{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9,
  "eval_steps": 100,
  "global_step": 4500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.002,
      "grad_norm": 366.0,
      "learning_rate": 1.7999999999999997e-05,
      "loss": 3.335772705078125,
      "step": 10
    },
    {
      "epoch": 0.004,
      "grad_norm": 76.0,
      "learning_rate": 3.8e-05,
      "loss": 3.325016403198242,
      "step": 20
    },
    {
      "epoch": 0.006,
      "grad_norm": 426.0,
      "learning_rate": 5.7999999999999994e-05,
      "loss": 3.355701446533203,
      "step": 30
    },
    {
      "epoch": 0.008,
      "grad_norm": 194.0,
      "learning_rate": 7.8e-05,
      "loss": 3.3106155395507812,
      "step": 40
    },
    {
      "epoch": 0.01,
      "grad_norm": 712.0,
      "learning_rate": 9.799999999999998e-05,
      "loss": 3.270039367675781,
      "step": 50
    },
    {
      "epoch": 0.012,
      "grad_norm": 164.0,
      "learning_rate": 0.00011799999999999998,
      "loss": 3.2886486053466797,
      "step": 60
    },
    {
      "epoch": 0.014,
      "grad_norm": 162.0,
      "learning_rate": 0.000138,
      "loss": 3.2564620971679688,
      "step": 70
    },
    {
      "epoch": 0.016,
      "grad_norm": 234.0,
      "learning_rate": 0.00015799999999999996,
      "loss": 3.267023468017578,
      "step": 80
    },
    {
      "epoch": 0.018,
      "grad_norm": 1536.0,
      "learning_rate": 0.000178,
      "loss": 3.231563949584961,
      "step": 90
    },
    {
      "epoch": 0.02,
      "grad_norm": 39.5,
      "learning_rate": 0.000198,
      "loss": 3.2340599060058595,
      "step": 100
    },
    {
      "epoch": 0.02,
      "eval_loss": 3.491870403289795,
      "eval_runtime": 4.5004,
      "eval_samples_per_second": 66.216,
      "eval_steps_per_second": 3.777,
      "step": 100
    },
    {
      "epoch": 0.022,
      "grad_norm": 23.0,
      "learning_rate": 0.00021799999999999999,
      "loss": 3.20191650390625,
      "step": 110
    },
    {
      "epoch": 0.024,
      "grad_norm": 32.0,
      "learning_rate": 0.00023799999999999998,
      "loss": 3.2040287017822267,
      "step": 120
    },
    {
      "epoch": 0.026,
      "grad_norm": 136.0,
      "learning_rate": 0.000258,
      "loss": 3.1589597702026366,
      "step": 130
    },
    {
      "epoch": 0.028,
      "grad_norm": 238.0,
      "learning_rate": 0.000278,
      "loss": 3.150320625305176,
      "step": 140
    },
    {
      "epoch": 0.03,
      "grad_norm": 540.0,
      "learning_rate": 0.000298,
      "loss": 3.172740364074707,
      "step": 150
    },
    {
      "epoch": 0.032,
      "grad_norm": 58.0,
      "learning_rate": 0.0003,
      "loss": 3.149807167053223,
      "step": 160
    },
    {
      "epoch": 0.034,
      "grad_norm": 196.0,
      "learning_rate": 0.0003,
      "loss": 3.169495391845703,
      "step": 170
    },
    {
      "epoch": 0.036,
      "grad_norm": 444.0,
      "learning_rate": 0.0003,
      "loss": 3.1639734268188477,
      "step": 180
    },
    {
      "epoch": 0.038,
      "grad_norm": 378.0,
      "learning_rate": 0.0003,
      "loss": 3.221640777587891,
      "step": 190
    },
    {
      "epoch": 0.04,
      "grad_norm": 2272.0,
      "learning_rate": 0.0003,
      "loss": 3.1916147232055665,
      "step": 200
    },
    {
      "epoch": 0.04,
      "eval_loss": 3.4648807048797607,
      "eval_runtime": 4.3534,
      "eval_samples_per_second": 68.452,
      "eval_steps_per_second": 3.905,
      "step": 200
    },
    {
      "epoch": 0.042,
      "grad_norm": 520.0,
      "learning_rate": 0.0003,
      "loss": 3.1814949035644533,
      "step": 210
    },
    {
      "epoch": 0.044,
      "grad_norm": 87.0,
      "learning_rate": 0.0003,
      "loss": 3.1746402740478517,
      "step": 220
    },
    {
      "epoch": 0.046,
      "grad_norm": 39.0,
      "learning_rate": 0.0003,
      "loss": 3.172597885131836,
      "step": 230
    },
    {
      "epoch": 0.048,
      "grad_norm": 50.5,
      "learning_rate": 0.0003,
      "loss": 3.1576799392700194,
      "step": 240
    },
    {
      "epoch": 0.05,
      "grad_norm": 29.0,
      "learning_rate": 0.0003,
      "loss": 3.1418006896972654,
      "step": 250
    },
    {
      "epoch": 0.052,
      "grad_norm": 50.5,
      "learning_rate": 0.0003,
      "loss": 3.1035797119140627,
      "step": 260
    },
    {
      "epoch": 0.054,
      "grad_norm": 568.0,
      "learning_rate": 0.0003,
      "loss": 3.095198059082031,
      "step": 270
    },
    {
      "epoch": 0.056,
      "grad_norm": 15.5625,
      "learning_rate": 0.0003,
      "loss": 3.079079818725586,
      "step": 280
    },
    {
      "epoch": 0.058,
      "grad_norm": 17.625,
      "learning_rate": 0.0003,
      "loss": 3.060195732116699,
      "step": 290
    },
    {
      "epoch": 0.06,
      "grad_norm": 56.25,
      "learning_rate": 0.0003,
      "loss": 3.0736625671386717,
      "step": 300
    },
    {
      "epoch": 0.06,
      "eval_loss": 3.349555015563965,
      "eval_runtime": 4.3918,
      "eval_samples_per_second": 67.854,
      "eval_steps_per_second": 3.871,
      "step": 300
    },
    {
      "epoch": 0.062,
      "grad_norm": 432.0,
      "learning_rate": 0.0003,
      "loss": 3.0508825302124025,
      "step": 310
    },
    {
      "epoch": 0.064,
      "grad_norm": 6.28125,
      "learning_rate": 0.0003,
      "loss": 3.021405792236328,
      "step": 320
    },
    {
      "epoch": 0.066,
      "grad_norm": 161.0,
      "learning_rate": 0.0003,
      "loss": 3.011042022705078,
      "step": 330
    },
    {
      "epoch": 0.068,
      "grad_norm": 29.625,
      "learning_rate": 0.0003,
      "loss": 2.9988969802856444,
      "step": 340
    },
    {
      "epoch": 0.07,
      "grad_norm": 24.625,
      "learning_rate": 0.0003,
      "loss": 2.9939882278442385,
      "step": 350
    },
    {
      "epoch": 0.072,
      "grad_norm": 61.0,
      "learning_rate": 0.0003,
      "loss": 2.979990768432617,
      "step": 360
    },
    {
      "epoch": 0.074,
      "grad_norm": 13.6875,
      "learning_rate": 0.0003,
      "loss": 2.9845876693725586,
      "step": 370
    },
    {
      "epoch": 0.076,
      "grad_norm": 280.0,
      "learning_rate": 0.0003,
      "loss": 2.975000762939453,
      "step": 380
    },
    {
      "epoch": 0.078,
      "grad_norm": 21.125,
      "learning_rate": 0.0003,
      "loss": 2.958791732788086,
      "step": 390
    },
    {
      "epoch": 0.08,
      "grad_norm": 12.8125,
      "learning_rate": 0.0003,
      "loss": 2.956746482849121,
      "step": 400
    },
    {
      "epoch": 0.08,
      "eval_loss": 3.2549893856048584,
      "eval_runtime": 4.4045,
      "eval_samples_per_second": 67.658,
      "eval_steps_per_second": 3.86,
      "step": 400
    },
    {
      "epoch": 0.082,
      "grad_norm": 314.0,
      "learning_rate": 0.0003,
      "loss": 2.9515405654907227,
      "step": 410
    },
    {
      "epoch": 0.084,
      "grad_norm": 2.125,
      "learning_rate": 0.0003,
      "loss": 2.910631561279297,
      "step": 420
    },
    {
      "epoch": 0.086,
      "grad_norm": 24.75,
      "learning_rate": 0.0003,
      "loss": 2.8589399337768553,
      "step": 430
    },
    {
      "epoch": 0.088,
      "grad_norm": 3.78125,
      "learning_rate": 0.0003,
      "loss": 2.844576835632324,
      "step": 440
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.6171875,
      "learning_rate": 0.0003,
      "loss": 2.7850717544555663,
      "step": 450
    },
    {
      "epoch": 0.092,
      "grad_norm": 6.5625,
      "learning_rate": 0.0003,
      "loss": 2.7605606079101563,
      "step": 460
    },
    {
      "epoch": 0.094,
      "grad_norm": 12.5625,
      "learning_rate": 0.0003,
      "loss": 2.747697639465332,
      "step": 470
    },
    {
      "epoch": 0.096,
      "grad_norm": 0.7734375,
      "learning_rate": 0.0003,
      "loss": 2.7024419784545897,
      "step": 480
    },
    {
      "epoch": 0.098,
      "grad_norm": 0.55078125,
      "learning_rate": 0.0003,
      "loss": 2.727059173583984,
      "step": 490
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.3046875,
      "learning_rate": 0.0003,
      "loss": 2.688610076904297,
      "step": 500
    },
    {
      "epoch": 0.1,
      "eval_loss": 3.0536108016967773,
      "eval_runtime": 4.4183,
      "eval_samples_per_second": 67.447,
      "eval_steps_per_second": 3.848,
      "step": 500
    },
    {
      "epoch": 0.102,
      "grad_norm": 0.66796875,
      "learning_rate": 0.0003,
      "loss": 2.6653621673583983,
      "step": 510
    },
    {
      "epoch": 0.104,
      "grad_norm": 2.53125,
      "learning_rate": 0.0003,
      "loss": 2.6938413619995116,
      "step": 520
    },
    {
      "epoch": 0.106,
      "grad_norm": 1.578125,
      "learning_rate": 0.0003,
      "loss": 2.6589494705200196,
      "step": 530
    },
    {
      "epoch": 0.108,
      "grad_norm": 1.453125,
      "learning_rate": 0.0003,
      "loss": 2.6775186538696287,
      "step": 540
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.1806640625,
      "learning_rate": 0.0003,
      "loss": 2.659619140625,
      "step": 550
    },
    {
      "epoch": 0.112,
      "grad_norm": 0.212890625,
      "learning_rate": 0.0003,
      "loss": 2.644103240966797,
      "step": 560
    },
    {
      "epoch": 0.114,
      "grad_norm": 0.73046875,
      "learning_rate": 0.0003,
      "loss": 2.65213565826416,
      "step": 570
    },
    {
      "epoch": 0.116,
      "grad_norm": 1.1796875,
      "learning_rate": 0.0003,
      "loss": 2.6230279922485353,
      "step": 580
    },
    {
      "epoch": 0.118,
      "grad_norm": 0.166015625,
      "learning_rate": 0.0003,
      "loss": 2.6443111419677736,
      "step": 590
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.8203125,
      "learning_rate": 0.0003,
      "loss": 2.6569570541381835,
      "step": 600
    },
    {
      "epoch": 0.12,
      "eval_loss": 3.0349245071411133,
      "eval_runtime": 4.4206,
      "eval_samples_per_second": 67.412,
      "eval_steps_per_second": 3.846,
      "step": 600
    },
    {
      "epoch": 0.122,
      "grad_norm": 0.1181640625,
      "learning_rate": 0.0003,
      "loss": 2.6595781326293944,
      "step": 610
    },
    {
      "epoch": 0.124,
      "grad_norm": 0.232421875,
      "learning_rate": 0.0003,
      "loss": 2.610936737060547,
      "step": 620
    },
    {
      "epoch": 0.126,
      "grad_norm": 0.1435546875,
      "learning_rate": 0.0003,
      "loss": 2.6387302398681642,
      "step": 630
    },
    {
      "epoch": 0.128,
      "grad_norm": 0.14453125,
      "learning_rate": 0.0003,
      "loss": 2.62979793548584,
      "step": 640
    },
    {
      "epoch": 0.13,
      "grad_norm": 2.34375,
      "learning_rate": 0.0003,
      "loss": 2.649636650085449,
      "step": 650
    },
    {
      "epoch": 0.132,
      "grad_norm": 0.2041015625,
      "learning_rate": 0.0003,
      "loss": 2.622520637512207,
      "step": 660
    },
    {
      "epoch": 0.134,
      "grad_norm": 0.97265625,
      "learning_rate": 0.0003,
      "loss": 2.646112823486328,
      "step": 670
    },
    {
      "epoch": 0.136,
      "grad_norm": 5.8125,
      "learning_rate": 0.0003,
      "loss": 2.6434356689453127,
      "step": 680
    },
    {
      "epoch": 0.138,
      "grad_norm": 0.10693359375,
      "learning_rate": 0.0003,
      "loss": 2.623065376281738,
      "step": 690
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.90625,
      "learning_rate": 0.0003,
      "loss": 2.6090648651123045,
      "step": 700
    },
    {
      "epoch": 0.14,
      "eval_loss": 3.0306036472320557,
      "eval_runtime": 4.4176,
      "eval_samples_per_second": 67.457,
      "eval_steps_per_second": 3.848,
      "step": 700
    },
    {
      "epoch": 0.142,
      "grad_norm": 0.6328125,
      "learning_rate": 0.0003,
      "loss": 2.6384559631347657,
      "step": 710
    },
    {
      "epoch": 0.144,
      "grad_norm": 0.640625,
      "learning_rate": 0.0003,
      "loss": 2.6472700119018553,
      "step": 720
    },
    {
      "epoch": 0.146,
      "grad_norm": 0.07177734375,
      "learning_rate": 0.0003,
      "loss": 2.6483619689941404,
      "step": 730
    },
    {
      "epoch": 0.148,
      "grad_norm": 0.322265625,
      "learning_rate": 0.0003,
      "loss": 2.626618766784668,
      "step": 740
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.099609375,
      "learning_rate": 0.0003,
      "loss": 2.6325525283813476,
      "step": 750
    },
    {
      "epoch": 0.152,
      "grad_norm": 0.1611328125,
      "learning_rate": 0.0003,
      "loss": 2.6193023681640626,
      "step": 760
    },
    {
      "epoch": 0.154,
      "grad_norm": 0.072265625,
      "learning_rate": 0.0003,
      "loss": 2.618332862854004,
      "step": 770
    },
    {
      "epoch": 0.156,
      "grad_norm": 0.87109375,
      "learning_rate": 0.0003,
      "loss": 2.610393524169922,
      "step": 780
    },
    {
      "epoch": 0.158,
      "grad_norm": 0.09716796875,
      "learning_rate": 0.0003,
      "loss": 2.6242008209228516,
      "step": 790
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.072265625,
      "learning_rate": 0.0003,
      "loss": 2.638929748535156,
      "step": 800
    },
    {
      "epoch": 0.16,
      "eval_loss": 3.0286974906921387,
      "eval_runtime": 4.4633,
      "eval_samples_per_second": 66.766,
      "eval_steps_per_second": 3.809,
      "step": 800
    },
    {
      "epoch": 0.162,
      "grad_norm": 0.29296875,
      "learning_rate": 0.0003,
      "loss": 2.638911247253418,
      "step": 810
    },
    {
      "epoch": 0.164,
      "grad_norm": 0.62890625,
      "learning_rate": 0.0003,
      "loss": 2.6282770156860353,
      "step": 820
    },
    {
      "epoch": 0.166,
      "grad_norm": 0.0859375,
      "learning_rate": 0.0003,
      "loss": 2.668540382385254,
      "step": 830
    },
    {
      "epoch": 0.168,
      "grad_norm": 0.08349609375,
      "learning_rate": 0.0003,
      "loss": 2.6025644302368165,
      "step": 840
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.251953125,
      "learning_rate": 0.0003,
      "loss": 2.614345932006836,
      "step": 850
    },
    {
      "epoch": 0.172,
      "grad_norm": 1.2421875,
      "learning_rate": 0.0003,
      "loss": 2.6097484588623048,
      "step": 860
    },
    {
      "epoch": 0.174,
      "grad_norm": 0.10107421875,
      "learning_rate": 0.0003,
      "loss": 2.5976139068603517,
      "step": 870
    },
    {
      "epoch": 0.176,
      "grad_norm": 0.0634765625,
      "learning_rate": 0.0003,
      "loss": 2.609291648864746,
      "step": 880
    },
    {
      "epoch": 0.178,
      "grad_norm": 0.060302734375,
      "learning_rate": 0.0003,
      "loss": 2.6136356353759767,
      "step": 890
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.306640625,
      "learning_rate": 0.0003,
      "loss": 2.6354192733764648,
      "step": 900
    },
    {
      "epoch": 0.18,
      "eval_loss": 3.028738021850586,
      "eval_runtime": 4.4125,
      "eval_samples_per_second": 67.536,
      "eval_steps_per_second": 3.853,
      "step": 900
    },
    {
      "epoch": 0.182,
      "grad_norm": 0.65234375,
      "learning_rate": 0.0003,
      "loss": 2.6325847625732424,
      "step": 910
    },
    {
      "epoch": 0.184,
      "grad_norm": 0.4921875,
      "learning_rate": 0.0003,
      "loss": 2.6379472732543947,
      "step": 920
    },
    {
      "epoch": 0.186,
      "grad_norm": 0.216796875,
      "learning_rate": 0.0003,
      "loss": 2.647774505615234,
      "step": 930
    },
    {
      "epoch": 0.188,
      "grad_norm": 0.083984375,
      "learning_rate": 0.0003,
      "loss": 2.628580665588379,
      "step": 940
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.09033203125,
      "learning_rate": 0.0003,
      "loss": 2.63638973236084,
      "step": 950
    },
    {
      "epoch": 0.192,
      "grad_norm": 0.08056640625,
      "learning_rate": 0.0003,
      "loss": 2.628683853149414,
      "step": 960
    },
    {
      "epoch": 0.194,
      "grad_norm": 0.06787109375,
      "learning_rate": 0.0003,
      "loss": 2.628560256958008,
      "step": 970
    },
    {
      "epoch": 0.196,
      "grad_norm": 0.1328125,
      "learning_rate": 0.0003,
      "loss": 2.627498435974121,
      "step": 980
    },
    {
      "epoch": 0.198,
      "grad_norm": 0.1796875,
      "learning_rate": 0.0003,
      "loss": 2.613533782958984,
      "step": 990
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.265625,
      "learning_rate": 0.0003,
      "loss": 2.615587615966797,
      "step": 1000
    },
    {
      "epoch": 0.2,
      "eval_loss": 3.0314836502075195,
      "eval_runtime": 4.4355,
      "eval_samples_per_second": 67.186,
      "eval_steps_per_second": 3.833,
      "step": 1000
    },
    {
      "epoch": 0.202,
      "grad_norm": 0.0859375,
      "learning_rate": 0.0003,
      "loss": 2.621230697631836,
      "step": 1010
    },
    {
      "epoch": 0.204,
      "grad_norm": 0.056640625,
      "learning_rate": 0.0003,
      "loss": 2.6316347122192383,
      "step": 1020
    },
    {
      "epoch": 0.206,
      "grad_norm": 0.2431640625,
      "learning_rate": 0.0003,
      "loss": 2.6155324935913087,
      "step": 1030
    },
    {
      "epoch": 0.208,
      "grad_norm": 0.06298828125,
      "learning_rate": 0.0003,
      "loss": 2.634638214111328,
      "step": 1040
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.1064453125,
      "learning_rate": 0.0003,
      "loss": 2.6131080627441405,
      "step": 1050
    },
    {
      "epoch": 0.212,
      "grad_norm": 0.06689453125,
      "learning_rate": 0.0003,
      "loss": 2.616171646118164,
      "step": 1060
    },
    {
      "epoch": 0.214,
      "grad_norm": 0.08984375,
      "learning_rate": 0.0003,
      "loss": 2.603512191772461,
      "step": 1070
    },
    {
      "epoch": 0.216,
      "grad_norm": 0.0625,
      "learning_rate": 0.0003,
      "loss": 2.6277271270751954,
      "step": 1080
    },
    {
      "epoch": 0.218,
      "grad_norm": 0.0673828125,
      "learning_rate": 0.0003,
      "loss": 2.615810775756836,
      "step": 1090
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.060791015625,
      "learning_rate": 0.0003,
      "loss": 2.6153131484985352,
      "step": 1100
    },
    {
      "epoch": 0.22,
      "eval_loss": 3.0302107334136963,
      "eval_runtime": 4.4032,
      "eval_samples_per_second": 67.678,
      "eval_steps_per_second": 3.861,
      "step": 1100
    },
    {
      "epoch": 0.222,
      "grad_norm": 0.064453125,
      "learning_rate": 0.0003,
      "loss": 2.623016357421875,
      "step": 1110
    },
    {
      "epoch": 0.224,
      "grad_norm": 0.0712890625,
      "learning_rate": 0.0003,
      "loss": 2.6170467376708983,
      "step": 1120
    },
    {
      "epoch": 0.226,
      "grad_norm": 0.59375,
      "learning_rate": 0.0003,
      "loss": 2.620796775817871,
      "step": 1130
    },
    {
      "epoch": 0.228,
      "grad_norm": 0.30078125,
      "learning_rate": 0.0003,
      "loss": 2.6082935333251953,
      "step": 1140
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.80078125,
      "learning_rate": 0.0003,
      "loss": 2.6210090637207033,
      "step": 1150
    },
    {
      "epoch": 0.232,
      "grad_norm": 0.0634765625,
      "learning_rate": 0.0003,
      "loss": 2.636818695068359,
      "step": 1160
    },
    {
      "epoch": 0.234,
      "grad_norm": 0.07080078125,
      "learning_rate": 0.0003,
      "loss": 2.5969371795654297,
      "step": 1170
    },
    {
      "epoch": 0.236,
      "grad_norm": 0.107421875,
      "learning_rate": 0.0003,
      "loss": 2.6085494995117187,
      "step": 1180
    },
    {
      "epoch": 0.238,
      "grad_norm": 0.1826171875,
      "learning_rate": 0.0003,
      "loss": 2.6118932723999024,
      "step": 1190
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.0625,
      "learning_rate": 0.0003,
      "loss": 2.6277448654174806,
      "step": 1200
    },
    {
      "epoch": 0.24,
      "eval_loss": 3.0262646675109863,
      "eval_runtime": 4.3768,
      "eval_samples_per_second": 68.087,
      "eval_steps_per_second": 3.884,
      "step": 1200
    },
    {
      "epoch": 0.242,
      "grad_norm": 0.08837890625,
      "learning_rate": 0.0003,
      "loss": 2.6076292037963866,
      "step": 1210
    },
    {
      "epoch": 0.244,
      "grad_norm": 0.0869140625,
      "learning_rate": 0.0003,
      "loss": 2.6239038467407227,
      "step": 1220
    },
    {
      "epoch": 0.246,
      "grad_norm": 0.06201171875,
      "learning_rate": 0.0003,
      "loss": 2.598759651184082,
      "step": 1230
    },
    {
      "epoch": 0.248,
      "grad_norm": 0.054443359375,
      "learning_rate": 0.0003,
      "loss": 2.623333549499512,
      "step": 1240
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.056884765625,
      "learning_rate": 0.0003,
      "loss": 2.595660400390625,
      "step": 1250
    },
    {
      "epoch": 0.252,
      "grad_norm": 0.07470703125,
      "learning_rate": 0.0003,
      "loss": 2.6085866928100585,
      "step": 1260
    },
    {
      "epoch": 0.254,
      "grad_norm": 0.06201171875,
      "learning_rate": 0.0003,
      "loss": 2.6313194274902343,
      "step": 1270
    },
    {
      "epoch": 0.256,
      "grad_norm": 0.05126953125,
      "learning_rate": 0.0003,
      "loss": 2.624466323852539,
      "step": 1280
    },
    {
      "epoch": 0.258,
      "grad_norm": 0.0693359375,
      "learning_rate": 0.0003,
      "loss": 2.61032657623291,
      "step": 1290
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.95703125,
      "learning_rate": 0.0003,
      "loss": 2.611441230773926,
      "step": 1300
    },
    {
      "epoch": 0.26,
      "eval_loss": 3.0270800590515137,
      "eval_runtime": 4.4364,
      "eval_samples_per_second": 67.171,
      "eval_steps_per_second": 3.832,
      "step": 1300
    },
    {
      "epoch": 0.262,
      "grad_norm": 0.416015625,
      "learning_rate": 0.0003,
      "loss": 2.622879981994629,
      "step": 1310
    },
    {
      "epoch": 0.264,
      "grad_norm": 2.015625,
      "learning_rate": 0.0003,
      "loss": 2.6255945205688476,
      "step": 1320
    },
    {
      "epoch": 0.266,
      "grad_norm": 0.419921875,
      "learning_rate": 0.0003,
      "loss": 2.619590950012207,
      "step": 1330
    },
    {
      "epoch": 0.268,
      "grad_norm": 0.057861328125,
      "learning_rate": 0.0003,
      "loss": 2.618213081359863,
      "step": 1340
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.054443359375,
      "learning_rate": 0.0003,
      "loss": 2.6290687561035155,
      "step": 1350
    },
    {
      "epoch": 0.272,
      "grad_norm": 0.2080078125,
      "learning_rate": 0.0003,
      "loss": 2.5912906646728517,
      "step": 1360
    },
    {
      "epoch": 0.274,
      "grad_norm": 0.06787109375,
      "learning_rate": 0.0003,
      "loss": 2.608586311340332,
      "step": 1370
    },
    {
      "epoch": 0.276,
      "grad_norm": 0.2734375,
      "learning_rate": 0.0003,
      "loss": 2.6140621185302733,
      "step": 1380
    },
    {
      "epoch": 0.278,
      "grad_norm": 0.060546875,
      "learning_rate": 0.0003,
      "loss": 2.6242074966430664,
      "step": 1390
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.056640625,
      "learning_rate": 0.0003,
      "loss": 2.6025035858154295,
      "step": 1400
    },
    {
      "epoch": 0.28,
      "eval_loss": 3.0289111137390137,
      "eval_runtime": 4.3992,
      "eval_samples_per_second": 67.739,
      "eval_steps_per_second": 3.864,
      "step": 1400
    },
    {
      "epoch": 0.282,
      "grad_norm": 0.09228515625,
      "learning_rate": 0.0003,
      "loss": 2.6210390090942384,
      "step": 1410
    },
    {
      "epoch": 0.284,
      "grad_norm": 0.087890625,
      "learning_rate": 0.0003,
      "loss": 2.6299697875976564,
      "step": 1420
    },
    {
      "epoch": 0.286,
      "grad_norm": 0.060546875,
      "learning_rate": 0.0003,
      "loss": 2.60406494140625,
      "step": 1430
    },
    {
      "epoch": 0.288,
      "grad_norm": 0.060546875,
      "learning_rate": 0.0003,
      "loss": 2.599949264526367,
      "step": 1440
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.212890625,
      "learning_rate": 0.0003,
      "loss": 2.5918996810913084,
      "step": 1450
    },
    {
      "epoch": 0.292,
      "grad_norm": 0.51953125,
      "learning_rate": 0.0003,
      "loss": 2.5961977005004884,
      "step": 1460
    },
    {
      "epoch": 0.294,
      "grad_norm": 0.0634765625,
      "learning_rate": 0.0003,
      "loss": 2.5982738494873048,
      "step": 1470
    },
    {
      "epoch": 0.296,
      "grad_norm": 0.0771484375,
      "learning_rate": 0.0003,
      "loss": 2.6246980667114257,
      "step": 1480
    },
    {
      "epoch": 0.298,
      "grad_norm": 0.515625,
      "learning_rate": 0.0003,
      "loss": 2.583383560180664,
      "step": 1490
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.068359375,
      "learning_rate": 0.0003,
      "loss": 2.6035650253295897,
      "step": 1500
    },
    {
      "epoch": 0.3,
      "eval_loss": 3.0258114337921143,
      "eval_runtime": 4.3825,
      "eval_samples_per_second": 67.998,
      "eval_steps_per_second": 3.879,
      "step": 1500
    },
    {
      "epoch": 0.302,
      "grad_norm": 0.058349609375,
      "learning_rate": 0.0003,
      "loss": 2.601143646240234,
      "step": 1510
    },
    {
      "epoch": 0.304,
      "grad_norm": 0.177734375,
      "learning_rate": 0.0003,
      "loss": 2.6228967666625977,
      "step": 1520
    },
    {
      "epoch": 0.306,
      "grad_norm": 0.12158203125,
      "learning_rate": 0.0003,
      "loss": 2.611318588256836,
      "step": 1530
    },
    {
      "epoch": 0.308,
      "grad_norm": 0.06201171875,
      "learning_rate": 0.0003,
      "loss": 2.61364860534668,
      "step": 1540
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.05419921875,
      "learning_rate": 0.0003,
      "loss": 2.6139429092407225,
      "step": 1550
    },
    {
      "epoch": 0.312,
      "grad_norm": 0.08837890625,
      "learning_rate": 0.0003,
      "loss": 2.6160503387451173,
      "step": 1560
    },
    {
      "epoch": 0.314,
      "grad_norm": 0.0673828125,
      "learning_rate": 0.0003,
      "loss": 2.615163230895996,
      "step": 1570
    },
    {
      "epoch": 0.316,
      "grad_norm": 0.055419921875,
      "learning_rate": 0.0003,
      "loss": 2.592479705810547,
      "step": 1580
    },
    {
      "epoch": 0.318,
      "grad_norm": 0.41796875,
      "learning_rate": 0.0003,
      "loss": 2.587255668640137,
      "step": 1590
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.11181640625,
      "learning_rate": 0.0003,
      "loss": 2.6103662490844726,
      "step": 1600
    },
    {
      "epoch": 0.32,
      "eval_loss": 3.0303030014038086,
      "eval_runtime": 4.4554,
      "eval_samples_per_second": 66.885,
      "eval_steps_per_second": 3.816,
      "step": 1600
    },
    {
      "epoch": 0.322,
      "grad_norm": 0.056884765625,
      "learning_rate": 0.0003,
      "loss": 2.602743911743164,
      "step": 1610
    },
    {
      "epoch": 0.324,
      "grad_norm": 0.08642578125,
      "learning_rate": 0.0003,
      "loss": 2.604319953918457,
      "step": 1620
    },
    {
      "epoch": 0.326,
      "grad_norm": 0.5859375,
      "learning_rate": 0.0003,
      "loss": 2.602446174621582,
      "step": 1630
    },
    {
      "epoch": 0.328,
      "grad_norm": 0.056396484375,
      "learning_rate": 0.0003,
      "loss": 2.6054025650024415,
      "step": 1640
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.0791015625,
      "learning_rate": 0.0003,
      "loss": 2.597785758972168,
      "step": 1650
    },
    {
      "epoch": 0.332,
      "grad_norm": 0.1259765625,
      "learning_rate": 0.0003,
      "loss": 2.5736499786376954,
      "step": 1660
    },
    {
      "epoch": 0.334,
      "grad_norm": 0.07470703125,
      "learning_rate": 0.0003,
      "loss": 2.5932741165161133,
      "step": 1670
    },
    {
      "epoch": 0.336,
      "grad_norm": 0.058837890625,
      "learning_rate": 0.0003,
      "loss": 2.6116256713867188,
      "step": 1680
    },
    {
      "epoch": 0.338,
      "grad_norm": 0.057861328125,
      "learning_rate": 0.0003,
      "loss": 2.625630187988281,
      "step": 1690
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.06640625,
      "learning_rate": 0.0003,
      "loss": 2.596027374267578,
      "step": 1700
    },
    {
      "epoch": 0.34,
      "eval_loss": 3.028986692428589,
      "eval_runtime": 4.4483,
      "eval_samples_per_second": 66.992,
      "eval_steps_per_second": 3.822,
      "step": 1700
    },
    {
      "epoch": 0.342,
      "grad_norm": 0.25,
      "learning_rate": 0.0003,
      "loss": 2.6015995025634764,
      "step": 1710
    },
    {
      "epoch": 0.344,
      "grad_norm": 0.057373046875,
      "learning_rate": 0.0003,
      "loss": 2.598787307739258,
      "step": 1720
    },
    {
      "epoch": 0.346,
      "grad_norm": 0.08544921875,
      "learning_rate": 0.0003,
      "loss": 2.6199289321899415,
      "step": 1730
    },
    {
      "epoch": 0.348,
      "grad_norm": 0.060302734375,
      "learning_rate": 0.0003,
      "loss": 2.602091598510742,
      "step": 1740
    },
    {
      "epoch": 0.35,
      "grad_norm": 4.875,
      "learning_rate": 0.0003,
      "loss": 2.5984159469604493,
      "step": 1750
    },
    {
      "epoch": 0.352,
      "grad_norm": 0.05224609375,
      "learning_rate": 0.0003,
      "loss": 2.6147043228149416,
      "step": 1760
    },
    {
      "epoch": 0.354,
      "grad_norm": 0.1982421875,
      "learning_rate": 0.0003,
      "loss": 2.6081768035888673,
      "step": 1770
    },
    {
      "epoch": 0.356,
      "grad_norm": 0.1328125,
      "learning_rate": 0.0003,
      "loss": 2.6030120849609375,
      "step": 1780
    },
    {
      "epoch": 0.358,
      "grad_norm": 0.08642578125,
      "learning_rate": 0.0003,
      "loss": 2.6104217529296876,
      "step": 1790
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.051025390625,
      "learning_rate": 0.0003,
      "loss": 2.5706939697265625,
      "step": 1800
    },
    {
      "epoch": 0.36,
      "eval_loss": 3.0339813232421875,
      "eval_runtime": 4.452,
      "eval_samples_per_second": 66.937,
      "eval_steps_per_second": 3.819,
      "step": 1800
    },
    {
      "epoch": 0.362,
      "grad_norm": 0.1376953125,
      "learning_rate": 0.0003,
      "loss": 2.576882743835449,
      "step": 1810
    },
    {
      "epoch": 0.364,
      "grad_norm": 0.1162109375,
      "learning_rate": 0.0003,
      "loss": 2.601199913024902,
      "step": 1820
    },
    {
      "epoch": 0.366,
      "grad_norm": 0.062255859375,
      "learning_rate": 0.0003,
      "loss": 2.570973777770996,
      "step": 1830
    },
    {
      "epoch": 0.368,
      "grad_norm": 0.05859375,
      "learning_rate": 0.0003,
      "loss": 2.618685722351074,
      "step": 1840
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.05859375,
      "learning_rate": 0.0003,
      "loss": 2.613678550720215,
      "step": 1850
    },
    {
      "epoch": 0.372,
      "grad_norm": 0.06640625,
      "learning_rate": 0.0003,
      "loss": 2.595960235595703,
      "step": 1860
    },
    {
      "epoch": 0.374,
      "grad_norm": 0.0595703125,
      "learning_rate": 0.0003,
      "loss": 2.6117223739624023,
      "step": 1870
    },
    {
      "epoch": 0.376,
      "grad_norm": 0.09423828125,
      "learning_rate": 0.0003,
      "loss": 2.5977685928344725,
      "step": 1880
    },
    {
      "epoch": 0.378,
      "grad_norm": 0.060791015625,
      "learning_rate": 0.0003,
      "loss": 2.619519805908203,
      "step": 1890
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.12060546875,
      "learning_rate": 0.0003,
      "loss": 2.606584930419922,
      "step": 1900
    },
    {
      "epoch": 0.38,
      "eval_loss": 3.0274596214294434,
      "eval_runtime": 4.4258,
      "eval_samples_per_second": 67.332,
      "eval_steps_per_second": 3.841,
      "step": 1900
    },
    {
      "epoch": 0.382,
      "grad_norm": 0.052734375,
      "learning_rate": 0.0003,
      "loss": 2.598057174682617,
      "step": 1910
    },
    {
      "epoch": 0.384,
      "grad_norm": 0.09228515625,
      "learning_rate": 0.0003,
      "loss": 2.5836353302001953,
      "step": 1920
    },
    {
      "epoch": 0.386,
      "grad_norm": 0.68359375,
      "learning_rate": 0.0003,
      "loss": 2.5969900131225585,
      "step": 1930
    },
    {
      "epoch": 0.388,
      "grad_norm": 0.06982421875,
      "learning_rate": 0.0003,
      "loss": 2.610474395751953,
      "step": 1940
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.16015625,
      "learning_rate": 0.0003,
      "loss": 2.589932441711426,
      "step": 1950
    },
    {
      "epoch": 0.392,
      "grad_norm": 0.05908203125,
      "learning_rate": 0.0003,
      "loss": 2.616892433166504,
      "step": 1960
    },
    {
      "epoch": 0.394,
      "grad_norm": 0.054931640625,
      "learning_rate": 0.0003,
      "loss": 2.5918600082397463,
      "step": 1970
    },
    {
      "epoch": 0.396,
      "grad_norm": 0.1064453125,
      "learning_rate": 0.0003,
      "loss": 2.6085411071777345,
      "step": 1980
    },
    {
      "epoch": 0.398,
      "grad_norm": 0.0576171875,
      "learning_rate": 0.0003,
      "loss": 2.598824691772461,
      "step": 1990
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.07861328125,
      "learning_rate": 0.0003,
      "loss": 2.5938999176025392,
      "step": 2000
    },
    {
      "epoch": 0.4,
      "eval_loss": 3.030641794204712,
      "eval_runtime": 4.4468,
      "eval_samples_per_second": 67.014,
      "eval_steps_per_second": 3.823,
      "step": 2000
    },
    {
      "epoch": 0.402,
      "grad_norm": 0.059326171875,
      "learning_rate": 0.0003,
      "loss": 2.608153533935547,
      "step": 2010
    },
    {
      "epoch": 0.404,
      "grad_norm": 0.20703125,
      "learning_rate": 0.0003,
      "loss": 2.6085268020629884,
      "step": 2020
    },
    {
      "epoch": 0.406,
      "grad_norm": 0.9296875,
      "learning_rate": 0.0003,
      "loss": 2.605429267883301,
      "step": 2030
    },
    {
      "epoch": 0.408,
      "grad_norm": 0.08056640625,
      "learning_rate": 0.0003,
      "loss": 2.6161428451538087,
      "step": 2040
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.0576171875,
      "learning_rate": 0.0003,
      "loss": 2.587337112426758,
      "step": 2050
    },
    {
      "epoch": 0.412,
      "grad_norm": 1.5546875,
      "learning_rate": 0.0003,
      "loss": 2.613199996948242,
      "step": 2060
    },
    {
      "epoch": 0.414,
      "grad_norm": 0.4375,
      "learning_rate": 0.0003,
      "loss": 2.595570945739746,
      "step": 2070
    },
    {
      "epoch": 0.416,
      "grad_norm": 0.059326171875,
      "learning_rate": 0.0003,
      "loss": 2.614238166809082,
      "step": 2080
    },
    {
      "epoch": 0.418,
      "grad_norm": 0.050537109375,
      "learning_rate": 0.0003,
      "loss": 2.5990161895751953,
      "step": 2090
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.05126953125,
      "learning_rate": 0.0003,
      "loss": 2.6036991119384765,
      "step": 2100
    },
    {
      "epoch": 0.42,
      "eval_loss": 3.0276641845703125,
      "eval_runtime": 4.4443,
      "eval_samples_per_second": 67.052,
      "eval_steps_per_second": 3.825,
      "step": 2100
    },
    {
      "epoch": 0.422,
      "grad_norm": 0.05126953125,
      "learning_rate": 0.0003,
      "loss": 2.5789426803588866,
      "step": 2110
    },
    {
      "epoch": 0.424,
      "grad_norm": 0.083984375,
      "learning_rate": 0.0003,
      "loss": 2.592574119567871,
      "step": 2120
    },
    {
      "epoch": 0.426,
      "grad_norm": 0.049072265625,
      "learning_rate": 0.0003,
      "loss": 2.592014122009277,
      "step": 2130
    },
    {
      "epoch": 0.428,
      "grad_norm": 0.0859375,
      "learning_rate": 0.0003,
      "loss": 2.5774700164794924,
      "step": 2140
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.05615234375,
      "learning_rate": 0.0003,
      "loss": 2.613265037536621,
      "step": 2150
    },
    {
      "epoch": 0.432,
      "grad_norm": 0.068359375,
      "learning_rate": 0.0003,
      "loss": 2.5970441818237306,
      "step": 2160
    },
    {
      "epoch": 0.434,
      "grad_norm": 0.08447265625,
      "learning_rate": 0.0003,
      "loss": 2.6015340805053713,
      "step": 2170
    },
    {
      "epoch": 0.436,
      "grad_norm": 0.640625,
      "learning_rate": 0.0003,
      "loss": 2.58035945892334,
      "step": 2180
    },
    {
      "epoch": 0.438,
      "grad_norm": 0.06640625,
      "learning_rate": 0.0003,
      "loss": 2.609104537963867,
      "step": 2190
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.057861328125,
      "learning_rate": 0.0003,
      "loss": 2.5889204025268553,
      "step": 2200
    },
    {
      "epoch": 0.44,
      "eval_loss": 3.0296566486358643,
      "eval_runtime": 4.4076,
      "eval_samples_per_second": 67.61,
      "eval_steps_per_second": 3.857,
      "step": 2200
    },
    {
      "epoch": 0.442,
      "grad_norm": 0.054931640625,
      "learning_rate": 0.0003,
      "loss": 2.582586097717285,
      "step": 2210
    },
    {
      "epoch": 0.444,
      "grad_norm": 0.06787109375,
      "learning_rate": 0.0003,
      "loss": 2.606948471069336,
      "step": 2220
    },
    {
      "epoch": 0.446,
      "grad_norm": 0.7421875,
      "learning_rate": 0.0003,
      "loss": 2.5844865798950196,
      "step": 2230
    },
    {
      "epoch": 0.448,
      "grad_norm": 0.1025390625,
      "learning_rate": 0.0003,
      "loss": 2.5672168731689453,
      "step": 2240
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.05810546875,
      "learning_rate": 0.0003,
      "loss": 2.588663673400879,
      "step": 2250
    },
    {
      "epoch": 0.452,
      "grad_norm": 0.05224609375,
      "learning_rate": 0.0003,
      "loss": 2.594828987121582,
      "step": 2260
    },
    {
      "epoch": 0.454,
      "grad_norm": 0.060546875,
      "learning_rate": 0.0003,
      "loss": 2.6015480041503904,
      "step": 2270
    },
    {
      "epoch": 0.456,
      "grad_norm": 0.0576171875,
      "learning_rate": 0.0003,
      "loss": 2.6068796157836913,
      "step": 2280
    },
    {
      "epoch": 0.458,
      "grad_norm": 0.5859375,
      "learning_rate": 0.0003,
      "loss": 2.59091854095459,
      "step": 2290
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.1142578125,
      "learning_rate": 0.0003,
      "loss": 2.5947988510131834,
      "step": 2300
    },
    {
      "epoch": 0.46,
      "eval_loss": 3.028251886367798,
      "eval_runtime": 4.4313,
      "eval_samples_per_second": 67.249,
      "eval_steps_per_second": 3.836,
      "step": 2300
    },
    {
      "epoch": 0.462,
      "grad_norm": 0.0673828125,
      "learning_rate": 0.0003,
      "loss": 2.604340744018555,
      "step": 2310
    },
    {
      "epoch": 0.464,
      "grad_norm": 0.08203125,
      "learning_rate": 0.0003,
      "loss": 2.610575866699219,
      "step": 2320
    },
    {
      "epoch": 0.466,
      "grad_norm": 0.26953125,
      "learning_rate": 0.0003,
      "loss": 2.5987600326538085,
      "step": 2330
    },
    {
      "epoch": 0.468,
      "grad_norm": 0.1572265625,
      "learning_rate": 0.0003,
      "loss": 2.5745519638061523,
      "step": 2340
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.0751953125,
      "learning_rate": 0.0003,
      "loss": 2.5737417221069334,
      "step": 2350
    },
    {
      "epoch": 0.472,
      "grad_norm": 0.061279296875,
      "learning_rate": 0.0003,
      "loss": 2.586330795288086,
      "step": 2360
    },
    {
      "epoch": 0.474,
      "grad_norm": 0.04931640625,
      "learning_rate": 0.0003,
      "loss": 2.6139583587646484,
      "step": 2370
    },
    {
      "epoch": 0.476,
      "grad_norm": 0.05908203125,
      "learning_rate": 0.0003,
      "loss": 2.55531005859375,
      "step": 2380
    },
    {
      "epoch": 0.478,
      "grad_norm": 0.060302734375,
      "learning_rate": 0.0003,
      "loss": 2.601852226257324,
      "step": 2390
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.05712890625,
      "learning_rate": 0.0003,
      "loss": 2.6031288146972655,
      "step": 2400
    },
    {
      "epoch": 0.48,
      "eval_loss": 3.02764892578125,
      "eval_runtime": 4.4578,
      "eval_samples_per_second": 66.849,
      "eval_steps_per_second": 3.814,
      "step": 2400
    },
    {
      "epoch": 0.482,
      "grad_norm": 0.053466796875,
      "learning_rate": 0.0003,
      "loss": 2.5988203048706056,
      "step": 2410
    },
    {
      "epoch": 0.484,
      "grad_norm": 0.05517578125,
      "learning_rate": 0.0003,
      "loss": 2.5995391845703124,
      "step": 2420
    },
    {
      "epoch": 0.486,
      "grad_norm": 0.054443359375,
      "learning_rate": 0.0003,
      "loss": 2.5926639556884767,
      "step": 2430
    },
    {
      "epoch": 0.488,
      "grad_norm": 0.050537109375,
      "learning_rate": 0.0003,
      "loss": 2.615771484375,
      "step": 2440
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.216796875,
      "learning_rate": 0.0003,
      "loss": 2.5967260360717774,
      "step": 2450
    },
    {
      "epoch": 0.492,
      "grad_norm": 0.1171875,
      "learning_rate": 0.0003,
      "loss": 2.6070770263671874,
      "step": 2460
    },
    {
      "epoch": 0.494,
      "grad_norm": 0.04931640625,
      "learning_rate": 0.0003,
      "loss": 2.632620620727539,
      "step": 2470
    },
    {
      "epoch": 0.496,
      "grad_norm": 0.08056640625,
      "learning_rate": 0.0003,
      "loss": 2.6182647705078126,
      "step": 2480
    },
    {
      "epoch": 0.498,
      "grad_norm": 0.0537109375,
      "learning_rate": 0.0003,
      "loss": 2.6368099212646485,
      "step": 2490
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.80078125,
      "learning_rate": 0.0003,
      "loss": 2.607630729675293,
      "step": 2500
    },
    {
      "epoch": 0.5,
      "eval_loss": 3.0262646675109863,
      "eval_runtime": 4.4503,
      "eval_samples_per_second": 66.962,
      "eval_steps_per_second": 3.82,
      "step": 2500
    },
    {
      "epoch": 0.502,
      "grad_norm": 0.048095703125,
      "learning_rate": 0.0003,
      "loss": 2.600408363342285,
      "step": 2510
    },
    {
      "epoch": 0.504,
      "grad_norm": 0.052001953125,
      "learning_rate": 0.0003,
      "loss": 2.6103189468383787,
      "step": 2520
    },
    {
      "epoch": 0.506,
      "grad_norm": 0.05712890625,
      "learning_rate": 0.0003,
      "loss": 2.5979167938232424,
      "step": 2530
    },
    {
      "epoch": 0.508,
      "grad_norm": 0.049560546875,
      "learning_rate": 0.0003,
      "loss": 2.5972728729248047,
      "step": 2540
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.050048828125,
      "learning_rate": 0.0003,
      "loss": 2.608352851867676,
      "step": 2550
    },
    {
      "epoch": 0.512,
      "grad_norm": 0.07275390625,
      "learning_rate": 0.0003,
      "loss": 2.600900650024414,
      "step": 2560
    },
    {
      "epoch": 0.514,
      "grad_norm": 0.052734375,
      "learning_rate": 0.0003,
      "loss": 2.638876533508301,
      "step": 2570
    },
    {
      "epoch": 0.516,
      "grad_norm": 0.05126953125,
      "learning_rate": 0.0003,
      "loss": 2.610831069946289,
      "step": 2580
    },
    {
      "epoch": 0.518,
      "grad_norm": 0.051513671875,
      "learning_rate": 0.0003,
      "loss": 2.6283756256103517,
      "step": 2590
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.0791015625,
      "learning_rate": 0.0003,
      "loss": 2.6440673828125,
      "step": 2600
    },
    {
      "epoch": 0.52,
      "eval_loss": 3.0254361629486084,
      "eval_runtime": 4.3572,
      "eval_samples_per_second": 68.392,
      "eval_steps_per_second": 3.902,
      "step": 2600
    },
    {
      "epoch": 0.522,
      "grad_norm": 0.05029296875,
      "learning_rate": 0.0003,
      "loss": 2.625893211364746,
      "step": 2610
    },
    {
      "epoch": 0.524,
      "grad_norm": 0.061767578125,
      "learning_rate": 0.0003,
      "loss": 2.6046930313110352,
      "step": 2620
    },
    {
      "epoch": 0.526,
      "grad_norm": 0.055419921875,
      "learning_rate": 0.0003,
      "loss": 2.623915100097656,
      "step": 2630
    },
    {
      "epoch": 0.528,
      "grad_norm": 0.058349609375,
      "learning_rate": 0.0003,
      "loss": 2.6163360595703127,
      "step": 2640
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.04736328125,
      "learning_rate": 0.0003,
      "loss": 2.5871484756469725,
      "step": 2650
    },
    {
      "epoch": 0.532,
      "grad_norm": 0.059326171875,
      "learning_rate": 0.0003,
      "loss": 2.6428932189941405,
      "step": 2660
    },
    {
      "epoch": 0.534,
      "grad_norm": 0.0498046875,
      "learning_rate": 0.0003,
      "loss": 2.608722686767578,
      "step": 2670
    },
    {
      "epoch": 0.536,
      "grad_norm": 0.224609375,
      "learning_rate": 0.0003,
      "loss": 2.6207298278808593,
      "step": 2680
    },
    {
      "epoch": 0.538,
      "grad_norm": 0.056884765625,
      "learning_rate": 0.0003,
      "loss": 2.598964309692383,
      "step": 2690
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.185546875,
      "learning_rate": 0.0003,
      "loss": 2.6367502212524414,
      "step": 2700
    },
    {
      "epoch": 0.54,
      "eval_loss": 3.0251870155334473,
      "eval_runtime": 4.446,
      "eval_samples_per_second": 67.027,
      "eval_steps_per_second": 3.824,
      "step": 2700
    },
    {
      "epoch": 0.542,
      "grad_norm": 0.0732421875,
      "learning_rate": 0.0003,
      "loss": 2.625163269042969,
      "step": 2710
    },
    {
      "epoch": 0.544,
      "grad_norm": 0.054931640625,
      "learning_rate": 0.0003,
      "loss": 2.641737365722656,
      "step": 2720
    },
    {
      "epoch": 0.546,
      "grad_norm": 0.1845703125,
      "learning_rate": 0.0003,
      "loss": 2.6120086669921876,
      "step": 2730
    },
    {
      "epoch": 0.548,
      "grad_norm": 0.049560546875,
      "learning_rate": 0.0003,
      "loss": 2.5945173263549806,
      "step": 2740
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.060302734375,
      "learning_rate": 0.0003,
      "loss": 2.635161590576172,
      "step": 2750
    },
    {
      "epoch": 0.552,
      "grad_norm": 1.8046875,
      "learning_rate": 0.0003,
      "loss": 2.6317985534667967,
      "step": 2760
    },
    {
      "epoch": 0.554,
      "grad_norm": 0.0517578125,
      "learning_rate": 0.0003,
      "loss": 2.6268796920776367,
      "step": 2770
    },
    {
      "epoch": 0.556,
      "grad_norm": 2.015625,
      "learning_rate": 0.0003,
      "loss": 2.639748382568359,
      "step": 2780
    },
    {
      "epoch": 0.558,
      "grad_norm": 0.061767578125,
      "learning_rate": 0.0003,
      "loss": 2.6535463333129883,
      "step": 2790
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.06396484375,
      "learning_rate": 0.0003,
      "loss": 2.609045219421387,
      "step": 2800
    },
    {
      "epoch": 0.56,
      "eval_loss": 3.024867534637451,
      "eval_runtime": 4.4522,
      "eval_samples_per_second": 66.933,
      "eval_steps_per_second": 3.818,
      "step": 2800
    },
    {
      "epoch": 0.562,
      "grad_norm": 0.064453125,
      "learning_rate": 0.0003,
      "loss": 2.6145769119262696,
      "step": 2810
    },
    {
      "epoch": 0.564,
      "grad_norm": 0.052978515625,
      "learning_rate": 0.0003,
      "loss": 2.603018379211426,
      "step": 2820
    },
    {
      "epoch": 0.566,
      "grad_norm": 0.05615234375,
      "learning_rate": 0.0003,
      "loss": 2.6036045074462892,
      "step": 2830
    },
    {
      "epoch": 0.568,
      "grad_norm": 0.049560546875,
      "learning_rate": 0.0003,
      "loss": 2.610326385498047,
      "step": 2840
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.060546875,
      "learning_rate": 0.0003,
      "loss": 2.5892526626586916,
      "step": 2850
    },
    {
      "epoch": 0.572,
      "grad_norm": 0.06103515625,
      "learning_rate": 0.0003,
      "loss": 2.6054714202880858,
      "step": 2860
    },
    {
      "epoch": 0.574,
      "grad_norm": 0.05810546875,
      "learning_rate": 0.0003,
      "loss": 2.622623062133789,
      "step": 2870
    },
    {
      "epoch": 0.576,
      "grad_norm": 0.384765625,
      "learning_rate": 0.0003,
      "loss": 2.6183904647827148,
      "step": 2880
    },
    {
      "epoch": 0.578,
      "grad_norm": 4.875,
      "learning_rate": 0.0003,
      "loss": 2.5992977142333986,
      "step": 2890
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.154296875,
      "learning_rate": 0.0003,
      "loss": 2.6220735549926757,
      "step": 2900
    },
    {
      "epoch": 0.58,
      "eval_loss": 3.024597644805908,
      "eval_runtime": 4.4485,
      "eval_samples_per_second": 66.989,
      "eval_steps_per_second": 3.822,
      "step": 2900
    },
    {
      "epoch": 0.582,
      "grad_norm": 0.046875,
      "learning_rate": 0.0003,
      "loss": 2.6097631454467773,
      "step": 2910
    },
    {
      "epoch": 0.584,
      "grad_norm": 0.435546875,
      "learning_rate": 0.0003,
      "loss": 2.6088932037353514,
      "step": 2920
    },
    {
      "epoch": 0.586,
      "grad_norm": 0.06494140625,
      "learning_rate": 0.0003,
      "loss": 2.6292213439941405,
      "step": 2930
    },
    {
      "epoch": 0.588,
      "grad_norm": 0.052734375,
      "learning_rate": 0.0003,
      "loss": 2.6273422241210938,
      "step": 2940
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.09716796875,
      "learning_rate": 0.0003,
      "loss": 2.622287368774414,
      "step": 2950
    },
    {
      "epoch": 0.592,
      "grad_norm": 0.0498046875,
      "learning_rate": 0.0003,
      "loss": 2.5906503677368162,
      "step": 2960
    },
    {
      "epoch": 0.594,
      "grad_norm": 0.057373046875,
      "learning_rate": 0.0003,
      "loss": 2.631303596496582,
      "step": 2970
    },
    {
      "epoch": 0.596,
      "grad_norm": 0.054443359375,
      "learning_rate": 0.0003,
      "loss": 2.6114727020263673,
      "step": 2980
    },
    {
      "epoch": 0.598,
      "grad_norm": 0.0908203125,
      "learning_rate": 0.0003,
      "loss": 2.603567886352539,
      "step": 2990
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.0556640625,
      "learning_rate": 0.0003,
      "loss": 2.585314178466797,
      "step": 3000
    },
    {
      "epoch": 0.6,
      "eval_loss": 3.0278046131134033,
      "eval_runtime": 4.4444,
      "eval_samples_per_second": 67.051,
      "eval_steps_per_second": 3.825,
      "step": 3000
    },
    {
      "epoch": 0.602,
      "grad_norm": 0.055908203125,
      "learning_rate": 0.0003,
      "loss": 2.6127801895141602,
      "step": 3010
    },
    {
      "epoch": 0.604,
      "grad_norm": 0.1455078125,
      "learning_rate": 0.0003,
      "loss": 2.5975704193115234,
      "step": 3020
    },
    {
      "epoch": 0.606,
      "grad_norm": 15.0625,
      "learning_rate": 0.0003,
      "loss": 2.5873924255371095,
      "step": 3030
    },
    {
      "epoch": 0.608,
      "grad_norm": 0.0478515625,
      "learning_rate": 0.0003,
      "loss": 2.5901187896728515,
      "step": 3040
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.166015625,
      "learning_rate": 0.0003,
      "loss": 2.585931968688965,
      "step": 3050
    },
    {
      "epoch": 0.612,
      "grad_norm": 0.053466796875,
      "learning_rate": 0.0003,
      "loss": 2.6187633514404296,
      "step": 3060
    },
    {
      "epoch": 0.614,
      "grad_norm": 0.0556640625,
      "learning_rate": 0.0003,
      "loss": 2.6473546981811524,
      "step": 3070
    },
    {
      "epoch": 0.616,
      "grad_norm": 0.055908203125,
      "learning_rate": 0.0003,
      "loss": 2.606519317626953,
      "step": 3080
    },
    {
      "epoch": 0.618,
      "grad_norm": 12.0,
      "learning_rate": 0.0003,
      "loss": 2.6005083084106446,
      "step": 3090
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.0615234375,
      "learning_rate": 0.0003,
      "loss": 2.6309852600097656,
      "step": 3100
    },
    {
      "epoch": 0.62,
      "eval_loss": 3.0271658897399902,
      "eval_runtime": 4.4454,
      "eval_samples_per_second": 67.036,
      "eval_steps_per_second": 3.824,
      "step": 3100
    },
    {
      "epoch": 0.622,
      "grad_norm": 0.107421875,
      "learning_rate": 0.0003,
      "loss": 2.6178503036499023,
      "step": 3110
    },
    {
      "epoch": 0.624,
      "grad_norm": 0.06640625,
      "learning_rate": 0.0003,
      "loss": 2.614837646484375,
      "step": 3120
    },
    {
      "epoch": 0.626,
      "grad_norm": 1.203125,
      "learning_rate": 0.0003,
      "loss": 2.62316837310791,
      "step": 3130
    },
    {
      "epoch": 0.628,
      "grad_norm": 0.0517578125,
      "learning_rate": 0.0003,
      "loss": 2.59710693359375,
      "step": 3140
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.058349609375,
      "learning_rate": 0.0003,
      "loss": 2.6030540466308594,
      "step": 3150
    },
    {
      "epoch": 0.632,
      "grad_norm": 0.05029296875,
      "learning_rate": 0.0003,
      "loss": 2.5992559432983398,
      "step": 3160
    },
    {
      "epoch": 0.634,
      "grad_norm": 0.052490234375,
      "learning_rate": 0.0003,
      "loss": 2.6360668182373046,
      "step": 3170
    },
    {
      "epoch": 0.636,
      "grad_norm": 0.11279296875,
      "learning_rate": 0.0003,
      "loss": 2.6165666580200195,
      "step": 3180
    },
    {
      "epoch": 0.638,
      "grad_norm": 0.052978515625,
      "learning_rate": 0.0003,
      "loss": 2.636544036865234,
      "step": 3190
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.0947265625,
      "learning_rate": 0.0003,
      "loss": 2.6081783294677736,
      "step": 3200
    },
    {
      "epoch": 0.64,
      "eval_loss": 3.0295300483703613,
      "eval_runtime": 4.4243,
      "eval_samples_per_second": 67.356,
      "eval_steps_per_second": 3.842,
      "step": 3200
    },
    {
      "epoch": 0.642,
      "grad_norm": 0.07470703125,
      "learning_rate": 0.0003,
      "loss": 2.6136301040649412,
      "step": 3210
    },
    {
      "epoch": 0.644,
      "grad_norm": 5.59375,
      "learning_rate": 0.0003,
      "loss": 2.611122703552246,
      "step": 3220
    },
    {
      "epoch": 0.646,
      "grad_norm": 0.0888671875,
      "learning_rate": 0.0003,
      "loss": 2.611610984802246,
      "step": 3230
    },
    {
      "epoch": 0.648,
      "grad_norm": 6.4375,
      "learning_rate": 0.0003,
      "loss": 2.6010536193847655,
      "step": 3240
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.051513671875,
      "learning_rate": 0.0003,
      "loss": 2.623811149597168,
      "step": 3250
    },
    {
      "epoch": 0.652,
      "grad_norm": 0.06005859375,
      "learning_rate": 0.0003,
      "loss": 2.5976247787475586,
      "step": 3260
    },
    {
      "epoch": 0.654,
      "grad_norm": 0.05859375,
      "learning_rate": 0.0003,
      "loss": 2.593853569030762,
      "step": 3270
    },
    {
      "epoch": 0.656,
      "grad_norm": 0.060546875,
      "learning_rate": 0.0003,
      "loss": 2.645317268371582,
      "step": 3280
    },
    {
      "epoch": 0.658,
      "grad_norm": 0.0595703125,
      "learning_rate": 0.0003,
      "loss": 2.592297172546387,
      "step": 3290
    },
    {
      "epoch": 0.66,
      "grad_norm": 2.21875,
      "learning_rate": 0.0003,
      "loss": 2.60241641998291,
      "step": 3300
    },
    {
      "epoch": 0.66,
      "eval_loss": 3.0250847339630127,
      "eval_runtime": 4.4783,
      "eval_samples_per_second": 66.543,
      "eval_steps_per_second": 3.796,
      "step": 3300
    },
    {
      "epoch": 0.662,
      "grad_norm": 0.052490234375,
      "learning_rate": 0.0003,
      "loss": 2.600591278076172,
      "step": 3310
    },
    {
      "epoch": 0.664,
      "grad_norm": 0.052978515625,
      "learning_rate": 0.0003,
      "loss": 2.6144927978515624,
      "step": 3320
    },
    {
      "epoch": 0.666,
      "grad_norm": 0.057861328125,
      "learning_rate": 0.0003,
      "loss": 2.593229866027832,
      "step": 3330
    },
    {
      "epoch": 0.668,
      "grad_norm": 0.0546875,
      "learning_rate": 0.0003,
      "loss": 2.6041738510131838,
      "step": 3340
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.052978515625,
      "learning_rate": 0.0003,
      "loss": 2.607752227783203,
      "step": 3350
    },
    {
      "epoch": 0.672,
      "grad_norm": 1.671875,
      "learning_rate": 0.0003,
      "loss": 2.6296701431274414,
      "step": 3360
    },
    {
      "epoch": 0.674,
      "grad_norm": 0.55859375,
      "learning_rate": 0.0003,
      "loss": 2.5772432327270507,
      "step": 3370
    },
    {
      "epoch": 0.676,
      "grad_norm": 0.052734375,
      "learning_rate": 0.0003,
      "loss": 2.6260414123535156,
      "step": 3380
    },
    {
      "epoch": 0.678,
      "grad_norm": 0.062255859375,
      "learning_rate": 0.0003,
      "loss": 2.5910011291503907,
      "step": 3390
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.0546875,
      "learning_rate": 0.0003,
      "loss": 2.5773406982421876,
      "step": 3400
    },
    {
      "epoch": 0.68,
      "eval_loss": 3.0262410640716553,
      "eval_runtime": 4.4195,
      "eval_samples_per_second": 67.428,
      "eval_steps_per_second": 3.847,
      "step": 3400
    },
    {
      "epoch": 0.682,
      "grad_norm": 0.056640625,
      "learning_rate": 0.0003,
      "loss": 2.602094268798828,
      "step": 3410
    },
    {
      "epoch": 0.684,
      "grad_norm": 0.04931640625,
      "learning_rate": 0.0003,
      "loss": 2.6283613204956056,
      "step": 3420
    },
    {
      "epoch": 0.686,
      "grad_norm": 0.26171875,
      "learning_rate": 0.0003,
      "loss": 2.58358211517334,
      "step": 3430
    },
    {
      "epoch": 0.688,
      "grad_norm": 0.052978515625,
      "learning_rate": 0.0003,
      "loss": 2.5964986801147463,
      "step": 3440
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.05322265625,
      "learning_rate": 0.0003,
      "loss": 2.5851970672607423,
      "step": 3450
    },
    {
      "epoch": 0.692,
      "grad_norm": 0.051025390625,
      "learning_rate": 0.0003,
      "loss": 2.6003726959228515,
      "step": 3460
    },
    {
      "epoch": 0.694,
      "grad_norm": 0.050048828125,
      "learning_rate": 0.0003,
      "loss": 2.6186376571655274,
      "step": 3470
    },
    {
      "epoch": 0.696,
      "grad_norm": 0.052001953125,
      "learning_rate": 0.0003,
      "loss": 2.624268341064453,
      "step": 3480
    },
    {
      "epoch": 0.698,
      "grad_norm": 0.05029296875,
      "learning_rate": 0.0003,
      "loss": 2.5755191802978517,
      "step": 3490
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.244140625,
      "learning_rate": 0.0003,
      "loss": 2.593659782409668,
      "step": 3500
    },
    {
      "epoch": 0.7,
      "eval_loss": 3.0274429321289062,
      "eval_runtime": 4.4166,
      "eval_samples_per_second": 67.472,
      "eval_steps_per_second": 3.849,
      "step": 3500
    },
    {
      "epoch": 0.702,
      "grad_norm": 0.053466796875,
      "learning_rate": 0.0003,
      "loss": 2.599693489074707,
      "step": 3510
    },
    {
      "epoch": 0.704,
      "grad_norm": 0.07275390625,
      "learning_rate": 0.0003,
      "loss": 2.599070167541504,
      "step": 3520
    },
    {
      "epoch": 0.706,
      "grad_norm": 0.055419921875,
      "learning_rate": 0.0003,
      "loss": 2.6012365341186525,
      "step": 3530
    },
    {
      "epoch": 0.708,
      "grad_norm": 0.05224609375,
      "learning_rate": 0.0003,
      "loss": 2.576295852661133,
      "step": 3540
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.0517578125,
      "learning_rate": 0.0003,
      "loss": 2.6024765014648437,
      "step": 3550
    },
    {
      "epoch": 0.712,
      "grad_norm": 0.0693359375,
      "learning_rate": 0.0003,
      "loss": 2.5685857772827148,
      "step": 3560
    },
    {
      "epoch": 0.714,
      "grad_norm": 0.1220703125,
      "learning_rate": 0.0003,
      "loss": 2.584480094909668,
      "step": 3570
    },
    {
      "epoch": 0.716,
      "grad_norm": 0.06396484375,
      "learning_rate": 0.0003,
      "loss": 2.6187564849853517,
      "step": 3580
    },
    {
      "epoch": 0.718,
      "grad_norm": 0.05126953125,
      "learning_rate": 0.0003,
      "loss": 2.608148384094238,
      "step": 3590
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.056396484375,
      "learning_rate": 0.0003,
      "loss": 2.596156883239746,
      "step": 3600
    },
    {
      "epoch": 0.72,
      "eval_loss": 3.0285751819610596,
      "eval_runtime": 4.4537,
      "eval_samples_per_second": 66.91,
      "eval_steps_per_second": 3.817,
      "step": 3600
    },
    {
      "epoch": 0.722,
      "grad_norm": 2.265625,
      "learning_rate": 0.0003,
      "loss": 2.6053913116455076,
      "step": 3610
    },
    {
      "epoch": 0.724,
      "grad_norm": 0.1259765625,
      "learning_rate": 0.0003,
      "loss": 2.5906614303588866,
      "step": 3620
    },
    {
      "epoch": 0.726,
      "grad_norm": 0.058349609375,
      "learning_rate": 0.0003,
      "loss": 2.6229223251342773,
      "step": 3630
    },
    {
      "epoch": 0.728,
      "grad_norm": 0.054443359375,
      "learning_rate": 0.0003,
      "loss": 2.6100414276123045,
      "step": 3640
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.05078125,
      "learning_rate": 0.0003,
      "loss": 2.5861778259277344,
      "step": 3650
    },
    {
      "epoch": 0.732,
      "grad_norm": 0.06884765625,
      "learning_rate": 0.0003,
      "loss": 2.6088991165161133,
      "step": 3660
    },
    {
      "epoch": 0.734,
      "grad_norm": 0.0546875,
      "learning_rate": 0.0003,
      "loss": 2.5957805633544924,
      "step": 3670
    },
    {
      "epoch": 0.736,
      "grad_norm": 0.08447265625,
      "learning_rate": 0.0003,
      "loss": 2.5968517303466796,
      "step": 3680
    },
    {
      "epoch": 0.738,
      "grad_norm": 0.06396484375,
      "learning_rate": 0.0003,
      "loss": 2.590867614746094,
      "step": 3690
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.057373046875,
      "learning_rate": 0.0003,
      "loss": 2.6020263671875,
      "step": 3700
    },
    {
      "epoch": 0.74,
      "eval_loss": 3.033440113067627,
      "eval_runtime": 4.4438,
      "eval_samples_per_second": 67.059,
      "eval_steps_per_second": 3.826,
      "step": 3700
    },
    {
      "epoch": 0.742,
      "grad_norm": 1.8359375,
      "learning_rate": 0.0003,
      "loss": 2.624915885925293,
      "step": 3710
    },
    {
      "epoch": 0.744,
      "grad_norm": 0.05322265625,
      "learning_rate": 0.0003,
      "loss": 2.6158367156982423,
      "step": 3720
    },
    {
      "epoch": 0.746,
      "grad_norm": 0.052978515625,
      "learning_rate": 0.0003,
      "loss": 2.5968292236328123,
      "step": 3730
    },
    {
      "epoch": 0.748,
      "grad_norm": 0.0849609375,
      "learning_rate": 0.0003,
      "loss": 2.61802978515625,
      "step": 3740
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.05517578125,
      "learning_rate": 0.0003,
      "loss": 2.5647909164428713,
      "step": 3750
    },
    {
      "epoch": 0.752,
      "grad_norm": 0.060791015625,
      "learning_rate": 0.0003,
      "loss": 2.601895332336426,
      "step": 3760
    },
    {
      "epoch": 0.754,
      "grad_norm": 0.0556640625,
      "learning_rate": 0.0003,
      "loss": 2.605121612548828,
      "step": 3770
    },
    {
      "epoch": 0.756,
      "grad_norm": 0.09716796875,
      "learning_rate": 0.0003,
      "loss": 2.6213621139526366,
      "step": 3780
    },
    {
      "epoch": 0.758,
      "grad_norm": 0.05078125,
      "learning_rate": 0.0003,
      "loss": 2.5848875045776367,
      "step": 3790
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.05859375,
      "learning_rate": 0.0003,
      "loss": 2.585398483276367,
      "step": 3800
    },
    {
      "epoch": 0.76,
      "eval_loss": 3.0310301780700684,
      "eval_runtime": 4.4352,
      "eval_samples_per_second": 67.19,
      "eval_steps_per_second": 3.833,
      "step": 3800
    },
    {
      "epoch": 0.762,
      "grad_norm": 0.0830078125,
      "learning_rate": 0.0003,
      "loss": 2.5891979217529295,
      "step": 3810
    },
    {
      "epoch": 0.764,
      "grad_norm": 0.048828125,
      "learning_rate": 0.0003,
      "loss": 2.587088203430176,
      "step": 3820
    },
    {
      "epoch": 0.766,
      "grad_norm": 0.056884765625,
      "learning_rate": 0.0003,
      "loss": 2.6202329635620116,
      "step": 3830
    },
    {
      "epoch": 0.768,
      "grad_norm": 0.058349609375,
      "learning_rate": 0.0003,
      "loss": 2.588342475891113,
      "step": 3840
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.051513671875,
      "learning_rate": 0.0003,
      "loss": 2.570342445373535,
      "step": 3850
    },
    {
      "epoch": 0.772,
      "grad_norm": 0.054443359375,
      "learning_rate": 0.0003,
      "loss": 2.5928205490112304,
      "step": 3860
    },
    {
      "epoch": 0.774,
      "grad_norm": 1.1484375,
      "learning_rate": 0.0003,
      "loss": 2.562236785888672,
      "step": 3870
    },
    {
      "epoch": 0.776,
      "grad_norm": 0.05322265625,
      "learning_rate": 0.0003,
      "loss": 2.5821876525878906,
      "step": 3880
    },
    {
      "epoch": 0.778,
      "grad_norm": 0.05029296875,
      "learning_rate": 0.0003,
      "loss": 2.583372688293457,
      "step": 3890
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.197265625,
      "learning_rate": 0.0003,
      "loss": 2.5738046646118162,
      "step": 3900
    },
    {
      "epoch": 0.78,
      "eval_loss": 3.029505491256714,
      "eval_runtime": 4.3771,
      "eval_samples_per_second": 68.082,
      "eval_steps_per_second": 3.884,
      "step": 3900
    },
    {
      "epoch": 0.782,
      "grad_norm": 0.173828125,
      "learning_rate": 0.0003,
      "loss": 2.576494598388672,
      "step": 3910
    },
    {
      "epoch": 0.784,
      "grad_norm": 0.380859375,
      "learning_rate": 0.0003,
      "loss": 2.5919593811035155,
      "step": 3920
    },
    {
      "epoch": 0.786,
      "grad_norm": 0.048095703125,
      "learning_rate": 0.0003,
      "loss": 2.621496391296387,
      "step": 3930
    },
    {
      "epoch": 0.788,
      "grad_norm": 0.052490234375,
      "learning_rate": 0.0003,
      "loss": 2.584731101989746,
      "step": 3940
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.0498046875,
      "learning_rate": 0.0003,
      "loss": 2.568536567687988,
      "step": 3950
    },
    {
      "epoch": 0.792,
      "grad_norm": 0.0615234375,
      "learning_rate": 0.0003,
      "loss": 2.590066909790039,
      "step": 3960
    },
    {
      "epoch": 0.794,
      "grad_norm": 0.060791015625,
      "learning_rate": 0.0003,
      "loss": 2.613643264770508,
      "step": 3970
    },
    {
      "epoch": 0.796,
      "grad_norm": 0.052734375,
      "learning_rate": 0.0003,
      "loss": 2.614491081237793,
      "step": 3980
    },
    {
      "epoch": 0.798,
      "grad_norm": 0.0517578125,
      "learning_rate": 0.0003,
      "loss": 2.591320037841797,
      "step": 3990
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.04931640625,
      "learning_rate": 0.0003,
      "loss": 2.605324554443359,
      "step": 4000
    },
    {
      "epoch": 0.8,
      "eval_loss": 3.0305187702178955,
      "eval_runtime": 4.4657,
      "eval_samples_per_second": 66.73,
      "eval_steps_per_second": 3.807,
      "step": 4000
    },
    {
      "epoch": 0.802,
      "grad_norm": 0.054931640625,
      "learning_rate": 0.0003,
      "loss": 2.5756288528442384,
      "step": 4010
    },
    {
      "epoch": 0.804,
      "grad_norm": 0.5546875,
      "learning_rate": 0.0003,
      "loss": 2.591953468322754,
      "step": 4020
    },
    {
      "epoch": 0.806,
      "grad_norm": 0.05029296875,
      "learning_rate": 0.0003,
      "loss": 2.5956727981567385,
      "step": 4030
    },
    {
      "epoch": 0.808,
      "grad_norm": 0.05224609375,
      "learning_rate": 0.0003,
      "loss": 2.60211067199707,
      "step": 4040
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.049072265625,
      "learning_rate": 0.0003,
      "loss": 2.5871608734130858,
      "step": 4050
    },
    {
      "epoch": 0.812,
      "grad_norm": 0.10693359375,
      "learning_rate": 0.0003,
      "loss": 2.5698373794555662,
      "step": 4060
    },
    {
      "epoch": 0.814,
      "grad_norm": 0.05810546875,
      "learning_rate": 0.0003,
      "loss": 2.6126802444458006,
      "step": 4070
    },
    {
      "epoch": 0.816,
      "grad_norm": 0.05029296875,
      "learning_rate": 0.0003,
      "loss": 2.5981868743896483,
      "step": 4080
    },
    {
      "epoch": 0.818,
      "grad_norm": 0.052978515625,
      "learning_rate": 0.0003,
      "loss": 2.580386161804199,
      "step": 4090
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.07080078125,
      "learning_rate": 0.0003,
      "loss": 2.5840375900268553,
      "step": 4100
    },
    {
      "epoch": 0.82,
      "eval_loss": 3.0312201976776123,
      "eval_runtime": 4.4429,
      "eval_samples_per_second": 67.074,
      "eval_steps_per_second": 3.826,
      "step": 4100
    },
    {
      "epoch": 0.822,
      "grad_norm": 0.052978515625,
      "learning_rate": 0.0003,
      "loss": 2.6040040969848635,
      "step": 4110
    },
    {
      "epoch": 0.824,
      "grad_norm": 0.05224609375,
      "learning_rate": 0.0003,
      "loss": 2.5784271240234373,
      "step": 4120
    },
    {
      "epoch": 0.826,
      "grad_norm": 0.05078125,
      "learning_rate": 0.0003,
      "loss": 2.5992313385009767,
      "step": 4130
    },
    {
      "epoch": 0.828,
      "grad_norm": 0.62109375,
      "learning_rate": 0.0003,
      "loss": 2.607274055480957,
      "step": 4140
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.05126953125,
      "learning_rate": 0.0003,
      "loss": 2.5900325775146484,
      "step": 4150
    },
    {
      "epoch": 0.832,
      "grad_norm": 0.0556640625,
      "learning_rate": 0.0003,
      "loss": 2.584280586242676,
      "step": 4160
    },
    {
      "epoch": 0.834,
      "grad_norm": 0.06103515625,
      "learning_rate": 0.0003,
      "loss": 2.578823471069336,
      "step": 4170
    },
    {
      "epoch": 0.836,
      "grad_norm": 0.05908203125,
      "learning_rate": 0.0003,
      "loss": 2.600706672668457,
      "step": 4180
    },
    {
      "epoch": 0.838,
      "grad_norm": 0.054931640625,
      "learning_rate": 0.0003,
      "loss": 2.5908893585205077,
      "step": 4190
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.08740234375,
      "learning_rate": 0.0003,
      "loss": 2.6078845977783205,
      "step": 4200
    },
    {
      "epoch": 0.84,
      "eval_loss": 3.0311102867126465,
      "eval_runtime": 4.4714,
      "eval_samples_per_second": 66.645,
      "eval_steps_per_second": 3.802,
      "step": 4200
    },
    {
      "epoch": 0.842,
      "grad_norm": 0.0556640625,
      "learning_rate": 0.0003,
      "loss": 2.630947303771973,
      "step": 4210
    },
    {
      "epoch": 0.844,
      "grad_norm": 0.09423828125,
      "learning_rate": 0.0003,
      "loss": 2.623428726196289,
      "step": 4220
    },
    {
      "epoch": 0.846,
      "grad_norm": 0.111328125,
      "learning_rate": 0.0003,
      "loss": 2.5977907180786133,
      "step": 4230
    },
    {
      "epoch": 0.848,
      "grad_norm": 0.455078125,
      "learning_rate": 0.0003,
      "loss": 2.6008005142211914,
      "step": 4240
    },
    {
      "epoch": 0.85,
      "grad_norm": 1.6015625,
      "learning_rate": 0.0003,
      "loss": 2.5678030014038087,
      "step": 4250
    },
    {
      "epoch": 0.852,
      "grad_norm": 0.052490234375,
      "learning_rate": 0.0003,
      "loss": 2.613857650756836,
      "step": 4260
    },
    {
      "epoch": 0.854,
      "grad_norm": 0.08349609375,
      "learning_rate": 0.0003,
      "loss": 2.579903221130371,
      "step": 4270
    },
    {
      "epoch": 0.856,
      "grad_norm": 0.049560546875,
      "learning_rate": 0.0003,
      "loss": 2.579732131958008,
      "step": 4280
    },
    {
      "epoch": 0.858,
      "grad_norm": 0.181640625,
      "learning_rate": 0.0003,
      "loss": 2.5955440521240236,
      "step": 4290
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.05712890625,
      "learning_rate": 0.0003,
      "loss": 2.592930793762207,
      "step": 4300
    },
    {
      "epoch": 0.86,
      "eval_loss": 3.031736373901367,
      "eval_runtime": 4.4299,
      "eval_samples_per_second": 67.27,
      "eval_steps_per_second": 3.838,
      "step": 4300
    },
    {
      "epoch": 0.862,
      "grad_norm": 0.05419921875,
      "learning_rate": 0.0003,
      "loss": 2.592413902282715,
      "step": 4310
    },
    {
      "epoch": 0.864,
      "grad_norm": 0.10986328125,
      "learning_rate": 0.0003,
      "loss": 2.569998931884766,
      "step": 4320
    },
    {
      "epoch": 0.866,
      "grad_norm": 0.048583984375,
      "learning_rate": 0.0003,
      "loss": 2.5922664642333983,
      "step": 4330
    },
    {
      "epoch": 0.868,
      "grad_norm": 0.052734375,
      "learning_rate": 0.0003,
      "loss": 2.6012556076049806,
      "step": 4340
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.07177734375,
      "learning_rate": 0.0003,
      "loss": 2.6031923294067383,
      "step": 4350
    },
    {
      "epoch": 0.872,
      "grad_norm": 0.07763671875,
      "learning_rate": 0.0003,
      "loss": 2.59698543548584,
      "step": 4360
    },
    {
      "epoch": 0.874,
      "grad_norm": 0.051513671875,
      "learning_rate": 0.0003,
      "loss": 2.6096282958984376,
      "step": 4370
    },
    {
      "epoch": 0.876,
      "grad_norm": 0.052734375,
      "learning_rate": 0.0003,
      "loss": 2.6013412475585938,
      "step": 4380
    },
    {
      "epoch": 0.878,
      "grad_norm": 0.051025390625,
      "learning_rate": 0.0003,
      "loss": 2.570247268676758,
      "step": 4390
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.19140625,
      "learning_rate": 0.0003,
      "loss": 2.6004173278808596,
      "step": 4400
    },
    {
      "epoch": 0.88,
      "eval_loss": 3.032118320465088,
      "eval_runtime": 4.4552,
      "eval_samples_per_second": 66.887,
      "eval_steps_per_second": 3.816,
      "step": 4400
    },
    {
      "epoch": 0.882,
      "grad_norm": 0.050048828125,
      "learning_rate": 0.0003,
      "loss": 2.60999755859375,
      "step": 4410
    },
    {
      "epoch": 0.884,
      "grad_norm": 0.1044921875,
      "learning_rate": 0.0003,
      "loss": 2.5999217987060548,
      "step": 4420
    },
    {
      "epoch": 0.886,
      "grad_norm": 0.255859375,
      "learning_rate": 0.0003,
      "loss": 2.57828369140625,
      "step": 4430
    },
    {
      "epoch": 0.888,
      "grad_norm": 0.0556640625,
      "learning_rate": 0.0003,
      "loss": 2.5876300811767576,
      "step": 4440
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.052490234375,
      "learning_rate": 0.0003,
      "loss": 2.5927480697631835,
      "step": 4450
    },
    {
      "epoch": 0.892,
      "grad_norm": 0.076171875,
      "learning_rate": 0.0003,
      "loss": 2.6045318603515626,
      "step": 4460
    },
    {
      "epoch": 0.894,
      "grad_norm": 0.06201171875,
      "learning_rate": 0.0003,
      "loss": 2.585988235473633,
      "step": 4470
    },
    {
      "epoch": 0.896,
      "grad_norm": 0.050048828125,
      "learning_rate": 0.0003,
      "loss": 2.613155555725098,
      "step": 4480
    },
    {
      "epoch": 0.898,
      "grad_norm": 4.125,
      "learning_rate": 0.0003,
      "loss": 2.592078971862793,
      "step": 4490
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.048095703125,
      "learning_rate": 0.0003,
      "loss": 2.591631317138672,
      "step": 4500
    },
    {
      "epoch": 0.9,
      "eval_loss": 3.028334617614746,
      "eval_runtime": 4.4441,
      "eval_samples_per_second": 67.055,
      "eval_steps_per_second": 3.825,
      "step": 4500
    }
  ],
  "logging_steps": 10,
  "max_steps": 5000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.467695316860928e+18,
  "train_batch_size": 18,
  "trial_name": null,
  "trial_params": null
}