
      "grad_norm": 0.7212499976158142,
      "learning_rate": 0.00010896,
      "loss": 0.2445,
      "step": 1140
    },
    {
      "epoch": 1.84,
      "grad_norm": 0.9492518901824951,
      "learning_rate": 0.00010816,
      "loss": 0.2704,
      "step": 1150
    },
    {
      "epoch": 1.86,
      "grad_norm": 0.4344118535518646,
      "learning_rate": 0.00010736000000000002,
      "loss": 0.1624,
      "step": 1160
    },
    {
      "epoch": 1.87,
      "grad_norm": 0.4115823209285736,
      "learning_rate": 0.00010656000000000001,
      "loss": 0.2069,
      "step": 1170
    },
    {
      "epoch": 1.89,
      "grad_norm": 0.6738015413284302,
      "learning_rate": 0.00010576000000000002,
      "loss": 0.208,
      "step": 1180
    },
    {
      "epoch": 1.9,
      "grad_norm": 0.9090007543563843,
      "learning_rate": 0.00010496000000000001,
      "loss": 0.1793,
      "step": 1190
    },
    {
      "epoch": 1.92,
      "grad_norm": 1.1480025053024292,
      "learning_rate": 0.00010416000000000002,
      "loss": 0.2013,
      "step": 1200
    },
    {
      "epoch": 1.92,
      "eval_loss": 0.29365527629852295,
      "eval_na_accuracy": 0.941,
      "eval_ordinal_accuracy": 0.5905420991926182,
      "eval_runtime": 39.6842,
      "eval_samples_per_second": 25.199,
      "eval_steps_per_second": 3.15,
      "step": 1200
    },
    {
      "epoch": 1.94,
      "grad_norm": 0.62380051612854,
      "learning_rate": 0.00010336000000000001,
      "loss": 0.192,
      "step": 1210
    },
    {
      "epoch": 1.95,
      "grad_norm": 0.9949710965156555,
      "learning_rate": 0.00010256000000000001,
      "loss": 0.226,
      "step": 1220
    },
    {
      "epoch": 1.97,
      "grad_norm": 1.0634446144104004,
      "learning_rate": 0.00010176000000000002,
      "loss": 0.142,
      "step": 1230
    },
    {
      "epoch": 1.98,
      "grad_norm": 0.8875225782394409,
      "learning_rate": 0.00010096000000000001,
      "loss": 0.1729,
      "step": 1240
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.6193259358406067,
      "learning_rate": 0.00010016,
      "loss": 0.102,
      "step": 1250
    },
    {
      "epoch": 2.02,
      "grad_norm": 0.4042517840862274,
      "learning_rate": 9.936000000000001e-05,
      "loss": 0.0976,
      "step": 1260
    },
    {
      "epoch": 2.03,
      "grad_norm": 0.4051195979118347,
      "learning_rate": 9.856e-05,
      "loss": 0.1448,
      "step": 1270
    },
    {
      "epoch": 2.05,
      "grad_norm": 0.46061789989471436,
      "learning_rate": 9.776000000000001e-05,
      "loss": 0.0768,
      "step": 1280
    },
    {
      "epoch": 2.06,
      "grad_norm": 0.5934004783630371,
      "learning_rate": 9.696000000000001e-05,
      "loss": 0.1404,
      "step": 1290
    },
    {
      "epoch": 2.08,
      "grad_norm": 0.6819984316825867,
      "learning_rate": 9.616e-05,
      "loss": 0.0663,
      "step": 1300
    },
    {
      "epoch": 2.08,
      "eval_loss": 0.2954486608505249,
      "eval_na_accuracy": 0.938,
      "eval_ordinal_accuracy": 0.6251441753171857,
      "eval_runtime": 39.658,
      "eval_samples_per_second": 25.216,
      "eval_steps_per_second": 3.152,
      "step": 1300
    },
    {
      "epoch": 2.1,
      "grad_norm": 0.5849266052246094,
      "learning_rate": 9.536000000000001e-05,
      "loss": 0.1574,
      "step": 1310
    },
    {
      "epoch": 2.11,
      "grad_norm": 0.9393780827522278,
      "learning_rate": 9.456e-05,
      "loss": 0.0979,
      "step": 1320
    },
    {
      "epoch": 2.13,
      "grad_norm": 0.47529059648513794,
      "learning_rate": 9.376e-05,
      "loss": 0.1073,
      "step": 1330
    },
    {
      "epoch": 2.14,
      "grad_norm": 0.4079722464084625,
      "learning_rate": 9.296e-05,
      "loss": 0.0868,
      "step": 1340
    },
    {
      "epoch": 2.16,
      "grad_norm": 0.7292589545249939,
      "learning_rate": 9.216e-05,
      "loss": 0.1446,
      "step": 1350
    },
    {
      "epoch": 2.18,
      "grad_norm": 0.9205511212348938,
      "learning_rate": 9.136e-05,
      "loss": 0.0907,
      "step": 1360
    },
    {
      "epoch": 2.19,
      "grad_norm": 0.9218105673789978,
      "learning_rate": 9.056e-05,
      "loss": 0.1387,
      "step": 1370
    },
    {
      "epoch": 2.21,
      "grad_norm": 0.5730422139167786,
      "learning_rate": 8.976e-05,
      "loss": 0.0882,
      "step": 1380
    },
    {
      "epoch": 2.22,
      "grad_norm": 0.6922823190689087,
      "learning_rate": 8.896e-05,
      "loss": 0.0741,
      "step": 1390
    },
    {
      "epoch": 2.24,
      "grad_norm": 1.1872971057891846,
      "learning_rate": 8.816000000000001e-05,
      "loss": 0.1503,
      "step": 1400
    },
    {
      "epoch": 2.24,
      "eval_loss": 0.3187769651412964,
      "eval_na_accuracy": 0.937,
      "eval_ordinal_accuracy": 0.5986159169550173,
      "eval_runtime": 39.7243,
      "eval_samples_per_second": 25.173,
      "eval_steps_per_second": 3.147,
      "step": 1400
    },
    {
      "epoch": 2.26,
      "grad_norm": 2.3350443840026855,
      "learning_rate": 8.736e-05,
      "loss": 0.1083,
      "step": 1410
    },
    {
      "epoch": 2.27,
      "grad_norm": 0.8266046643257141,
      "learning_rate": 8.656000000000001e-05,
      "loss": 0.0684,
      "step": 1420
    },
    {
      "epoch": 2.29,
      "grad_norm": 0.41480687260627747,
      "learning_rate": 8.576e-05,
      "loss": 0.0809,
      "step": 1430
    },
    {
      "epoch": 2.3,
      "grad_norm": 0.4657377600669861,
      "learning_rate": 8.496e-05,
      "loss": 0.0718,
      "step": 1440
    },
    {
      "epoch": 2.32,
      "grad_norm": 0.5419800877571106,
      "learning_rate": 8.416000000000001e-05,
      "loss": 0.1322,
      "step": 1450
    },
    {
      "epoch": 2.34,
      "grad_norm": 1.167611837387085,
      "learning_rate": 8.336e-05,
      "loss": 0.1017,
      "step": 1460
    },
    {
      "epoch": 2.35,
      "grad_norm": 0.449034184217453,
      "learning_rate": 8.256000000000001e-05,
      "loss": 0.0636,
      "step": 1470
    },
    {
      "epoch": 2.37,
      "grad_norm": 0.6716451048851013,
      "learning_rate": 8.176e-05,
      "loss": 0.1109,
      "step": 1480
    },
    {
      "epoch": 2.38,
      "grad_norm": 4.306596755981445,
      "learning_rate": 8.096e-05,
      "loss": 0.0898,
      "step": 1490
    },
    {
      "epoch": 2.4,
      "grad_norm": 0.41288742423057556,
      "learning_rate": 8.016e-05,
      "loss": 0.0611,
      "step": 1500
    },
    {
      "epoch": 2.4,
      "eval_loss": 0.33932703733444214,
      "eval_na_accuracy": 0.945,
      "eval_ordinal_accuracy": 0.5997693194925029,
      "eval_runtime": 39.236,
      "eval_samples_per_second": 25.487,
      "eval_steps_per_second": 3.186,
      "step": 1500
    },
    {
      "epoch": 2.42,
      "grad_norm": 0.7951626181602478,
      "learning_rate": 7.936e-05,
      "loss": 0.0799,
      "step": 1510
    },
    {
      "epoch": 2.43,
      "grad_norm": 1.0197049379348755,
      "learning_rate": 7.856000000000001e-05,
      "loss": 0.0928,
      "step": 1520
    },
    {
      "epoch": 2.45,
      "grad_norm": 0.6486759185791016,
      "learning_rate": 7.776e-05,
      "loss": 0.0964,
      "step": 1530
    },
    {
      "epoch": 2.46,
      "grad_norm": 1.0220657587051392,
      "learning_rate": 7.696e-05,
      "loss": 0.1736,
      "step": 1540
    },
    {
      "epoch": 2.48,
      "grad_norm": 2.3006441593170166,
      "learning_rate": 7.616e-05,
      "loss": 0.0993,
      "step": 1550
    },
    {
      "epoch": 2.5,
      "grad_norm": 0.4701670706272125,
      "learning_rate": 7.536000000000001e-05,
      "loss": 0.0779,
      "step": 1560
    },
    {
      "epoch": 2.51,
      "grad_norm": 0.641832172870636,
      "learning_rate": 7.456e-05,
      "loss": 0.0586,
      "step": 1570
    },
    {
      "epoch": 2.53,
      "grad_norm": 0.5836305618286133,
      "learning_rate": 7.376000000000001e-05,
      "loss": 0.053,
      "step": 1580
    },
    {
      "epoch": 2.54,
      "grad_norm": 0.6500815153121948,
      "learning_rate": 7.296e-05,
      "loss": 0.0779,
      "step": 1590
    },
    {
      "epoch": 2.56,
      "grad_norm": 0.5682386755943298,
      "learning_rate": 7.216e-05,
      "loss": 0.0743,
      "step": 1600
    },
    {
      "epoch": 2.56,
      "eval_loss": 0.3182476758956909,
      "eval_na_accuracy": 0.942,
      "eval_ordinal_accuracy": 0.6482122260668973,
      "eval_runtime": 40.5507,
      "eval_samples_per_second": 24.66,
      "eval_steps_per_second": 3.083,
      "step": 1600
    },
    {
      "epoch": 2.58,
      "grad_norm": 0.5527540445327759,
      "learning_rate": 7.136000000000001e-05,
      "loss": 0.094,
      "step": 1610
    },
    {
      "epoch": 2.59,
      "grad_norm": 0.8710426092147827,
      "learning_rate": 7.056e-05,
      "loss": 0.0839,
      "step": 1620
    },
    {
      "epoch": 2.61,
      "grad_norm": 0.9312260746955872,
      "learning_rate": 6.976000000000001e-05,
      "loss": 0.155,
      "step": 1630
    },
    {
      "epoch": 2.62,
      "grad_norm": 0.48695412278175354,
      "learning_rate": 6.896e-05,
      "loss": 0.0606,
      "step": 1640
    },
    {
      "epoch": 2.64,
      "grad_norm": 0.525652289390564,
      "learning_rate": 6.816e-05,
      "loss": 0.0715,
      "step": 1650
    },
    {
      "epoch": 2.66,
      "grad_norm": 0.7670960426330566,
      "learning_rate": 6.736e-05,
      "loss": 0.1108,
      "step": 1660
    },
    {
      "epoch": 2.67,
      "grad_norm": 1.0041375160217285,
      "learning_rate": 6.656e-05,
      "loss": 0.1257,
      "step": 1670
    },
    {
      "epoch": 2.69,
      "grad_norm": 0.3819805681705475,
      "learning_rate": 6.576e-05,
      "loss": 0.1004,
      "step": 1680
    },
    {
      "epoch": 2.7,
      "grad_norm": 0.5372006893157959,
      "learning_rate": 6.496e-05,
      "loss": 0.0825,
      "step": 1690
    },
    {
      "epoch": 2.72,
      "grad_norm": 0.5835949182510376,
      "learning_rate": 6.416e-05,
      "loss": 0.0908,
      "step": 1700
    },
    {
      "epoch": 2.72,
      "eval_loss": 0.3332485854625702,
      "eval_na_accuracy": 0.942,
      "eval_ordinal_accuracy": 0.6482122260668973,
      "eval_runtime": 39.9642,
      "eval_samples_per_second": 25.022,
      "eval_steps_per_second": 3.128,
      "step": 1700
    },
    {
      "epoch": 2.74,
      "grad_norm": 0.677947461605072,
      "learning_rate": 6.336e-05,
      "loss": 0.1086,
      "step": 1710
    },
    {
      "epoch": 2.75,
      "grad_norm": 0.7373325228691101,
      "learning_rate": 6.256000000000001e-05,
      "loss": 0.0698,
      "step": 1720
    },
    {
      "epoch": 2.77,
      "grad_norm": 0.7738047242164612,
      "learning_rate": 6.176e-05,
      "loss": 0.1118,
      "step": 1730
    },
    {
      "epoch": 2.78,
      "grad_norm": 2.052891254425049,
      "learning_rate": 6.0960000000000006e-05,
      "loss": 0.1002,
      "step": 1740
    },
    {
      "epoch": 2.8,
      "grad_norm": 0.26311352849006653,
      "learning_rate": 6.016000000000001e-05,
      "loss": 0.0944,
      "step": 1750
    },
    {
      "epoch": 2.82,
      "grad_norm": 0.8190409541130066,
      "learning_rate": 5.936000000000001e-05,
      "loss": 0.0717,
      "step": 1760
    },
    {
      "epoch": 2.83,
      "grad_norm": 0.5824436545372009,
      "learning_rate": 5.856e-05,
      "loss": 0.0746,
      "step": 1770
    },
    {
      "epoch": 2.85,
      "grad_norm": 0.5489352941513062,
      "learning_rate": 5.776e-05,
      "loss": 0.1063,
      "step": 1780
    },
    {
      "epoch": 2.86,
      "grad_norm": 0.656225323677063,
      "learning_rate": 5.6960000000000004e-05,
      "loss": 0.0763,
      "step": 1790
    },
    {
      "epoch": 2.88,
      "grad_norm": 0.8495000600814819,
      "learning_rate": 5.6160000000000004e-05,
      "loss": 0.1108,
      "step": 1800
    },
    {
      "epoch": 2.88,
      "eval_loss": 0.32561835646629333,
      "eval_na_accuracy": 0.943,
      "eval_ordinal_accuracy": 0.6459054209919262,
      "eval_runtime": 39.4673,
      "eval_samples_per_second": 25.337,
      "eval_steps_per_second": 3.167,
      "step": 1800
    },
    {
      "epoch": 2.9,
      "grad_norm": 0.617258608341217,
      "learning_rate": 5.536e-05,
      "loss": 0.1203,
      "step": 1810
    },
    {
      "epoch": 2.91,
      "grad_norm": 0.4484919011592865,
      "learning_rate": 5.456e-05,
      "loss": 0.0573,
      "step": 1820
    },
    {
      "epoch": 2.93,
      "grad_norm": 0.533388614654541,
      "learning_rate": 5.376e-05,
      "loss": 0.0762,
      "step": 1830
    },
    {
      "epoch": 2.94,
      "grad_norm": 0.4078121483325958,
      "learning_rate": 5.296e-05,
      "loss": 0.0643,
      "step": 1840
    },
    {
      "epoch": 2.96,
      "grad_norm": 0.5678732395172119,
      "learning_rate": 5.2159999999999995e-05,
      "loss": 0.1126,
      "step": 1850
    },
    {
      "epoch": 2.98,
      "grad_norm": 0.6543716192245483,
      "learning_rate": 5.1359999999999996e-05,
      "loss": 0.0763,
      "step": 1860
    },
    {
      "epoch": 2.99,
      "grad_norm": 0.6005885601043701,
      "learning_rate": 5.056000000000001e-05,
      "loss": 0.0949,
      "step": 1870
    },
    {
      "epoch": 3.01,
      "grad_norm": 0.4467845559120178,
      "learning_rate": 4.976e-05,
      "loss": 0.0481,
      "step": 1880
    },
    {
      "epoch": 3.02,
      "grad_norm": 0.48746606707572937,
      "learning_rate": 4.896e-05,
      "loss": 0.0415,
      "step": 1890
    },
    {
      "epoch": 3.04,
      "grad_norm": 0.7011713981628418,
      "learning_rate": 4.816e-05,
      "loss": 0.0786,
      "step": 1900
    },
    {
      "epoch": 3.04,
      "eval_loss": 0.3222349286079407,
      "eval_na_accuracy": 0.944,
      "eval_ordinal_accuracy": 0.6539792387543253,
      "eval_runtime": 39.2271,
      "eval_samples_per_second": 25.493,
      "eval_steps_per_second": 3.187,
      "step": 1900
    },
    {
      "epoch": 3.06,
      "grad_norm": 0.5153183341026306,
      "learning_rate": 4.736000000000001e-05,
      "loss": 0.0387,
      "step": 1910
    },
    {
      "epoch": 3.07,
      "grad_norm": 0.43751129508018494,
      "learning_rate": 4.656e-05,
      "loss": 0.0364,
      "step": 1920
    },
    {
      "epoch": 3.09,
      "grad_norm": 0.7584701776504517,
      "learning_rate": 4.576e-05,
      "loss": 0.0397,
      "step": 1930
    },
    {
      "epoch": 3.1,
      "grad_norm": 0.20170661807060242,
      "learning_rate": 4.496e-05,
      "loss": 0.0288,
      "step": 1940
    },
    {
      "epoch": 3.12,
      "grad_norm": 0.2583639919757843,
      "learning_rate": 4.4160000000000004e-05,
      "loss": 0.0302,
      "step": 1950
    },
    {
      "epoch": 3.14,
      "grad_norm": 3.9720704555511475,
      "learning_rate": 4.336e-05,
      "loss": 0.0484,
      "step": 1960
    },
    {
      "epoch": 3.15,
      "grad_norm": 0.3367606997489929,
      "learning_rate": 4.256e-05,
      "loss": 0.0387,
      "step": 1970
    },
    {
      "epoch": 3.17,
      "grad_norm": 0.7610962986946106,
      "learning_rate": 4.176000000000001e-05,
      "loss": 0.0424,
      "step": 1980
    },
    {
      "epoch": 3.18,
      "grad_norm": 0.6901140213012695,
      "learning_rate": 4.096e-05,
      "loss": 0.035,
      "step": 1990
    },
    {
      "epoch": 3.2,
      "grad_norm": 0.2873363196849823,
      "learning_rate": 4.016e-05,
      "loss": 0.043,
      "step": 2000
    },
    {
      "epoch": 3.2,
      "eval_loss": 0.35012441873550415,
      "eval_na_accuracy": 0.941,
      "eval_ordinal_accuracy": 0.6482122260668973,
      "eval_runtime": 39.427,
      "eval_samples_per_second": 25.363,
      "eval_steps_per_second": 3.17,
      "step": 2000
    },
    {
      "epoch": 3.22,
      "grad_norm": 0.5277103781700134,
      "learning_rate": 3.936e-05,
      "loss": 0.0423,
      "step": 2010
    },
    {
      "epoch": 3.23,
      "grad_norm": 0.3088182508945465,
      "learning_rate": 3.8560000000000004e-05,
      "loss": 0.0305,
      "step": 2020
    },
    {
      "epoch": 3.25,
      "grad_norm": 0.3621159791946411,
      "learning_rate": 3.776e-05,
      "loss": 0.0398,
      "step": 2030
    },
    {
      "epoch": 3.26,
      "grad_norm": 0.6761226654052734,
      "learning_rate": 3.696e-05,
      "loss": 0.0606,
      "step": 2040
    },
    {
      "epoch": 3.28,
      "grad_norm": 0.9860779047012329,
      "learning_rate": 3.616e-05,
      "loss": 0.0437,
      "step": 2050
    },
    {
      "epoch": 3.3,
      "grad_norm": 0.6743874549865723,
      "learning_rate": 3.536000000000001e-05,
      "loss": 0.0328,
      "step": 2060
    },
    {
      "epoch": 3.31,
      "grad_norm": 1.5928354263305664,
      "learning_rate": 3.456e-05,
      "loss": 0.0261,
      "step": 2070
    },
    {
      "epoch": 3.33,
      "grad_norm": 0.9067389965057373,
      "learning_rate": 3.376e-05,
      "loss": 0.0268,
      "step": 2080
    },
    {
      "epoch": 3.34,
      "grad_norm": 0.5733221173286438,
      "learning_rate": 3.296e-05,
      "loss": 0.0335,
      "step": 2090
    },
    {
      "epoch": 3.36,
      "grad_norm": 0.7042862772941589,
      "learning_rate": 3.2160000000000004e-05,
      "loss": 0.0472,
      "step": 2100
    },
    {
      "epoch": 3.36,
      "eval_loss": 0.34554365277290344,
      "eval_na_accuracy": 0.943,
      "eval_ordinal_accuracy": 0.6608996539792388,
      "eval_runtime": 40.1655,
      "eval_samples_per_second": 24.897,
      "eval_steps_per_second": 3.112,
      "step": 2100
    },
    {
      "epoch": 3.38,
      "grad_norm": 0.8036783933639526,
      "learning_rate": 3.136e-05,
      "loss": 0.0402,
      "step": 2110
    },
    {
      "epoch": 3.39,
      "grad_norm": 0.4863825738430023,
      "learning_rate": 3.056e-05,
      "loss": 0.0507,
      "step": 2120
    },
    {
      "epoch": 3.41,
      "grad_norm": 0.5171158313751221,
      "learning_rate": 2.976e-05,
      "loss": 0.0333,
      "step": 2130
    },
    {
      "epoch": 3.42,
      "grad_norm": 0.21965381503105164,
      "learning_rate": 2.8960000000000004e-05,
      "loss": 0.0277,
      "step": 2140
    },
    {
      "epoch": 3.44,
      "grad_norm": 0.20841450989246368,
      "learning_rate": 2.816e-05,
      "loss": 0.0259,
      "step": 2150
    },
    {
      "epoch": 3.46,
      "grad_norm": 0.5015869736671448,
      "learning_rate": 2.7360000000000002e-05,
      "loss": 0.0316,
      "step": 2160
    },
    {
      "epoch": 3.47,
      "grad_norm": 0.7938678860664368,
      "learning_rate": 2.6560000000000003e-05,
      "loss": 0.0301,
      "step": 2170
    },
    {
      "epoch": 3.49,
      "grad_norm": 0.44840845465660095,
      "learning_rate": 2.576e-05,
      "loss": 0.0406,
      "step": 2180
    },
    {
      "epoch": 3.5,
      "grad_norm": 0.35510167479515076,
      "learning_rate": 2.496e-05,
      "loss": 0.0268,
      "step": 2190
    },
    {
      "epoch": 3.52,
      "grad_norm": 0.37328681349754333,
      "learning_rate": 2.4160000000000002e-05,
      "loss": 0.032,
      "step": 2200
    },
    {
      "epoch": 3.52,
      "eval_loss": 0.35616353154182434,
      "eval_na_accuracy": 0.94,
      "eval_ordinal_accuracy": 0.6516724336793541,
      "eval_runtime": 40.3825,
      "eval_samples_per_second": 24.763,
      "eval_steps_per_second": 3.095,
      "step": 2200
    },
    {
      "epoch": 3.54,
      "grad_norm": 0.24070143699645996,
      "learning_rate": 2.336e-05,
      "loss": 0.0262,
      "step": 2210
    },
    {
      "epoch": 3.55,
      "grad_norm": 1.0428861379623413,
      "learning_rate": 2.256e-05,
      "loss": 0.04,
      "step": 2220
    },
    {
      "epoch": 3.57,
      "grad_norm": 0.626348614692688,
      "learning_rate": 2.176e-05,
      "loss": 0.0275,
      "step": 2230
    },
    {
      "epoch": 3.58,
      "grad_norm": 0.47826460003852844,
      "learning_rate": 2.0960000000000003e-05,
      "loss": 0.0379,
      "step": 2240
    },
    {
      "epoch": 3.6,
      "grad_norm": 2.685340166091919,
      "learning_rate": 2.016e-05,
      "loss": 0.047,
      "step": 2250
    },
    {
      "epoch": 3.62,
      "grad_norm": 0.1495877057313919,
      "learning_rate": 1.936e-05,
      "loss": 0.0389,
      "step": 2260
    },
    {
      "epoch": 3.63,
      "grad_norm": 0.5789759755134583,
      "learning_rate": 1.856e-05,
      "loss": 0.0337,
      "step": 2270
    },
    {
      "epoch": 3.65,
      "grad_norm": 0.4255303144454956,
      "learning_rate": 1.7760000000000003e-05,
      "loss": 0.0282,
      "step": 2280
    },
    {
      "epoch": 3.66,
      "grad_norm": 0.7483348846435547,
      "learning_rate": 1.696e-05,
      "loss": 0.0668,
      "step": 2290
    },
    {
      "epoch": 3.68,
      "grad_norm": 0.23885439336299896,
      "learning_rate": 1.616e-05,
      "loss": 0.0434,
      "step": 2300
    },
    {
      "epoch": 3.68,
      "eval_loss": 0.34990155696868896,
      "eval_na_accuracy": 0.94,
      "eval_ordinal_accuracy": 0.6597462514417531,
      "eval_runtime": 40.407,
      "eval_samples_per_second": 24.748,
      "eval_steps_per_second": 3.094,
      "step": 2300
    },
    {
      "epoch": 3.7,
      "grad_norm": 0.3341818153858185,
      "learning_rate": 1.536e-05,
      "loss": 0.0358,
      "step": 2310
    },
    {
      "epoch": 3.71,
      "grad_norm": 0.6008884310722351,
      "learning_rate": 1.4560000000000001e-05,
      "loss": 0.0394,
      "step": 2320
    },
    {
      "epoch": 3.73,
      "grad_norm": 0.3966546654701233,
      "learning_rate": 1.376e-05,
      "loss": 0.0346,
      "step": 2330
    },
    {
      "epoch": 3.74,
      "grad_norm": 0.46933791041374207,
      "learning_rate": 1.296e-05,
      "loss": 0.0227,
      "step": 2340
    },
    {
      "epoch": 3.76,
      "grad_norm": 0.6652282476425171,
      "learning_rate": 1.216e-05,
      "loss": 0.0393,
      "step": 2350
    },
    {
      "epoch": 3.78,
      "grad_norm": 0.23938482999801636,
      "learning_rate": 1.1360000000000001e-05,
      "loss": 0.0267,
      "step": 2360
    },
    {
      "epoch": 3.79,
      "grad_norm": 0.6050881147384644,
      "learning_rate": 1.056e-05,
      "loss": 0.0287,
      "step": 2370
    },
    {
      "epoch": 3.81,
      "grad_norm": 0.22671189904212952,
      "learning_rate": 9.760000000000001e-06,
      "loss": 0.0491,
      "step": 2380
    },
    {
      "epoch": 3.82,
      "grad_norm": 0.5296955704689026,
      "learning_rate": 8.96e-06,
      "loss": 0.0266,
      "step": 2390
    },
    {
      "epoch": 3.84,
      "grad_norm": 0.5424560308456421,
      "learning_rate": 8.160000000000001e-06,
      "loss": 0.0341,
      "step": 2400
    },
    {
      "epoch": 3.84,
      "eval_loss": 0.3610887825489044,
      "eval_na_accuracy": 0.94,
      "eval_ordinal_accuracy": 0.6482122260668973,
      "eval_runtime": 40.741,
      "eval_samples_per_second": 24.545,
      "eval_steps_per_second": 3.068,
      "step": 2400
    },
    {
      "epoch": 3.86,
      "grad_norm": 0.4790421724319458,
      "learning_rate": 7.36e-06,
      "loss": 0.0319,
      "step": 2410
    },
    {
      "epoch": 3.87,
      "grad_norm": 0.4021483063697815,
      "learning_rate": 6.560000000000001e-06,
      "loss": 0.0551,
      "step": 2420
    },
    {
      "epoch": 3.89,
      "grad_norm": 0.43051794171333313,
      "learning_rate": 5.76e-06,
      "loss": 0.0281,
      "step": 2430
    },
    {
      "epoch": 3.9,
      "grad_norm": 0.23781944811344147,
      "learning_rate": 4.96e-06,
      "loss": 0.0306,
      "step": 2440
    },
    {
      "epoch": 3.92,
      "grad_norm": 0.6060004234313965,
      "learning_rate": 4.16e-06,
      "loss": 0.0326,
      "step": 2450
    },
    {
      "epoch": 3.94,
      "grad_norm": 0.5149852633476257,
      "learning_rate": 3.36e-06,
      "loss": 0.0266,
      "step": 2460
    },
    {
      "epoch": 3.95,
      "grad_norm": 0.579931914806366,
      "learning_rate": 2.56e-06,
      "loss": 0.0236,
      "step": 2470
    },
    {
      "epoch": 3.97,
      "grad_norm": 0.14379101991653442,
      "learning_rate": 1.76e-06,
      "loss": 0.0221,
      "step": 2480
    },
    {
      "epoch": 3.98,
      "grad_norm": 0.5184658765792847,
      "learning_rate": 9.6e-07,
      "loss": 0.0281,
      "step": 2490
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.5299363732337952,
      "learning_rate": 1.6e-07,
      "loss": 0.0305,
      "step": 2500
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.36354970932006836,
      "eval_na_accuracy": 0.939,
      "eval_ordinal_accuracy": 0.6608996539792388,
      "eval_runtime": 40.233,
      "eval_samples_per_second": 24.855,
      "eval_steps_per_second": 3.107,
      "step": 2500
    },
    {
      "epoch": 4.0,
      "step": 2500,
      "total_flos": 3.0997907103744e+18,
      "train_loss": 0.15650403581261635,
      "train_runtime": 3981.9696,
      "train_samples_per_second": 10.045,
      "train_steps_per_second": 0.628
    }
  ],
  "logging_steps": 10,
  "max_steps": 2500,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 100,
  "total_flos": 3.0997907103744e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}