{
  "best_metric": 0.03171728551387787,
  "best_model_checkpoint": "saves/psy-course/Qwen2.5-7B-Instruct/train/fold5/checkpoint-1850",
  "epoch": 4.994228549442093,
  "eval_steps": 50,
  "global_step": 3245,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.015390534821085032,
      "grad_norm": 3.703810453414917,
      "learning_rate": 3.0769230769230774e-06,
      "loss": 1.9992,
      "step": 10
    },
    {
      "epoch": 0.030781069642170065,
      "grad_norm": 4.092873573303223,
      "learning_rate": 6.153846153846155e-06,
      "loss": 1.9267,
      "step": 20
    },
    {
      "epoch": 0.0461716044632551,
      "grad_norm": 3.967761516571045,
      "learning_rate": 9.230769230769232e-06,
      "loss": 1.7205,
      "step": 30
    },
    {
      "epoch": 0.06156213928434013,
      "grad_norm": 3.3643269538879395,
      "learning_rate": 1.230769230769231e-05,
      "loss": 1.3791,
      "step": 40
    },
    {
      "epoch": 0.07695267410542517,
      "grad_norm": 2.090191602706909,
      "learning_rate": 1.5384615384615387e-05,
      "loss": 0.8754,
      "step": 50
    },
    {
      "epoch": 0.07695267410542517,
      "eval_loss": 0.6955563426017761,
      "eval_runtime": 155.819,
      "eval_samples_per_second": 7.419,
      "eval_steps_per_second": 7.419,
      "step": 50
    },
    {
      "epoch": 0.0923432089265102,
      "grad_norm": 0.8891865611076355,
      "learning_rate": 1.8461538461538465e-05,
      "loss": 0.5912,
      "step": 60
    },
    {
      "epoch": 0.10773374374759523,
      "grad_norm": 1.215137004852295,
      "learning_rate": 2.1538461538461542e-05,
      "loss": 0.4086,
      "step": 70
    },
    {
      "epoch": 0.12312427856868026,
      "grad_norm": 0.6001670360565186,
      "learning_rate": 2.461538461538462e-05,
      "loss": 0.2136,
      "step": 80
    },
    {
      "epoch": 0.1385148133897653,
      "grad_norm": 0.9708924293518066,
      "learning_rate": 2.7692307692307694e-05,
      "loss": 0.2183,
      "step": 90
    },
    {
      "epoch": 0.15390534821085033,
      "grad_norm": 0.6720219254493713,
      "learning_rate": 3.0769230769230774e-05,
      "loss": 0.1571,
      "step": 100
    },
    {
      "epoch": 0.15390534821085033,
      "eval_loss": 0.10784146934747696,
      "eval_runtime": 155.8128,
      "eval_samples_per_second": 7.419,
      "eval_steps_per_second": 7.419,
      "step": 100
    },
    {
      "epoch": 0.16929588303193535,
      "grad_norm": 0.8736152648925781,
      "learning_rate": 3.384615384615385e-05,
      "loss": 0.1103,
      "step": 110
    },
    {
      "epoch": 0.1846864178530204,
      "grad_norm": 0.8433762788772583,
      "learning_rate": 3.692307692307693e-05,
      "loss": 0.1089,
      "step": 120
    },
    {
      "epoch": 0.2000769526741054,
      "grad_norm": 1.2629233598709106,
      "learning_rate": 4e-05,
      "loss": 0.078,
      "step": 130
    },
    {
      "epoch": 0.21546748749519046,
      "grad_norm": 0.9535691142082214,
      "learning_rate": 4.3076923076923084e-05,
      "loss": 0.0833,
      "step": 140
    },
    {
      "epoch": 0.2308580223162755,
      "grad_norm": 0.7020408511161804,
      "learning_rate": 4.615384615384616e-05,
      "loss": 0.0871,
      "step": 150
    },
    {
      "epoch": 0.2308580223162755,
      "eval_loss": 0.07267580926418304,
      "eval_runtime": 155.8641,
      "eval_samples_per_second": 7.417,
      "eval_steps_per_second": 7.417,
      "step": 150
    },
    {
      "epoch": 0.24624855713736052,
      "grad_norm": 0.7574706077575684,
      "learning_rate": 4.923076923076924e-05,
      "loss": 0.0825,
      "step": 160
    },
    {
      "epoch": 0.26163909195844554,
      "grad_norm": 1.3506394624710083,
      "learning_rate": 5.230769230769231e-05,
      "loss": 0.0751,
      "step": 170
    },
    {
      "epoch": 0.2770296267795306,
      "grad_norm": 0.713176429271698,
      "learning_rate": 5.538461538461539e-05,
      "loss": 0.0595,
      "step": 180
    },
    {
      "epoch": 0.2924201616006156,
      "grad_norm": 0.655463695526123,
      "learning_rate": 5.846153846153847e-05,
      "loss": 0.0518,
      "step": 190
    },
    {
      "epoch": 0.30781069642170067,
      "grad_norm": 0.5643651485443115,
      "learning_rate": 6.153846153846155e-05,
      "loss": 0.073,
      "step": 200
    },
    {
      "epoch": 0.30781069642170067,
      "eval_loss": 0.055601589381694794,
      "eval_runtime": 155.9101,
      "eval_samples_per_second": 7.415,
      "eval_steps_per_second": 7.415,
      "step": 200
    },
    {
      "epoch": 0.3232012312427857,
      "grad_norm": 0.6759706139564514,
      "learning_rate": 6.461538461538462e-05,
      "loss": 0.0641,
      "step": 210
    },
    {
      "epoch": 0.3385917660638707,
      "grad_norm": 0.7807886004447937,
      "learning_rate": 6.76923076923077e-05,
      "loss": 0.062,
      "step": 220
    },
    {
      "epoch": 0.35398230088495575,
      "grad_norm": 0.3335706889629364,
      "learning_rate": 7.076923076923078e-05,
      "loss": 0.0467,
      "step": 230
    },
    {
      "epoch": 0.3693728357060408,
      "grad_norm": 0.6220697164535522,
      "learning_rate": 7.384615384615386e-05,
      "loss": 0.0581,
      "step": 240
    },
    {
      "epoch": 0.38476337052712584,
      "grad_norm": 0.4891945719718933,
      "learning_rate": 7.692307692307693e-05,
      "loss": 0.0552,
      "step": 250
    },
    {
      "epoch": 0.38476337052712584,
      "eval_loss": 0.052611254155635834,
      "eval_runtime": 155.71,
      "eval_samples_per_second": 7.424,
      "eval_steps_per_second": 7.424,
      "step": 250
    },
    {
      "epoch": 0.4001539053482108,
      "grad_norm": 0.3613339364528656,
      "learning_rate": 8e-05,
      "loss": 0.0524,
      "step": 260
    },
    {
      "epoch": 0.41554444016929587,
      "grad_norm": 0.8048490881919861,
      "learning_rate": 8.307692307692309e-05,
      "loss": 0.0514,
      "step": 270
    },
    {
      "epoch": 0.4309349749903809,
      "grad_norm": 0.5144039988517761,
      "learning_rate": 8.615384615384617e-05,
      "loss": 0.0577,
      "step": 280
    },
    {
      "epoch": 0.44632550981146596,
      "grad_norm": 0.34227821230888367,
      "learning_rate": 8.923076923076924e-05,
      "loss": 0.0621,
      "step": 290
    },
    {
      "epoch": 0.461716044632551,
      "grad_norm": 0.42484498023986816,
      "learning_rate": 9.230769230769232e-05,
      "loss": 0.0522,
      "step": 300
    },
    {
      "epoch": 0.461716044632551,
      "eval_loss": 0.05023057386279106,
      "eval_runtime": 155.809,
      "eval_samples_per_second": 7.419,
      "eval_steps_per_second": 7.419,
      "step": 300
    },
    {
      "epoch": 0.477106579453636,
      "grad_norm": 0.4658105671405792,
      "learning_rate": 9.53846153846154e-05,
      "loss": 0.0577,
      "step": 310
    },
    {
      "epoch": 0.49249711427472104,
      "grad_norm": 1.1311237812042236,
      "learning_rate": 9.846153846153848e-05,
      "loss": 0.0621,
      "step": 320
    },
    {
      "epoch": 0.5078876490958061,
      "grad_norm": 0.3455394506454468,
      "learning_rate": 9.999927654251793e-05,
      "loss": 0.058,
      "step": 330
    },
    {
      "epoch": 0.5232781839168911,
      "grad_norm": 0.4553656280040741,
      "learning_rate": 9.99934890082745e-05,
      "loss": 0.0536,
      "step": 340
    },
    {
      "epoch": 0.5386687187379762,
      "grad_norm": 0.39271876215934753,
      "learning_rate": 9.99819146097084e-05,
      "loss": 0.0464,
      "step": 350
    },
    {
      "epoch": 0.5386687187379762,
      "eval_loss": 0.04761577770113945,
      "eval_runtime": 156.0615,
      "eval_samples_per_second": 7.407,
      "eval_steps_per_second": 7.407,
      "step": 350
    },
    {
      "epoch": 0.5540592535590612,
      "grad_norm": 0.28057408332824707,
      "learning_rate": 9.996455468658355e-05,
      "loss": 0.0592,
      "step": 360
    },
    {
      "epoch": 0.5694497883801463,
      "grad_norm": 0.4577235281467438,
      "learning_rate": 9.994141124835203e-05,
      "loss": 0.0434,
      "step": 370
    },
    {
      "epoch": 0.5848403232012312,
      "grad_norm": 0.43431663513183594,
      "learning_rate": 9.991248697392142e-05,
      "loss": 0.0464,
      "step": 380
    },
    {
      "epoch": 0.6002308580223162,
      "grad_norm": 0.5893954038619995,
      "learning_rate": 9.987778521134476e-05,
      "loss": 0.0429,
      "step": 390
    },
    {
      "epoch": 0.6156213928434013,
      "grad_norm": 0.24359793961048126,
      "learning_rate": 9.983730997743294e-05,
      "loss": 0.0625,
      "step": 400
    },
    {
      "epoch": 0.6156213928434013,
      "eval_loss": 0.04480620101094246,
      "eval_runtime": 155.9589,
      "eval_samples_per_second": 7.412,
      "eval_steps_per_second": 7.412,
      "step": 400
    },
    {
      "epoch": 0.6310119276644863,
      "grad_norm": 0.18097874522209167,
      "learning_rate": 9.979106595728978e-05,
      "loss": 0.0427,
      "step": 410
    },
    {
      "epoch": 0.6464024624855714,
      "grad_norm": 0.47393399477005005,
      "learning_rate": 9.973905850376977e-05,
      "loss": 0.0546,
      "step": 420
    },
    {
      "epoch": 0.6617929973066564,
      "grad_norm": 0.5652933120727539,
      "learning_rate": 9.968129363685833e-05,
      "loss": 0.047,
      "step": 430
    },
    {
      "epoch": 0.6771835321277414,
      "grad_norm": 0.3586273193359375,
      "learning_rate": 9.961777804297514e-05,
      "loss": 0.0611,
      "step": 440
    },
    {
      "epoch": 0.6925740669488265,
      "grad_norm": 0.29289236664772034,
      "learning_rate": 9.954851907420009e-05,
      "loss": 0.0348,
      "step": 450
    },
    {
      "epoch": 0.6925740669488265,
      "eval_loss": 0.0438113771378994,
      "eval_runtime": 155.9991,
      "eval_samples_per_second": 7.41,
      "eval_steps_per_second": 7.41,
      "step": 450
    },
    {
      "epoch": 0.7079646017699115,
      "grad_norm": 0.344553679227829,
      "learning_rate": 9.947352474742222e-05,
      "loss": 0.0362,
      "step": 460
    },
    {
      "epoch": 0.7233551365909965,
      "grad_norm": 0.35288894176483154,
      "learning_rate": 9.93928037434118e-05,
      "loss": 0.0295,
      "step": 470
    },
    {
      "epoch": 0.7387456714120816,
      "grad_norm": 0.5358998775482178,
      "learning_rate": 9.930636540581556e-05,
      "loss": 0.0356,
      "step": 480
    },
    {
      "epoch": 0.7541362062331666,
      "grad_norm": 0.2639545202255249,
      "learning_rate": 9.921421974007506e-05,
      "loss": 0.0392,
      "step": 490
    },
    {
      "epoch": 0.7695267410542517,
      "grad_norm": 0.2737574577331543,
      "learning_rate": 9.911637741226849e-05,
      "loss": 0.0292,
      "step": 500
    },
    {
      "epoch": 0.7695267410542517,
      "eval_loss": 0.040529005229473114,
      "eval_runtime": 156.0231,
      "eval_samples_per_second": 7.409,
      "eval_steps_per_second": 7.409,
      "step": 500
    },
    {
      "epoch": 0.7849172758753367,
      "grad_norm": 0.5032161474227905,
      "learning_rate": 9.90128497478762e-05,
      "loss": 0.0311,
      "step": 510
    },
    {
      "epoch": 0.8003078106964217,
      "grad_norm": 0.20168349146842957,
      "learning_rate": 9.890364873046965e-05,
      "loss": 0.0495,
      "step": 520
    },
    {
      "epoch": 0.8156983455175068,
      "grad_norm": 0.43058785796165466,
      "learning_rate": 9.878878700032427e-05,
      "loss": 0.063,
      "step": 530
    },
    {
      "epoch": 0.8310888803385917,
      "grad_norm": 0.3579358160495758,
      "learning_rate": 9.866827785295638e-05,
      "loss": 0.0421,
      "step": 540
    },
    {
      "epoch": 0.8464794151596768,
      "grad_norm": 0.2530227303504944,
      "learning_rate": 9.854213523758413e-05,
      "loss": 0.0472,
      "step": 550
    },
    {
      "epoch": 0.8464794151596768,
      "eval_loss": 0.03791683539748192,
      "eval_runtime": 156.087,
      "eval_samples_per_second": 7.406,
      "eval_steps_per_second": 7.406,
      "step": 550
    },
    {
      "epoch": 0.8618699499807618,
      "grad_norm": 0.20525501668453217,
      "learning_rate": 9.841037375551294e-05,
      "loss": 0.0316,
      "step": 560
    },
    {
      "epoch": 0.8772604848018468,
      "grad_norm": 0.3436931371688843,
      "learning_rate": 9.827300865844527e-05,
      "loss": 0.0315,
      "step": 570
    },
    {
      "epoch": 0.8926510196229319,
      "grad_norm": 0.34183600544929504,
      "learning_rate": 9.813005584671522e-05,
      "loss": 0.0374,
      "step": 580
    },
    {
      "epoch": 0.9080415544440169,
      "grad_norm": 0.2613978087902069,
      "learning_rate": 9.79815318674481e-05,
      "loss": 0.0536,
      "step": 590
    },
    {
      "epoch": 0.923432089265102,
      "grad_norm": 0.3309175968170166,
      "learning_rate": 9.782745391264498e-05,
      "loss": 0.0367,
      "step": 600
    },
    {
      "epoch": 0.923432089265102,
      "eval_loss": 0.03722962364554405,
      "eval_runtime": 156.0122,
      "eval_samples_per_second": 7.41,
      "eval_steps_per_second": 7.41,
      "step": 600
    },
    {
      "epoch": 0.938822624086187,
      "grad_norm": 0.25939252972602844,
      "learning_rate": 9.76678398171927e-05,
      "loss": 0.036,
      "step": 610
    },
    {
      "epoch": 0.954213158907272,
      "grad_norm": 0.19500210881233215,
      "learning_rate": 9.750270805679945e-05,
      "loss": 0.0398,
      "step": 620
    },
    {
      "epoch": 0.9696036937283571,
      "grad_norm": 0.39204320311546326,
      "learning_rate": 9.733207774585618e-05,
      "loss": 0.0447,
      "step": 630
    },
    {
      "epoch": 0.9849942285494421,
      "grad_norm": 0.8868380784988403,
      "learning_rate": 9.715596863522398e-05,
      "loss": 0.0362,
      "step": 640
    },
    {
      "epoch": 1.0003847633705272,
      "grad_norm": 0.7710979580879211,
      "learning_rate": 9.697440110994801e-05,
      "loss": 0.0309,
      "step": 650
    },
    {
      "epoch": 1.0003847633705272,
      "eval_loss": 0.036069341003894806,
      "eval_runtime": 155.9667,
      "eval_samples_per_second": 7.412,
      "eval_steps_per_second": 7.412,
      "step": 650
    },
    {
      "epoch": 1.0157752981916122,
      "grad_norm": 0.414287805557251,
      "learning_rate": 9.678739618689775e-05,
      "loss": 0.0368,
      "step": 660
    },
    {
      "epoch": 1.0311658330126972,
      "grad_norm": 0.5874229669570923,
      "learning_rate": 9.659497551233431e-05,
      "loss": 0.0362,
      "step": 670
    },
    {
      "epoch": 1.0465563678337821,
      "grad_norm": 0.278563529253006,
      "learning_rate": 9.639716135940485e-05,
      "loss": 0.0343,
      "step": 680
    },
    {
      "epoch": 1.0619469026548674,
      "grad_norm": 0.4759167432785034,
      "learning_rate": 9.619397662556435e-05,
      "loss": 0.048,
      "step": 690
    },
    {
      "epoch": 1.0773374374759523,
      "grad_norm": 0.2320932000875473,
      "learning_rate": 9.598544482992517e-05,
      "loss": 0.0362,
      "step": 700
    },
    {
      "epoch": 1.0773374374759523,
      "eval_loss": 0.03848585486412048,
      "eval_runtime": 155.7909,
      "eval_samples_per_second": 7.42,
      "eval_steps_per_second": 7.42,
      "step": 700
    },
    {
      "epoch": 1.0927279722970373,
      "grad_norm": 0.2700707018375397,
      "learning_rate": 9.57715901105348e-05,
      "loss": 0.0318,
      "step": 710
    },
    {
      "epoch": 1.1081185071181223,
      "grad_norm": 0.18170036375522614,
      "learning_rate": 9.555243722158158e-05,
      "loss": 0.0262,
      "step": 720
    },
    {
      "epoch": 1.1235090419392073,
      "grad_norm": 0.27000892162323,
      "learning_rate": 9.532801153052958e-05,
      "loss": 0.0323,
      "step": 730
    },
    {
      "epoch": 1.1388995767602923,
      "grad_norm": 0.5233851671218872,
      "learning_rate": 9.509833901518209e-05,
      "loss": 0.039,
      "step": 740
    },
    {
      "epoch": 1.1542901115813775,
      "grad_norm": 0.22729772329330444,
      "learning_rate": 9.486344626067475e-05,
      "loss": 0.0293,
      "step": 750
    },
    {
      "epoch": 1.1542901115813775,
      "eval_loss": 0.035639576613903046,
      "eval_runtime": 156.0565,
      "eval_samples_per_second": 7.408,
      "eval_steps_per_second": 7.408,
      "step": 750
    },
    {
      "epoch": 1.1696806464024625,
      "grad_norm": 0.23746898770332336,
      "learning_rate": 9.462336045639815e-05,
      "loss": 0.0381,
      "step": 760
    },
    {
      "epoch": 1.1850711812235475,
      "grad_norm": 0.27130433917045593,
      "learning_rate": 9.437810939285068e-05,
      "loss": 0.0392,
      "step": 770
    },
    {
      "epoch": 1.2004617160446325,
      "grad_norm": 0.21095439791679382,
      "learning_rate": 9.412772145842165e-05,
      "loss": 0.033,
      "step": 780
    },
    {
      "epoch": 1.2158522508657175,
      "grad_norm": 0.2957763671875,
      "learning_rate": 9.387222563610535e-05,
      "loss": 0.0341,
      "step": 790
    },
    {
      "epoch": 1.2312427856868027,
      "grad_norm": 0.2919231653213501,
      "learning_rate": 9.361165150014606e-05,
      "loss": 0.0309,
      "step": 800
    },
    {
      "epoch": 1.2312427856868027,
      "eval_loss": 0.035071104764938354,
      "eval_runtime": 156.0582,
      "eval_samples_per_second": 7.407,
      "eval_steps_per_second": 7.407,
      "step": 800
    },
    {
      "epoch": 1.2466333205078877,
      "grad_norm": 0.16284899413585663,
      "learning_rate": 9.334602921261492e-05,
      "loss": 0.0347,
      "step": 810
    },
    {
      "epoch": 1.2620238553289727,
      "grad_norm": 0.24403131008148193,
      "learning_rate": 9.307538951991852e-05,
      "loss": 0.0267,
      "step": 820
    },
    {
      "epoch": 1.2774143901500576,
      "grad_norm": 0.3003334701061249,
      "learning_rate": 9.279976374923988e-05,
      "loss": 0.031,
      "step": 830
    },
    {
      "epoch": 1.2928049249711426,
      "grad_norm": 0.4727235436439514,
      "learning_rate": 9.251918380491238e-05,
      "loss": 0.0296,
      "step": 840
    },
    {
      "epoch": 1.3081954597922278,
      "grad_norm": 0.5926356911659241,
      "learning_rate": 9.223368216472668e-05,
      "loss": 0.0341,
      "step": 850
    },
    {
      "epoch": 1.3081954597922278,
      "eval_loss": 0.035179633647203445,
      "eval_runtime": 155.7495,
      "eval_samples_per_second": 7.422,
      "eval_steps_per_second": 7.422,
      "step": 850
    },
    {
      "epoch": 1.3235859946133128,
      "grad_norm": 0.1942935287952423,
      "learning_rate": 9.194329187617127e-05,
      "loss": 0.0338,
      "step": 860
    },
    {
      "epoch": 1.3389765294343978,
      "grad_norm": 0.37602418661117554,
      "learning_rate": 9.164804655260735e-05,
      "loss": 0.0305,
      "step": 870
    },
    {
      "epoch": 1.3543670642554828,
      "grad_norm": 0.13682475686073303,
      "learning_rate": 9.134798036937778e-05,
      "loss": 0.027,
      "step": 880
    },
    {
      "epoch": 1.3697575990765678,
      "grad_norm": 0.18598909676074982,
      "learning_rate": 9.104312805985134e-05,
      "loss": 0.0356,
      "step": 890
    },
    {
      "epoch": 1.385148133897653,
      "grad_norm": 0.2571493983268738,
      "learning_rate": 9.073352491140221e-05,
      "loss": 0.0211,
      "step": 900
    },
    {
      "epoch": 1.385148133897653,
      "eval_loss": 0.03406280279159546,
      "eval_runtime": 156.013,
      "eval_samples_per_second": 7.41,
      "eval_steps_per_second": 7.41,
      "step": 900
    },
    {
      "epoch": 1.400538668718738,
      "grad_norm": 0.21523404121398926,
      "learning_rate": 9.041920676132543e-05,
      "loss": 0.03,
      "step": 910
    },
    {
      "epoch": 1.415929203539823,
      "grad_norm": 0.2555162310600281,
      "learning_rate": 9.010020999268854e-05,
      "loss": 0.0388,
      "step": 920
    },
    {
      "epoch": 1.431319738360908,
      "grad_norm": 0.239962637424469,
      "learning_rate": 8.977657153012027e-05,
      "loss": 0.0275,
      "step": 930
    },
    {
      "epoch": 1.446710273181993,
      "grad_norm": 0.36492571234703064,
      "learning_rate": 8.944832883553633e-05,
      "loss": 0.0358,
      "step": 940
    },
    {
      "epoch": 1.4621008080030782,
      "grad_norm": 0.09042529761791229,
      "learning_rate": 8.91155199038032e-05,
      "loss": 0.0405,
      "step": 950
    },
    {
      "epoch": 1.4621008080030782,
      "eval_loss": 0.034839894622564316,
      "eval_runtime": 156.0152,
      "eval_samples_per_second": 7.41,
      "eval_steps_per_second": 7.41,
      "step": 950
    },
    {
      "epoch": 1.4774913428241632,
      "grad_norm": 0.3277328610420227,
      "learning_rate": 8.877818325834006e-05,
      "loss": 0.0297,
      "step": 960
    },
    {
      "epoch": 1.4928818776452482,
      "grad_norm": 0.09511667490005493,
      "learning_rate": 8.843635794665961e-05,
      "loss": 0.0246,
      "step": 970
    },
    {
      "epoch": 1.5082724124663334,
      "grad_norm": 0.2817135453224182,
      "learning_rate": 8.809008353584829e-05,
      "loss": 0.0353,
      "step": 980
    },
    {
      "epoch": 1.5236629472874181,
      "grad_norm": 0.18087555468082428,
      "learning_rate": 8.773940010798627e-05,
      "loss": 0.0339,
      "step": 990
    },
    {
      "epoch": 1.5390534821085033,
      "grad_norm": 0.2365584522485733,
      "learning_rate": 8.738434825550785e-05,
      "loss": 0.0325,
      "step": 1000
    },
    {
      "epoch": 1.5390534821085033,
      "eval_loss": 0.03327273949980736,
      "eval_runtime": 156.0055,
      "eval_samples_per_second": 7.41,
      "eval_steps_per_second": 7.41,
      "step": 1000
    },
    {
      "epoch": 1.5544440169295883,
      "grad_norm": 0.22156043350696564,
      "learning_rate": 8.702496907650274e-05,
      "loss": 0.0356,
      "step": 1010
    },
    {
      "epoch": 1.5698345517506733,
      "grad_norm": 0.2572099566459656,
      "learning_rate": 8.666130416995897e-05,
      "loss": 0.0401,
      "step": 1020
    },
    {
      "epoch": 1.5852250865717585,
      "grad_norm": 0.14971685409545898,
      "learning_rate": 8.629339563094758e-05,
      "loss": 0.0318,
      "step": 1030
    },
    {
      "epoch": 1.6006156213928433,
      "grad_norm": 0.13614961504936218,
      "learning_rate": 8.592128604575014e-05,
      "loss": 0.0323,
      "step": 1040
    },
    {
      "epoch": 1.6160061562139285,
      "grad_norm": 0.23914273083209991,
      "learning_rate": 8.554501848692921e-05,
      "loss": 0.0335,
      "step": 1050
    },
    {
      "epoch": 1.6160061562139285,
      "eval_loss": 0.03335578367114067,
      "eval_runtime": 155.8248,
      "eval_samples_per_second": 7.419,
      "eval_steps_per_second": 7.419,
      "step": 1050
    },
    {
      "epoch": 1.6313966910350135,
      "grad_norm": 0.19318076968193054,
      "learning_rate": 8.51646365083426e-05,
      "loss": 0.0271,
      "step": 1060
    },
    {
      "epoch": 1.6467872258560985,
      "grad_norm": 0.16463421285152435,
      "learning_rate": 8.478018414010191e-05,
      "loss": 0.0293,
      "step": 1070
    },
    {
      "epoch": 1.6621777606771835,
      "grad_norm": 0.15071162581443787,
      "learning_rate": 8.439170588347598e-05,
      "loss": 0.0252,
      "step": 1080
    },
    {
      "epoch": 1.6775682954982685,
      "grad_norm": 0.5070039629936218,
      "learning_rate": 8.399924670573969e-05,
      "loss": 0.0308,
      "step": 1090
    },
    {
      "epoch": 1.6929588303193537,
      "grad_norm": 0.17456522583961487,
      "learning_rate": 8.360285203496894e-05,
      "loss": 0.0318,
      "step": 1100
    },
    {
      "epoch": 1.6929588303193537,
      "eval_loss": 0.03446831926703453,
      "eval_runtime": 155.9663,
      "eval_samples_per_second": 7.412,
      "eval_steps_per_second": 7.412,
      "step": 1100
    },
    {
      "epoch": 1.7083493651404387,
      "grad_norm": 0.49715712666511536,
      "learning_rate": 8.320256775478228e-05,
      "loss": 0.037,
      "step": 1110
    },
    {
      "epoch": 1.7237398999615237,
      "grad_norm": 0.24198360741138458,
      "learning_rate": 8.279844019902968e-05,
      "loss": 0.0261,
      "step": 1120
    },
    {
      "epoch": 1.7391304347826086,
      "grad_norm": 0.2602682411670685,
      "learning_rate": 8.239051614642934e-05,
      "loss": 0.0399,
      "step": 1130
    },
    {
      "epoch": 1.7545209696036936,
      "grad_norm": 0.16040796041488647,
      "learning_rate": 8.197884281515292e-05,
      "loss": 0.0479,
      "step": 1140
    },
    {
      "epoch": 1.7699115044247788,
      "grad_norm": 0.2572268843650818,
      "learning_rate": 8.156346785735997e-05,
      "loss": 0.0205,
      "step": 1150
    },
    {
      "epoch": 1.7699115044247788,
      "eval_loss": 0.03324705362319946,
      "eval_runtime": 155.7309,
      "eval_samples_per_second": 7.423,
      "eval_steps_per_second": 7.423,
      "step": 1150
    },
    {
      "epoch": 1.7853020392458638,
      "grad_norm": 0.17412768304347992,
      "learning_rate": 8.114443935368198e-05,
      "loss": 0.0302,
      "step": 1160
    },
    {
      "epoch": 1.8006925740669488,
      "grad_norm": 0.15181364119052887,
      "learning_rate": 8.072180580765709e-05,
      "loss": 0.0237,
      "step": 1170
    },
    {
      "epoch": 1.8160831088880338,
      "grad_norm": 0.21758894622325897,
      "learning_rate": 8.02956161401155e-05,
      "loss": 0.0242,
      "step": 1180
    },
    {
      "epoch": 1.8314736437091188,
      "grad_norm": 0.3291786313056946,
      "learning_rate": 7.98659196835169e-05,
      "loss": 0.0378,
      "step": 1190
    },
    {
      "epoch": 1.846864178530204,
      "grad_norm": 0.09492679685354233,
      "learning_rate": 7.943276617624014e-05,
      "loss": 0.0218,
      "step": 1200
    },
    {
      "epoch": 1.846864178530204,
      "eval_loss": 0.034512847661972046,
      "eval_runtime": 155.933,
      "eval_samples_per_second": 7.413,
      "eval_steps_per_second": 7.413,
      "step": 1200
    },
    {
      "epoch": 1.862254713351289,
      "grad_norm": 0.258452832698822,
      "learning_rate": 7.899620575682578e-05,
      "loss": 0.0298,
      "step": 1210
    },
    {
      "epoch": 1.877645248172374,
      "grad_norm": 0.15415191650390625,
      "learning_rate": 7.85562889581725e-05,
      "loss": 0.026,
      "step": 1220
    },
    {
      "epoch": 1.893035782993459,
      "grad_norm": 0.10892465710639954,
      "learning_rate": 7.811306670168788e-05,
      "loss": 0.0302,
      "step": 1230
    },
    {
      "epoch": 1.908426317814544,
      "grad_norm": 0.2263966053724289,
      "learning_rate": 7.766659029139393e-05,
      "loss": 0.0353,
      "step": 1240
    },
    {
      "epoch": 1.9238168526356292,
      "grad_norm": 0.34887775778770447,
      "learning_rate": 7.72169114079887e-05,
      "loss": 0.0305,
      "step": 1250
    },
    {
      "epoch": 1.9238168526356292,
      "eval_loss": 0.03335217013955116,
      "eval_runtime": 156.0018,
      "eval_samples_per_second": 7.41,
      "eval_steps_per_second": 7.41,
      "step": 1250
    },
    {
      "epoch": 1.9392073874567142,
      "grad_norm": 0.3543195426464081,
      "learning_rate": 7.676408210286407e-05,
      "loss": 0.0319,
      "step": 1260
    },
    {
      "epoch": 1.9545979222777992,
      "grad_norm": 0.20112381875514984,
      "learning_rate": 7.630815479208066e-05,
      "loss": 0.028,
      "step": 1270
    },
    {
      "epoch": 1.9699884570988841,
      "grad_norm": 0.15991556644439697,
      "learning_rate": 7.584918225030053e-05,
      "loss": 0.0249,
      "step": 1280
    },
    {
      "epoch": 1.9853789919199691,
      "grad_norm": 0.21419313549995422,
      "learning_rate": 7.538721760467845e-05,
      "loss": 0.0317,
      "step": 1290
    },
    {
      "epoch": 2.0007695267410543,
      "grad_norm": 0.22169247269630432,
      "learning_rate": 7.49223143287122e-05,
      "loss": 0.0225,
      "step": 1300
    },
    {
      "epoch": 2.0007695267410543,
      "eval_loss": 0.03226001188158989,
      "eval_runtime": 155.7712,
      "eval_samples_per_second": 7.421,
      "eval_steps_per_second": 7.421,
      "step": 1300
    },
    {
      "epoch": 2.016160061562139,
      "grad_norm": 0.25050830841064453,
      "learning_rate": 7.445452623605307e-05,
      "loss": 0.0215,
      "step": 1310
    },
    {
      "epoch": 2.0315505963832243,
      "grad_norm": 0.2710731327533722,
      "learning_rate": 7.398390747427662e-05,
      "loss": 0.0229,
      "step": 1320
    },
    {
      "epoch": 2.0469411312043095,
      "grad_norm": 0.13344928622245789,
      "learning_rate": 7.351051251861501e-05,
      "loss": 0.0228,
      "step": 1330
    },
    {
      "epoch": 2.0623316660253943,
      "grad_norm": 0.13948418200016022,
      "learning_rate": 7.303439616565146e-05,
      "loss": 0.0207,
      "step": 1340
    },
    {
      "epoch": 2.0777222008464795,
      "grad_norm": 0.24005165696144104,
      "learning_rate": 7.255561352697732e-05,
      "loss": 0.0166,
      "step": 1350
    },
    {
      "epoch": 2.0777222008464795,
      "eval_loss": 0.03310776874423027,
      "eval_runtime": 155.7842,
      "eval_samples_per_second": 7.421,
      "eval_steps_per_second": 7.421,
      "step": 1350
    },
    {
      "epoch": 2.0931127356675643,
      "grad_norm": 0.12693017721176147,
      "learning_rate": 7.207422002281281e-05,
      "loss": 0.0206,
      "step": 1360
    },
    {
      "epoch": 2.1085032704886495,
      "grad_norm": 0.5274406671524048,
      "learning_rate": 7.159027137559197e-05,
      "loss": 0.0173,
      "step": 1370
    },
    {
      "epoch": 2.1238938053097347,
      "grad_norm": 0.34713730216026306,
      "learning_rate": 7.110382360351267e-05,
      "loss": 0.0135,
      "step": 1380
    },
    {
      "epoch": 2.1392843401308195,
      "grad_norm": 0.2378131002187729,
      "learning_rate": 7.061493301405244e-05,
      "loss": 0.0165,
      "step": 1390
    },
    {
      "epoch": 2.1546748749519047,
      "grad_norm": 0.6217334866523743,
      "learning_rate": 7.012365619745064e-05,
      "loss": 0.0325,
      "step": 1400
    },
    {
      "epoch": 2.1546748749519047,
      "eval_loss": 0.0352943055331707,
      "eval_runtime": 155.8301,
      "eval_samples_per_second": 7.418,
      "eval_steps_per_second": 7.418,
      "step": 1400
    },
    {
      "epoch": 2.1700654097729895,
      "grad_norm": 0.08532413840293884,
      "learning_rate": 6.963005002015808e-05,
      "loss": 0.0154,
      "step": 1410
    },
    {
      "epoch": 2.1854559445940747,
      "grad_norm": 0.30685949325561523,
      "learning_rate": 6.91341716182545e-05,
      "loss": 0.0242,
      "step": 1420
    },
    {
      "epoch": 2.20084647941516,
      "grad_norm": 0.2981208562850952,
      "learning_rate": 6.863607839083507e-05,
      "loss": 0.0152,
      "step": 1430
    },
    {
      "epoch": 2.2162370142362446,
      "grad_norm": 0.19159533083438873,
      "learning_rate": 6.813582799336624e-05,
      "loss": 0.021,
      "step": 1440
    },
    {
      "epoch": 2.23162754905733,
      "grad_norm": 0.10919353365898132,
      "learning_rate": 6.763347833101192e-05,
      "loss": 0.0127,
      "step": 1450
    },
    {
      "epoch": 2.23162754905733,
      "eval_loss": 0.036012399941682816,
      "eval_runtime": 155.8346,
      "eval_samples_per_second": 7.418,
      "eval_steps_per_second": 7.418,
      "step": 1450
    },
    {
      "epoch": 2.2470180838784146,
      "grad_norm": 0.20541125535964966,
      "learning_rate": 6.712908755193094e-05,
      "loss": 0.0179,
      "step": 1460
    },
    {
      "epoch": 2.2624086186995,
      "grad_norm": 0.4262455105781555,
      "learning_rate": 6.662271404054621e-05,
      "loss": 0.0231,
      "step": 1470
    },
    {
      "epoch": 2.2777991535205846,
      "grad_norm": 0.254803866147995,
      "learning_rate": 6.611441641078657e-05,
      "loss": 0.0241,
      "step": 1480
    },
    {
      "epoch": 2.29318968834167,
      "grad_norm": 0.12292696535587311,
      "learning_rate": 6.560425349930208e-05,
      "loss": 0.0244,
      "step": 1490
    },
    {
      "epoch": 2.308580223162755,
      "grad_norm": 0.2538405656814575,
      "learning_rate": 6.509228435865357e-05,
      "loss": 0.0203,
      "step": 1500
    },
    {
      "epoch": 2.308580223162755,
      "eval_loss": 0.03317645564675331,
      "eval_runtime": 155.7695,
      "eval_samples_per_second": 7.421,
      "eval_steps_per_second": 7.421,
      "step": 1500
    },
    {
      "epoch": 2.32397075798384,
      "grad_norm": 0.31867679953575134,
      "learning_rate": 6.45785682504772e-05,
      "loss": 0.0223,
      "step": 1510
    },
    {
      "epoch": 2.339361292804925,
      "grad_norm": 0.1164252907037735,
      "learning_rate": 6.406316463862467e-05,
      "loss": 0.0204,
      "step": 1520
    },
    {
      "epoch": 2.3547518276260098,
      "grad_norm": 0.2475312501192093,
      "learning_rate": 6.35461331822803e-05,
      "loss": 0.0194,
      "step": 1530
    },
    {
      "epoch": 2.370142362447095,
      "grad_norm": 0.3384310007095337,
      "learning_rate": 6.302753372905515e-05,
      "loss": 0.0184,
      "step": 1540
    },
    {
      "epoch": 2.38553289726818,
      "grad_norm": 0.20339253544807434,
      "learning_rate": 6.250742630805971e-05,
      "loss": 0.0152,
      "step": 1550
    },
    {
      "epoch": 2.38553289726818,
      "eval_loss": 0.034951258450746536,
      "eval_runtime": 155.7631,
      "eval_samples_per_second": 7.422,
      "eval_steps_per_second": 7.422,
      "step": 1550
    },
    {
      "epoch": 2.400923432089265,
      "grad_norm": 0.2379731386899948,
      "learning_rate": 6.198587112295526e-05,
      "loss": 0.0123,
      "step": 1560
    },
    {
      "epoch": 2.41631396691035,
      "grad_norm": 0.4188356101512909,
      "learning_rate": 6.14629285449852e-05,
      "loss": 0.0246,
      "step": 1570
    },
    {
      "epoch": 2.431704501731435,
      "grad_norm": 0.13910157978534698,
      "learning_rate": 6.093865910598688e-05,
      "loss": 0.0207,
      "step": 1580
    },
    {
      "epoch": 2.44709503655252,
      "grad_norm": 0.23834119737148285,
      "learning_rate": 6.0413123491385025e-05,
      "loss": 0.0189,
      "step": 1590
    },
    {
      "epoch": 2.4624855713736054,
      "grad_norm": 0.3100840747356415,
      "learning_rate": 5.988638253316713e-05,
      "loss": 0.0149,
      "step": 1600
    },
    {
      "epoch": 2.4624855713736054,
      "eval_loss": 0.0337284654378891,
      "eval_runtime": 155.7862,
      "eval_samples_per_second": 7.42,
      "eval_steps_per_second": 7.42,
      "step": 1600
    },
    {
      "epoch": 2.47787610619469,
      "grad_norm": 0.2668248414993286,
      "learning_rate": 5.935849720284206e-05,
      "loss": 0.0177,
      "step": 1610
    },
    {
      "epoch": 2.4932666410157753,
      "grad_norm": 0.1645742505788803,
      "learning_rate": 5.8829528604382436e-05,
      "loss": 0.0205,
      "step": 1620
    },
    {
      "epoch": 2.50865717583686,
      "grad_norm": 0.13473357260227203,
      "learning_rate": 5.829953796715176e-05,
      "loss": 0.0205,
      "step": 1630
    },
    {
      "epoch": 2.5240477106579453,
      "grad_norm": 0.26183220744132996,
      "learning_rate": 5.7768586638816924e-05,
      "loss": 0.0209,
      "step": 1640
    },
    {
      "epoch": 2.5394382454790305,
      "grad_norm": 0.25971588492393494,
      "learning_rate": 5.723673607824705e-05,
      "loss": 0.0235,
      "step": 1650
    },
    {
      "epoch": 2.5394382454790305,
      "eval_loss": 0.0318160355091095,
      "eval_runtime": 155.797,
      "eval_samples_per_second": 7.42,
      "eval_steps_per_second": 7.42,
      "step": 1650
    },
    {
      "epoch": 2.5548287803001153,
      "grad_norm": 0.4008631408214569,
      "learning_rate": 5.670404784839953e-05,
      "loss": 0.0175,
      "step": 1660
    },
    {
      "epoch": 2.5702193151212005,
      "grad_norm": 0.33550024032592773,
      "learning_rate": 5.6170583609193984e-05,
      "loss": 0.018,
      "step": 1670
    },
    {
      "epoch": 2.5856098499422853,
      "grad_norm": 0.2780638337135315,
      "learning_rate": 5.56364051103749e-05,
      "loss": 0.0241,
      "step": 1680
    },
    {
      "epoch": 2.6010003847633705,
      "grad_norm": 0.3534189462661743,
      "learning_rate": 5.5101574184364056e-05,
      "loss": 0.022,
      "step": 1690
    },
    {
      "epoch": 2.6163909195844557,
      "grad_norm": 0.2706807553768158,
      "learning_rate": 5.4566152739103116e-05,
      "loss": 0.0178,
      "step": 1700
    },
    {
      "epoch": 2.6163909195844557,
      "eval_loss": 0.03322064131498337,
      "eval_runtime": 155.8094,
      "eval_samples_per_second": 7.419,
      "eval_steps_per_second": 7.419,
      "step": 1700
    },
    {
      "epoch": 2.6317814544055405,
      "grad_norm": 0.24887484312057495,
      "learning_rate": 5.403020275088786e-05,
      "loss": 0.0194,
      "step": 1710
    },
    {
      "epoch": 2.6471719892266257,
      "grad_norm": 0.19714955985546112,
      "learning_rate": 5.349378625719408e-05,
      "loss": 0.0301,
      "step": 1720
    },
    {
      "epoch": 2.6625625240477104,
      "grad_norm": 0.39319732785224915,
      "learning_rate": 5.2956965349496743e-05,
      "loss": 0.023,
      "step": 1730
    },
    {
      "epoch": 2.6779530588687956,
      "grad_norm": 0.14886178076267242,
      "learning_rate": 5.241980216608261e-05,
      "loss": 0.0166,
      "step": 1740
    },
    {
      "epoch": 2.693343593689881,
      "grad_norm": 0.1679171472787857,
      "learning_rate": 5.188235888485773e-05,
      "loss": 0.027,
      "step": 1750
    },
    {
      "epoch": 2.693343593689881,
      "eval_loss": 0.032434236258268356,
      "eval_runtime": 156.066,
      "eval_samples_per_second": 7.407,
      "eval_steps_per_second": 7.407,
      "step": 1750
    },
    {
      "epoch": 2.7087341285109656,
      "grad_norm": 0.32193097472190857,
      "learning_rate": 5.13446977161501e-05,
      "loss": 0.0157,
      "step": 1760
    },
    {
      "epoch": 2.724124663332051,
      "grad_norm": 0.3438890278339386,
      "learning_rate": 5.0806880895508766e-05,
      "loss": 0.0208,
      "step": 1770
    },
    {
      "epoch": 2.7395151981531356,
      "grad_norm": 0.35433468222618103,
      "learning_rate": 5.026897067649976e-05,
      "loss": 0.0257,
      "step": 1780
    },
    {
      "epoch": 2.754905732974221,
      "grad_norm": 0.08436018228530884,
      "learning_rate": 4.973102932350026e-05,
      "loss": 0.0182,
      "step": 1790
    },
    {
      "epoch": 2.770296267795306,
      "grad_norm": 0.27937933802604675,
      "learning_rate": 4.9193119104491245e-05,
      "loss": 0.0246,
      "step": 1800
    },
    {
      "epoch": 2.770296267795306,
      "eval_loss": 0.03323720023036003,
      "eval_runtime": 156.0855,
      "eval_samples_per_second": 7.406,
      "eval_steps_per_second": 7.406,
      "step": 1800
    },
    {
      "epoch": 2.785686802616391,
      "grad_norm": 0.22944512963294983,
      "learning_rate": 4.86553022838499e-05,
      "loss": 0.0201,
      "step": 1810
    },
    {
      "epoch": 2.801077337437476,
      "grad_norm": 0.5866400003433228,
      "learning_rate": 4.8117641115142285e-05,
      "loss": 0.0244,
      "step": 1820
    },
    {
      "epoch": 2.8164678722585608,
      "grad_norm": 0.6666045784950256,
      "learning_rate": 4.758019783391741e-05,
      "loss": 0.0248,
      "step": 1830
    },
    {
      "epoch": 2.831858407079646,
      "grad_norm": 0.22676365077495575,
      "learning_rate": 4.704303465050327e-05,
      "loss": 0.0258,
      "step": 1840
    },
    {
      "epoch": 2.847248941900731,
      "grad_norm": 0.21843405067920685,
      "learning_rate": 4.6506213742805924e-05,
      "loss": 0.0232,
      "step": 1850
    },
    {
      "epoch": 2.847248941900731,
      "eval_loss": 0.03171728551387787,
      "eval_runtime": 156.0633,
      "eval_samples_per_second": 7.407,
      "eval_steps_per_second": 7.407,
      "step": 1850
    },
    {
      "epoch": 2.862639476721816,
      "grad_norm": 0.2720729112625122,
      "learning_rate": 4.5969797249112146e-05,
      "loss": 0.0142,
      "step": 1860
    },
    {
      "epoch": 2.878030011542901,
      "grad_norm": 0.27897199988365173,
      "learning_rate": 4.54338472608969e-05,
      "loss": 0.0278,
      "step": 1870
    },
    {
      "epoch": 2.893420546363986,
      "grad_norm": 0.4508466422557831,
      "learning_rate": 4.489842581563597e-05,
      "loss": 0.0126,
      "step": 1880
    },
    {
      "epoch": 2.908811081185071,
      "grad_norm": 0.19606506824493408,
      "learning_rate": 4.436359488962511e-05,
      "loss": 0.0182,
      "step": 1890
    },
    {
      "epoch": 2.9242016160061564,
      "grad_norm": 0.24699777364730835,
      "learning_rate": 4.382941639080602e-05,
      "loss": 0.016,
      "step": 1900
    },
    {
      "epoch": 2.9242016160061564,
      "eval_loss": 0.03240085020661354,
      "eval_runtime": 156.1891,
      "eval_samples_per_second": 7.401,
      "eval_steps_per_second": 7.401,
      "step": 1900
    },
    {
      "epoch": 2.939592150827241,
      "grad_norm": 0.36841627955436707,
      "learning_rate": 4.3295952151600475e-05,
      "loss": 0.0235,
      "step": 1910
    },
    {
      "epoch": 2.9549826856483263,
      "grad_norm": 0.4242415726184845,
      "learning_rate": 4.276326392175296e-05,
      "loss": 0.0201,
      "step": 1920
    },
    {
      "epoch": 2.970373220469411,
      "grad_norm": 0.11797887086868286,
      "learning_rate": 4.2231413361183095e-05,
      "loss": 0.0173,
      "step": 1930
    },
    {
      "epoch": 2.9857637552904963,
      "grad_norm": 0.23386447131633759,
      "learning_rate": 4.1700462032848244e-05,
      "loss": 0.0138,
      "step": 1940
    },
    {
      "epoch": 3.0011542901115815,
      "grad_norm": 0.3257446587085724,
      "learning_rate": 4.117047139561758e-05,
      "loss": 0.0287,
      "step": 1950
    },
    {
      "epoch": 3.0011542901115815,
      "eval_loss": 0.0324116013944149,
      "eval_runtime": 156.1481,
      "eval_samples_per_second": 7.403,
      "eval_steps_per_second": 7.403,
      "step": 1950
    },
    {
      "epoch": 3.0165448249326663,
      "grad_norm": 0.12218138575553894,
      "learning_rate": 4.0641502797157966e-05,
      "loss": 0.0068,
      "step": 1960
    },
    {
      "epoch": 3.0319353597537515,
      "grad_norm": 0.19436976313591003,
      "learning_rate": 4.011361746683288e-05,
      "loss": 0.0106,
      "step": 1970
    },
    {
      "epoch": 3.0473258945748363,
      "grad_norm": 0.23910145461559296,
      "learning_rate": 3.958687650861499e-05,
      "loss": 0.0095,
      "step": 1980
    },
    {
      "epoch": 3.0627164293959215,
      "grad_norm": 0.29927709698677063,
      "learning_rate": 3.906134089401313e-05,
      "loss": 0.0095,
      "step": 1990
    },
    {
      "epoch": 3.0781069642170067,
      "grad_norm": 0.15977710485458374,
      "learning_rate": 3.853707145501482e-05,
      "loss": 0.0108,
      "step": 2000
    },
    {
      "epoch": 3.0781069642170067,
      "eval_loss": 0.03532339632511139,
      "eval_runtime": 155.7516,
      "eval_samples_per_second": 7.422,
      "eval_steps_per_second": 7.422,
      "step": 2000
    },
    {
      "epoch": 3.0934974990380915,
      "grad_norm": 0.5743552446365356,
      "learning_rate": 3.801412887704475e-05,
      "loss": 0.0086,
      "step": 2010
    },
    {
      "epoch": 3.1088880338591767,
      "grad_norm": 0.15589486062526703,
      "learning_rate": 3.749257369194029e-05,
      "loss": 0.0122,
      "step": 2020
    },
    {
      "epoch": 3.1242785686802614,
      "grad_norm": 0.12079951167106628,
      "learning_rate": 3.697246627094487e-05,
      "loss": 0.0113,
      "step": 2030
    },
    {
      "epoch": 3.1396691035013466,
      "grad_norm": 0.08928914368152618,
      "learning_rate": 3.6453866817719726e-05,
      "loss": 0.0081,
      "step": 2040
    },
    {
      "epoch": 3.155059638322432,
      "grad_norm": 0.1578502357006073,
      "learning_rate": 3.593683536137533e-05,
      "loss": 0.0092,
      "step": 2050
    },
    {
      "epoch": 3.155059638322432,
      "eval_loss": 0.03737669810652733,
      "eval_runtime": 156.185,
      "eval_samples_per_second": 7.401,
      "eval_steps_per_second": 7.401,
      "step": 2050
    },
    {
      "epoch": 3.1704501731435166,
      "grad_norm": 0.08487873524427414,
      "learning_rate": 3.542143174952282e-05,
      "loss": 0.0077,
      "step": 2060
    },
    {
      "epoch": 3.185840707964602,
      "grad_norm": 1.4453635215759277,
      "learning_rate": 3.490771564134643e-05,
      "loss": 0.0102,
      "step": 2070
    },
    {
      "epoch": 3.2012312427856866,
      "grad_norm": 0.3025449514389038,
      "learning_rate": 3.4395746500697925e-05,
      "loss": 0.0115,
      "step": 2080
    },
    {
      "epoch": 3.216621777606772,
      "grad_norm": 0.3175906240940094,
      "learning_rate": 3.388558358921343e-05,
      "loss": 0.0091,
      "step": 2090
    },
    {
      "epoch": 3.232012312427857,
      "grad_norm": 0.7257821559906006,
      "learning_rate": 3.337728595945378e-05,
      "loss": 0.0078,
      "step": 2100
    },
    {
      "epoch": 3.232012312427857,
      "eval_loss": 0.03916662558913231,
      "eval_runtime": 155.8879,
      "eval_samples_per_second": 7.416,
      "eval_steps_per_second": 7.416,
      "step": 2100
    },
    {
      "epoch": 3.247402847248942,
      "grad_norm": 0.1518784463405609,
      "learning_rate": 3.287091244806907e-05,
      "loss": 0.0106,
      "step": 2110
    },
    {
      "epoch": 3.262793382070027,
      "grad_norm": 0.2841784954071045,
      "learning_rate": 3.2366521668988106e-05,
      "loss": 0.0116,
      "step": 2120
    },
    {
      "epoch": 3.2781839168911118,
      "grad_norm": 0.4791579842567444,
      "learning_rate": 3.186417200663379e-05,
      "loss": 0.0081,
      "step": 2130
    },
    {
      "epoch": 3.293574451712197,
      "grad_norm": 0.21681740880012512,
      "learning_rate": 3.136392160916495e-05,
      "loss": 0.005,
      "step": 2140
    },
    {
      "epoch": 3.308964986533282,
      "grad_norm": 0.4098879098892212,
      "learning_rate": 3.086582838174551e-05,
      "loss": 0.0101,
      "step": 2150
    },
    {
      "epoch": 3.308964986533282,
      "eval_loss": 0.03898944705724716,
      "eval_runtime": 155.8763,
      "eval_samples_per_second": 7.416,
      "eval_steps_per_second": 7.416,
      "step": 2150
    },
    {
      "epoch": 3.324355521354367,
      "grad_norm": 0.18891297280788422,
      "learning_rate": 3.036994997984194e-05,
      "loss": 0.0097,
      "step": 2160
    },
    {
      "epoch": 3.339746056175452,
      "grad_norm": 0.48369404673576355,
      "learning_rate": 2.9876343802549357e-05,
      "loss": 0.0191,
      "step": 2170
    },
    {
      "epoch": 3.355136590996537,
      "grad_norm": 0.06741505116224289,
      "learning_rate": 2.938506698594755e-05,
      "loss": 0.009,
      "step": 2180
    },
    {
      "epoch": 3.370527125817622,
      "grad_norm": 0.47319892048835754,
      "learning_rate": 2.8896176396487334e-05,
      "loss": 0.0088,
      "step": 2190
    },
    {
      "epoch": 3.3859176606387074,
      "grad_norm": 0.06409157067537308,
      "learning_rate": 2.8409728624408067e-05,
      "loss": 0.0151,
      "step": 2200
    },
    {
      "epoch": 3.3859176606387074,
      "eval_loss": 0.03757644072175026,
      "eval_runtime": 156.0618,
      "eval_samples_per_second": 7.407,
      "eval_steps_per_second": 7.407,
      "step": 2200
    },
    {
      "epoch": 3.401308195459792,
      "grad_norm": 0.1982724368572235,
      "learning_rate": 2.7925779977187215e-05,
      "loss": 0.0088,
      "step": 2210
    },
    {
      "epoch": 3.4166987302808773,
      "grad_norm": 0.28932785987854004,
      "learning_rate": 2.7444386473022686e-05,
      "loss": 0.0085,
      "step": 2220
    },
    {
      "epoch": 3.432089265101962,
      "grad_norm": 0.2503265142440796,
      "learning_rate": 2.696560383434854e-05,
      "loss": 0.0118,
      "step": 2230
    },
    {
      "epoch": 3.4474797999230473,
      "grad_norm": 0.2704777121543884,
      "learning_rate": 2.6489487481384988e-05,
      "loss": 0.0127,
      "step": 2240
    },
    {
      "epoch": 3.4628703347441325,
      "grad_norm": 0.2654518187046051,
      "learning_rate": 2.601609252572339e-05,
      "loss": 0.0091,
      "step": 2250
    },
    {
      "epoch": 3.4628703347441325,
      "eval_loss": 0.039073023945093155,
      "eval_runtime": 155.8311,
      "eval_samples_per_second": 7.418,
      "eval_steps_per_second": 7.418,
      "step": 2250
    },
    {
      "epoch": 3.4782608695652173,
      "grad_norm": 0.1407390534877777,
      "learning_rate": 2.554547376394692e-05,
      "loss": 0.0101,
      "step": 2260
    },
    {
      "epoch": 3.4936514043863025,
      "grad_norm": 0.29681631922721863,
      "learning_rate": 2.5077685671287808e-05,
      "loss": 0.0056,
      "step": 2270
    },
    {
      "epoch": 3.5090419392073873,
      "grad_norm": 0.3207552433013916,
      "learning_rate": 2.4612782395321593e-05,
      "loss": 0.0124,
      "step": 2280
    },
    {
      "epoch": 3.5244324740284725,
      "grad_norm": 0.33818677067756653,
      "learning_rate": 2.4150817749699496e-05,
      "loss": 0.0063,
      "step": 2290
    },
    {
      "epoch": 3.5398230088495577,
      "grad_norm": 0.13887569308280945,
      "learning_rate": 2.3691845207919357e-05,
      "loss": 0.0098,
      "step": 2300
    },
    {
      "epoch": 3.5398230088495577,
      "eval_loss": 0.039753951132297516,
      "eval_runtime": 155.9312,
      "eval_samples_per_second": 7.414,
      "eval_steps_per_second": 7.414,
      "step": 2300
    },
    {
      "epoch": 3.5552135436706425,
      "grad_norm": 0.13252702355384827,
      "learning_rate": 2.3235917897135934e-05,
      "loss": 0.0124,
      "step": 2310
    },
    {
      "epoch": 3.5706040784917277,
      "grad_norm": 0.5990612506866455,
      "learning_rate": 2.2783088592011303e-05,
      "loss": 0.0142,
      "step": 2320
    },
    {
      "epoch": 3.5859946133128124,
      "grad_norm": 0.1325896829366684,
      "learning_rate": 2.233340970860608e-05,
      "loss": 0.0079,
      "step": 2330
    },
    {
      "epoch": 3.6013851481338977,
      "grad_norm": 0.23623527586460114,
      "learning_rate": 2.1886933298312125e-05,
      "loss": 0.0114,
      "step": 2340
    },
    {
      "epoch": 3.616775682954983,
      "grad_norm": 0.13087904453277588,
      "learning_rate": 2.1443711041827503e-05,
      "loss": 0.0048,
      "step": 2350
    },
    {
      "epoch": 3.616775682954983,
      "eval_loss": 0.04036819934844971,
      "eval_runtime": 155.9337,
      "eval_samples_per_second": 7.413,
      "eval_steps_per_second": 7.413,
      "step": 2350
    },
    {
      "epoch": 3.6321662177760676,
      "grad_norm": 0.3286195695400238,
      "learning_rate": 2.1003794243174252e-05,
      "loss": 0.0121,
      "step": 2360
    },
    {
      "epoch": 3.647556752597153,
      "grad_norm": 0.1689368337392807,
      "learning_rate": 2.0567233823759886e-05,
      "loss": 0.0085,
      "step": 2370
    },
    {
      "epoch": 3.6629472874182376,
      "grad_norm": 0.2292981743812561,
      "learning_rate": 2.0134080316483113e-05,
      "loss": 0.0098,
      "step": 2380
    },
    {
      "epoch": 3.678337822239323,
      "grad_norm": 0.31897374987602234,
      "learning_rate": 1.970438385988452e-05,
      "loss": 0.0125,
      "step": 2390
    },
    {
      "epoch": 3.693728357060408,
      "grad_norm": 0.10046721994876862,
      "learning_rate": 1.927819419234292e-05,
      "loss": 0.0127,
      "step": 2400
    },
    {
      "epoch": 3.693728357060408,
      "eval_loss": 0.03930658847093582,
      "eval_runtime": 155.945,
      "eval_samples_per_second": 7.413,
      "eval_steps_per_second": 7.413,
      "step": 2400
    },
    {
      "epoch": 3.709118891881493,
      "grad_norm": 0.45897823572158813,
      "learning_rate": 1.885556064631801e-05,
      "loss": 0.0105,
      "step": 2410
    },
    {
      "epoch": 3.724509426702578,
      "grad_norm": 0.2278304398059845,
      "learning_rate": 1.843653214264003e-05,
      "loss": 0.0072,
      "step": 2420
    },
    {
      "epoch": 3.7398999615236628,
      "grad_norm": 0.29349052906036377,
      "learning_rate": 1.8021157184847087e-05,
      "loss": 0.0093,
      "step": 2430
    },
    {
      "epoch": 3.755290496344748,
      "grad_norm": 0.4859161376953125,
      "learning_rate": 1.7609483853570668e-05,
      "loss": 0.0082,
      "step": 2440
    },
    {
      "epoch": 3.770681031165833,
      "grad_norm": 0.09254691749811172,
      "learning_rate": 1.720155980097034e-05,
      "loss": 0.0062,
      "step": 2450
    },
    {
      "epoch": 3.770681031165833,
      "eval_loss": 0.03990287333726883,
      "eval_runtime": 155.9259,
      "eval_samples_per_second": 7.414,
      "eval_steps_per_second": 7.414,
      "step": 2450
    },
    {
      "epoch": 3.786071565986918,
      "grad_norm": 0.14296738803386688,
      "learning_rate": 1.6797432245217736e-05,
      "loss": 0.0092,
      "step": 2460
    },
    {
      "epoch": 3.801462100808003,
      "grad_norm": 0.15422552824020386,
      "learning_rate": 1.6397147965031072e-05,
      "loss": 0.0043,
      "step": 2470
    },
    {
      "epoch": 3.816852635629088,
      "grad_norm": 0.16916704177856445,
      "learning_rate": 1.600075329426033e-05,
      "loss": 0.0122,
      "step": 2480
    },
    {
      "epoch": 3.832243170450173,
      "grad_norm": 0.2736268937587738,
      "learning_rate": 1.560829411652403e-05,
      "loss": 0.0115,
      "step": 2490
    },
    {
      "epoch": 3.8476337052712584,
      "grad_norm": 0.16690406203269958,
      "learning_rate": 1.521981585989809e-05,
      "loss": 0.01,
      "step": 2500
    },
    {
      "epoch": 3.8476337052712584,
      "eval_loss": 0.040496647357940674,
      "eval_runtime": 156.0387,
      "eval_samples_per_second": 7.408,
      "eval_steps_per_second": 7.408,
      "step": 2500
    },
    {
      "epoch": 3.863024240092343,
      "grad_norm": 0.04446037486195564,
      "learning_rate": 1.4835363491657412e-05,
      "loss": 0.0087,
      "step": 2510
    },
    {
      "epoch": 3.8784147749134283,
      "grad_norm": 0.4967706799507141,
      "learning_rate": 1.4454981513070791e-05,
      "loss": 0.0041,
      "step": 2520
    },
    {
      "epoch": 3.893805309734513,
      "grad_norm": 0.14164391160011292,
      "learning_rate": 1.4078713954249873e-05,
      "loss": 0.0068,
      "step": 2530
    },
    {
      "epoch": 3.9091958445555983,
      "grad_norm": 0.17527906596660614,
      "learning_rate": 1.3706604369052434e-05,
      "loss": 0.0055,
      "step": 2540
    },
    {
      "epoch": 3.9245863793766835,
      "grad_norm": 0.514822244644165,
      "learning_rate": 1.3338695830041047e-05,
      "loss": 0.0124,
      "step": 2550
    },
    {
      "epoch": 3.9245863793766835,
      "eval_loss": 0.04217739403247833,
      "eval_runtime": 156.0435,
      "eval_samples_per_second": 7.408,
      "eval_steps_per_second": 7.408,
      "step": 2550
    },
    {
      "epoch": 3.9399769141977683,
      "grad_norm": 0.4367717504501343,
      "learning_rate": 1.2975030923497262e-05,
      "loss": 0.007,
      "step": 2560
    },
    {
      "epoch": 3.9553674490188535,
      "grad_norm": 0.38048142194747925,
      "learning_rate": 1.2615651744492151e-05,
      "loss": 0.0056,
      "step": 2570
    },
    {
      "epoch": 3.9707579838399383,
      "grad_norm": 0.8933420777320862,
      "learning_rate": 1.2260599892013719e-05,
      "loss": 0.0136,
      "step": 2580
    },
    {
      "epoch": 3.9861485186610235,
      "grad_norm": 0.41310974955558777,
      "learning_rate": 1.1909916464151715e-05,
      "loss": 0.0075,
      "step": 2590
    },
    {
      "epoch": 4.001539053482109,
      "grad_norm": 0.05755435675382614,
      "learning_rate": 1.1563642053340406e-05,
      "loss": 0.0094,
      "step": 2600
    },
    {
      "epoch": 4.001539053482109,
      "eval_loss": 0.0415499210357666,
      "eval_runtime": 155.9861,
      "eval_samples_per_second": 7.411,
      "eval_steps_per_second": 7.411,
      "step": 2600
    },
    {
      "epoch": 4.016929588303194,
      "grad_norm": 0.1198483407497406,
      "learning_rate": 1.1221816741659964e-05,
      "loss": 0.0052,
      "step": 2610
    },
    {
      "epoch": 4.032320123124278,
      "grad_norm": 0.17106081545352936,
      "learning_rate": 1.0884480096196803e-05,
      "loss": 0.0017,
      "step": 2620
    },
    {
      "epoch": 4.047710657945363,
      "grad_norm": 0.07802654802799225,
      "learning_rate": 1.0551671164463666e-05,
      "loss": 0.002,
      "step": 2630
    },
    {
      "epoch": 4.063101192766449,
      "grad_norm": 0.04642901569604874,
      "learning_rate": 1.0223428469879736e-05,
      "loss": 0.0065,
      "step": 2640
    },
    {
      "epoch": 4.078491727587534,
      "grad_norm": 0.05049436166882515,
      "learning_rate": 9.899790007311455e-06,
      "loss": 0.0019,
      "step": 2650
    },
    {
      "epoch": 4.078491727587534,
      "eval_loss": 0.04245912283658981,
      "eval_runtime": 155.7617,
      "eval_samples_per_second": 7.422,
      "eval_steps_per_second": 7.422,
      "step": 2650
    },
    {
      "epoch": 4.093882262408619,
      "grad_norm": 0.2789390981197357,
      "learning_rate": 9.580793238674573e-06,
      "loss": 0.0048,
      "step": 2660
    },
    {
      "epoch": 4.109272797229703,
      "grad_norm": 0.31272974610328674,
      "learning_rate": 9.266475088597781e-06,
      "loss": 0.0023,
      "step": 2670
    },
    {
      "epoch": 4.124663332050789,
      "grad_norm": 0.07020661234855652,
      "learning_rate": 8.956871940148681e-06,
      "loss": 0.0022,
      "step": 2680
    },
    {
      "epoch": 4.140053866871874,
      "grad_norm": 0.041926730424165726,
      "learning_rate": 8.65201963062225e-06,
      "loss": 0.0041,
      "step": 2690
    },
    {
      "epoch": 4.155444401692959,
      "grad_norm": 0.050659824162721634,
      "learning_rate": 8.351953447392674e-06,
      "loss": 0.0059,
      "step": 2700
    },
    {
      "epoch": 4.155444401692959,
      "eval_loss": 0.04506076127290726,
      "eval_runtime": 156.0638,
      "eval_samples_per_second": 7.407,
      "eval_steps_per_second": 7.407,
      "step": 2700
    },
    {
      "epoch": 4.170834936514044,
      "grad_norm": 0.04850563779473305,
      "learning_rate": 8.056708123828738e-06,
      "loss": 0.004,
      "step": 2710
    },
    {
      "epoch": 4.186225471335129,
      "grad_norm": 0.19528210163116455,
      "learning_rate": 7.76631783527334e-06,
      "loss": 0.0037,
      "step": 2720
    },
    {
      "epoch": 4.201616006156214,
      "grad_norm": 0.04227529838681221,
      "learning_rate": 7.4808161950876145e-06,
      "loss": 0.0024,
      "step": 2730
    },
    {
      "epoch": 4.217006540977299,
      "grad_norm": 0.3444930613040924,
      "learning_rate": 7.200236250760111e-06,
      "loss": 0.0043,
      "step": 2740
    },
    {
      "epoch": 4.232397075798384,
      "grad_norm": 0.1485668569803238,
      "learning_rate": 6.924610480081478e-06,
      "loss": 0.0025,
      "step": 2750
    },
    {
      "epoch": 4.232397075798384,
      "eval_loss": 0.046710480004549026,
      "eval_runtime": 155.7929,
      "eval_samples_per_second": 7.42,
      "eval_steps_per_second": 7.42,
      "step": 2750
    },
    {
      "epoch": 4.247787610619469,
      "grad_norm": 0.17770326137542725,
      "learning_rate": 6.6539707873850774e-06,
      "loss": 0.0021,
      "step": 2760
    },
    {
      "epoch": 4.263178145440554,
      "grad_norm": 0.06237461790442467,
      "learning_rate": 6.3883484998539525e-06,
      "loss": 0.0017,
      "step": 2770
    },
    {
      "epoch": 4.278568680261639,
      "grad_norm": 0.3299518823623657,
      "learning_rate": 6.127774363894667e-06,
      "loss": 0.0031,
      "step": 2780
    },
    {
      "epoch": 4.293959215082724,
      "grad_norm": 0.11566298454999924,
      "learning_rate": 5.872278541578352e-06,
      "loss": 0.0036,
      "step": 2790
    },
    {
      "epoch": 4.309349749903809,
      "grad_norm": 0.037256695330142975,
      "learning_rate": 5.6218906071493275e-06,
      "loss": 0.0026,
      "step": 2800
    },
    {
      "epoch": 4.309349749903809,
      "eval_loss": 0.0478338748216629,
      "eval_runtime": 155.839,
      "eval_samples_per_second": 7.418,
      "eval_steps_per_second": 7.418,
      "step": 2800
    },
    {
      "epoch": 4.324740284724895,
      "grad_norm": 0.016917018219828606,
      "learning_rate": 5.376639543601858e-06,
      "loss": 0.0024,
      "step": 2810
    },
    {
      "epoch": 4.340130819545979,
      "grad_norm": 0.027036458253860474,
      "learning_rate": 5.136553739325256e-06,
      "loss": 0.002,
      "step": 2820
    },
    {
      "epoch": 4.355521354367064,
      "grad_norm": 0.09365002810955048,
      "learning_rate": 4.901660984817908e-06,
      "loss": 0.0026,
      "step": 2830
    },
    {
      "epoch": 4.370911889188149,
      "grad_norm": 0.5041496753692627,
      "learning_rate": 4.671988469470434e-06,
      "loss": 0.0042,
      "step": 2840
    },
    {
      "epoch": 4.3863024240092345,
      "grad_norm": 0.05433773994445801,
      "learning_rate": 4.447562778418435e-06,
      "loss": 0.0026,
      "step": 2850
    },
    {
      "epoch": 4.3863024240092345,
      "eval_loss": 0.048599518835544586,
      "eval_runtime": 156.1256,
      "eval_samples_per_second": 7.404,
      "eval_steps_per_second": 7.404,
      "step": 2850
    },
    {
      "epoch": 4.40169295883032,
      "grad_norm": 0.25091418623924255,
      "learning_rate": 4.228409889465218e-06,
      "loss": 0.0018,
      "step": 2860
    },
    {
      "epoch": 4.417083493651404,
      "grad_norm": 0.037694767117500305,
      "learning_rate": 4.014555170074824e-06,
      "loss": 0.0012,
      "step": 2870
    },
    {
      "epoch": 4.432474028472489,
      "grad_norm": 0.07268691807985306,
      "learning_rate": 3.8060233744356633e-06,
      "loss": 0.007,
      "step": 2880
    },
    {
      "epoch": 4.4478645632935745,
      "grad_norm": 0.03574419766664505,
      "learning_rate": 3.602838640595152e-06,
      "loss": 0.0019,
      "step": 2890
    },
    {
      "epoch": 4.46325509811466,
      "grad_norm": 0.3148428201675415,
      "learning_rate": 3.4050244876656866e-06,
      "loss": 0.0032,
      "step": 2900
    },
    {
      "epoch": 4.46325509811466,
      "eval_loss": 0.049422282725572586,
      "eval_runtime": 156.0803,
      "eval_samples_per_second": 7.406,
      "eval_steps_per_second": 7.406,
      "step": 2900
    },
    {
      "epoch": 4.478645632935745,
      "grad_norm": 0.07241390645503998,
      "learning_rate": 3.2126038131022552e-06,
      "loss": 0.007,
      "step": 2910
    },
    {
      "epoch": 4.494036167756829,
      "grad_norm": 0.07479214668273926,
      "learning_rate": 3.025598890051995e-06,
      "loss": 0.004,
      "step": 2920
    },
    {
      "epoch": 4.5094267025779144,
      "grad_norm": 0.2251429706811905,
      "learning_rate": 2.8440313647760288e-06,
      "loss": 0.0023,
      "step": 2930
    },
    {
      "epoch": 4.524817237399,
      "grad_norm": 0.5468044281005859,
      "learning_rate": 2.6679222541438397e-06,
      "loss": 0.0028,
      "step": 2940
    },
    {
      "epoch": 4.540207772220085,
      "grad_norm": 0.015656841918826103,
      "learning_rate": 2.4972919432005582e-06,
      "loss": 0.0022,
      "step": 2950
    },
    {
      "epoch": 4.540207772220085,
      "eval_loss": 0.049510955810546875,
      "eval_runtime": 156.08,
      "eval_samples_per_second": 7.406,
      "eval_steps_per_second": 7.406,
      "step": 2950
    },
    {
      "epoch": 4.555598307041169,
      "grad_norm": 0.05784641206264496,
      "learning_rate": 2.3321601828073135e-06,
      "loss": 0.0007,
      "step": 2960
    },
    {
      "epoch": 4.570988841862254,
      "grad_norm": 0.19153526425361633,
      "learning_rate": 2.1725460873550317e-06,
      "loss": 0.0024,
      "step": 2970
    },
    {
      "epoch": 4.58637937668334,
      "grad_norm": 0.015539661981165409,
      "learning_rate": 2.0184681325519096e-06,
      "loss": 0.0025,
      "step": 2980
    },
    {
      "epoch": 4.601769911504425,
      "grad_norm": 0.27713876962661743,
      "learning_rate": 1.869944153284786e-06,
      "loss": 0.0034,
      "step": 2990
    },
    {
      "epoch": 4.61716044632551,
      "grad_norm": 0.20282918214797974,
      "learning_rate": 1.7269913415547456e-06,
      "loss": 0.0028,
      "step": 3000
    },
    {
      "epoch": 4.61716044632551,
      "eval_loss": 0.04998774826526642,
      "eval_runtime": 156.1348,
      "eval_samples_per_second": 7.404,
      "eval_steps_per_second": 7.404,
      "step": 3000
    },
    {
      "epoch": 4.632550981146595,
      "grad_norm": 0.0793176218867302,
      "learning_rate": 1.589626244487069e-06,
      "loss": 0.0015,
      "step": 3010
    },
    {
      "epoch": 4.64794151596768,
      "grad_norm": 0.07218535989522934,
      "learning_rate": 1.4578647624158726e-06,
      "loss": 0.0035,
      "step": 3020
    },
    {
      "epoch": 4.663332050788765,
      "grad_norm": 0.10875333845615387,
      "learning_rate": 1.3317221470436293e-06,
      "loss": 0.0064,
      "step": 3030
    },
    {
      "epoch": 4.67872258560985,
      "grad_norm": 0.08289933204650879,
      "learning_rate": 1.2112129996757315e-06,
      "loss": 0.0033,
      "step": 3040
    },
    {
      "epoch": 4.694113120430935,
      "grad_norm": 0.017408736050128937,
      "learning_rate": 1.0963512695303546e-06,
      "loss": 0.0023,
      "step": 3050
    },
    {
      "epoch": 4.694113120430935,
      "eval_loss": 0.05005095899105072,
      "eval_runtime": 156.1135,
      "eval_samples_per_second": 7.405,
      "eval_steps_per_second": 7.405,
      "step": 3050
    },
    {
      "epoch": 4.7095036552520195,
      "grad_norm": 0.11436308920383453,
      "learning_rate": 9.871502521237975e-07,
      "loss": 0.0056,
      "step": 3060
    },
    {
      "epoch": 4.724894190073105,
      "grad_norm": 0.5728814601898193,
      "learning_rate": 8.83622587731514e-07,
      "loss": 0.0031,
      "step": 3070
    },
    {
      "epoch": 4.74028472489419,
      "grad_norm": 0.7811269760131836,
      "learning_rate": 7.857802599249564e-07,
      "loss": 0.0043,
      "step": 3080
    },
    {
      "epoch": 4.755675259715275,
      "grad_norm": 0.04214154928922653,
      "learning_rate": 6.936345941844335e-07,
      "loss": 0.0042,
      "step": 3090
    },
    {
      "epoch": 4.77106579453636,
      "grad_norm": 0.10212664306163788,
      "learning_rate": 6.071962565881939e-07,
      "loss": 0.0028,
      "step": 3100
    },
    {
      "epoch": 4.77106579453636,
      "eval_loss": 0.05000366270542145,
      "eval_runtime": 156.1486,
      "eval_samples_per_second": 7.403,
      "eval_steps_per_second": 7.403,
      "step": 3100
    },
    {
      "epoch": 4.786456329357446,
      "grad_norm": 0.0864560678601265,
      "learning_rate": 5.264752525777961e-07,
      "loss": 0.0048,
      "step": 3110
    },
    {
      "epoch": 4.80184686417853,
      "grad_norm": 0.12185970693826675,
      "learning_rate": 4.5148092579991863e-07,
      "loss": 0.002,
      "step": 3120
    },
    {
      "epoch": 4.817237398999615,
      "grad_norm": 0.010840833187103271,
      "learning_rate": 3.822219570248531e-07,
      "loss": 0.0038,
      "step": 3130
    },
    {
      "epoch": 4.8326279338207,
      "grad_norm": 0.007412767969071865,
      "learning_rate": 3.187063631416742e-07,
      "loss": 0.0016,
      "step": 3140
    },
    {
      "epoch": 4.8480184686417855,
      "grad_norm": 0.2218623012304306,
      "learning_rate": 2.6094149623024365e-07,
      "loss": 0.0059,
      "step": 3150
    },
    {
      "epoch": 4.8480184686417855,
      "eval_loss": 0.05003925785422325,
      "eval_runtime": 155.8146,
      "eval_samples_per_second": 7.419,
      "eval_steps_per_second": 7.419,
      "step": 3150
    },
    {
      "epoch": 4.86340900346287,
      "grad_norm": 0.09100077301263809,
      "learning_rate": 2.0893404271022376e-07,
      "loss": 0.005,
      "step": 3160
    },
    {
      "epoch": 4.878799538283955,
      "grad_norm": 0.4347333014011383,
      "learning_rate": 1.626900225670691e-07,
      "loss": 0.0034,
      "step": 3170
    },
    {
      "epoch": 4.89419007310504,
      "grad_norm": 0.09482406079769135,
      "learning_rate": 1.2221478865524493e-07,
      "loss": 0.0029,
      "step": 3180
    },
    {
      "epoch": 4.9095806079261255,
      "grad_norm": 0.13639184832572937,
      "learning_rate": 8.751302607857836e-08,
      "loss": 0.0023,
      "step": 3190
    },
    {
      "epoch": 4.924971142747211,
      "grad_norm": 0.02655295468866825,
      "learning_rate": 5.8588751647975506e-08,
      "loss": 0.0026,
      "step": 3200
    },
    {
      "epoch": 4.924971142747211,
      "eval_loss": 0.049996137619018555,
      "eval_runtime": 155.7806,
      "eval_samples_per_second": 7.421,
      "eval_steps_per_second": 7.421,
      "step": 3200
    },
    {
      "epoch": 4.940361677568296,
      "grad_norm": 0.2657541334629059,
      "learning_rate": 3.544531341646007e-08,
      "loss": 0.0036,
      "step": 3210
    },
    {
      "epoch": 4.95575221238938,
      "grad_norm": 0.02715928852558136,
      "learning_rate": 1.808539029161116e-08,
      "loss": 0.0029,
      "step": 3220
    },
    {
      "epoch": 4.9711427472104655,
      "grad_norm": 0.45431065559387207,
      "learning_rate": 6.510991725500182e-09,
      "loss": 0.0051,
      "step": 3230
    },
    {
      "epoch": 4.986533282031551,
      "grad_norm": 0.07858604937791824,
      "learning_rate": 7.234574820769169e-10,
      "loss": 0.0035,
      "step": 3240
    },
    {
      "epoch": 4.994228549442093,
      "step": 3245,
      "total_flos": 7.696760527880847e+17,
      "train_loss": 0.05130562667788016,
      "train_runtime": 36886.3364,
      "train_samples_per_second": 1.409,
      "train_steps_per_second": 0.088
    }
  ],
  "logging_steps": 10,
  "max_steps": 3245,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 7.696760527880847e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}