{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.44,
  "eval_steps": 1000,
  "global_step": 22000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.002,
      "grad_norm": 0.8063191175460815,
      "learning_rate": 7.920000000000001e-05,
      "loss": 8.2537,
      "step": 100
    },
    {
      "epoch": 0.004,
      "grad_norm": 0.6161019206047058,
      "learning_rate": 0.00015920000000000002,
      "loss": 6.5731,
      "step": 200
    },
    {
      "epoch": 0.006,
      "grad_norm": 0.8347169160842896,
      "learning_rate": 0.00023920000000000001,
      "loss": 6.1639,
      "step": 300
    },
    {
      "epoch": 0.008,
      "grad_norm": 0.9150793552398682,
      "learning_rate": 0.0003192,
      "loss": 5.8682,
      "step": 400
    },
    {
      "epoch": 0.01,
      "grad_norm": 0.6698479652404785,
      "learning_rate": 0.0003992,
      "loss": 5.5708,
      "step": 500
    },
    {
      "epoch": 0.012,
      "grad_norm": 0.5323518514633179,
      "learning_rate": 0.00047920000000000005,
      "loss": 5.3157,
      "step": 600
    },
    {
      "epoch": 0.014,
      "grad_norm": 0.7472386956214905,
      "learning_rate": 0.0005592,
      "loss": 5.0907,
      "step": 700
    },
    {
      "epoch": 0.016,
      "grad_norm": 0.6033297181129456,
      "learning_rate": 0.0006392,
      "loss": 4.8805,
      "step": 800
    },
    {
      "epoch": 0.018,
      "grad_norm": 0.497113436460495,
      "learning_rate": 0.0007191999999999999,
      "loss": 4.7207,
      "step": 900
    },
    {
      "epoch": 0.02,
      "grad_norm": 0.47698545455932617,
      "learning_rate": 0.0007992,
      "loss": 4.6068,
      "step": 1000
    },
    {
      "epoch": 0.02,
      "eval_accuracy": 0.2703913894324853,
      "eval_loss": 4.5053935050964355,
      "eval_runtime": 13.0992,
      "eval_samples_per_second": 76.341,
      "eval_steps_per_second": 0.611,
      "step": 1000
    },
    {
      "epoch": 0.022,
      "grad_norm": 0.44341349601745605,
      "learning_rate": 0.0008792,
      "loss": 4.4904,
      "step": 1100
    },
    {
      "epoch": 0.024,
      "grad_norm": 0.4383724331855774,
      "learning_rate": 0.0009592000000000001,
      "loss": 4.439,
      "step": 1200
    },
    {
      "epoch": 0.026,
      "grad_norm": 0.3894674777984619,
      "learning_rate": 0.0010391999999999999,
      "loss": 4.3735,
      "step": 1300
    },
    {
      "epoch": 0.028,
      "grad_norm": 0.41903820633888245,
      "learning_rate": 0.0011192,
      "loss": 4.3406,
      "step": 1400
    },
    {
      "epoch": 0.03,
      "grad_norm": 0.36601418256759644,
      "learning_rate": 0.0011992,
      "loss": 4.2879,
      "step": 1500
    },
    {
      "epoch": 0.032,
      "grad_norm": 0.4282030165195465,
      "learning_rate": 0.0012791999999999999,
      "loss": 4.2625,
      "step": 1600
    },
    {
      "epoch": 0.034,
      "grad_norm": 0.3431214690208435,
      "learning_rate": 0.0013592,
      "loss": 4.2316,
      "step": 1700
    },
    {
      "epoch": 0.036,
      "grad_norm": 0.3303460478782654,
      "learning_rate": 0.0014392,
      "loss": 4.1976,
      "step": 1800
    },
    {
      "epoch": 0.038,
      "grad_norm": 0.3030431866645813,
      "learning_rate": 0.0015192,
      "loss": 4.1494,
      "step": 1900
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.3076256513595581,
      "learning_rate": 0.0015992,
      "loss": 4.1688,
      "step": 2000
    },
    {
      "epoch": 0.04,
      "eval_accuracy": 0.30658512720156555,
      "eval_loss": 4.094621658325195,
      "eval_runtime": 12.4723,
      "eval_samples_per_second": 80.178,
      "eval_steps_per_second": 0.641,
      "step": 2000
    },
    {
      "epoch": 0.042,
      "grad_norm": 0.2820931077003479,
      "learning_rate": 0.0016792,
      "loss": 4.5201,
      "step": 2100
    },
    {
      "epoch": 0.044,
      "grad_norm": 0.23724448680877686,
      "learning_rate": 0.0017592,
      "loss": 4.1333,
      "step": 2200
    },
    {
      "epoch": 0.046,
      "grad_norm": 0.2325439602136612,
      "learning_rate": 0.0018392,
      "loss": 4.1072,
      "step": 2300
    },
    {
      "epoch": 0.048,
      "grad_norm": 0.23736131191253662,
      "learning_rate": 0.0019192,
      "loss": 4.0682,
      "step": 2400
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.2313733696937561,
      "learning_rate": 0.0019992,
      "loss": 4.0571,
      "step": 2500
    },
    {
      "epoch": 0.052,
      "grad_norm": 0.22318924963474274,
      "learning_rate": 0.001999978563623903,
      "loss": 4.0153,
      "step": 2600
    },
    {
      "epoch": 0.054,
      "grad_norm": 0.20275355875492096,
      "learning_rate": 0.0019999133871331223,
      "loss": 4.0163,
      "step": 2700
    },
    {
      "epoch": 0.056,
      "grad_norm": 0.21306632459163666,
      "learning_rate": 0.0019998044711915177,
      "loss": 3.9777,
      "step": 2800
    },
    {
      "epoch": 0.058,
      "grad_norm": 0.19690750539302826,
      "learning_rate": 0.0019996518205634257,
      "loss": 3.9603,
      "step": 2900
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.20464760065078735,
      "learning_rate": 0.0019994554419262797,
      "loss": 3.9566,
      "step": 3000
    },
    {
      "epoch": 0.06,
      "eval_accuracy": 0.3226555772994129,
      "eval_loss": 3.9112932682037354,
      "eval_runtime": 11.7171,
      "eval_samples_per_second": 85.345,
      "eval_steps_per_second": 0.683,
      "step": 3000
    },
    {
      "epoch": 0.062,
      "grad_norm": 0.18151508271694183,
      "learning_rate": 0.001999215343870317,
      "loss": 3.9501,
      "step": 3100
    },
    {
      "epoch": 0.064,
      "grad_norm": 0.18748600780963898,
      "learning_rate": 0.0019989315368982054,
      "loss": 3.9228,
      "step": 3200
    },
    {
      "epoch": 0.066,
      "grad_norm": 0.17895972728729248,
      "learning_rate": 0.0019986040334245797,
      "loss": 3.9041,
      "step": 3300
    },
    {
      "epoch": 0.068,
      "grad_norm": 0.18976759910583496,
      "learning_rate": 0.001998232847775504,
      "loss": 3.9123,
      "step": 3400
    },
    {
      "epoch": 0.07,
      "grad_norm": 0.1871727555990219,
      "learning_rate": 0.0019978179961878404,
      "loss": 3.8846,
      "step": 3500
    },
    {
      "epoch": 0.072,
      "grad_norm": 0.17986708879470825,
      "learning_rate": 0.001997359496808541,
      "loss": 3.8794,
      "step": 3600
    },
    {
      "epoch": 0.074,
      "grad_norm": 0.17090782523155212,
      "learning_rate": 0.001996857369693855,
      "loss": 3.8593,
      "step": 3700
    },
    {
      "epoch": 0.076,
      "grad_norm": 0.19015273451805115,
      "learning_rate": 0.0019963116368084486,
      "loss": 3.8642,
      "step": 3800
    },
    {
      "epoch": 0.078,
      "grad_norm": 0.1871633529663086,
      "learning_rate": 0.001995722322024446,
      "loss": 3.8479,
      "step": 3900
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.18049727380275726,
      "learning_rate": 0.001995089451120385,
      "loss": 3.8226,
      "step": 4000
    },
    {
      "epoch": 0.08,
      "eval_accuracy": 0.33194520547945205,
      "eval_loss": 3.8037452697753906,
      "eval_runtime": 12.6853,
      "eval_samples_per_second": 78.832,
      "eval_steps_per_second": 0.631,
      "step": 4000
    },
    {
      "epoch": 0.082,
      "grad_norm": 0.14800746738910675,
      "learning_rate": 0.0019944130517800893,
      "loss": 3.8319,
      "step": 4100
    },
    {
      "epoch": 0.084,
      "grad_norm": 0.14713813364505768,
      "learning_rate": 0.001993693153591457,
      "loss": 3.8319,
      "step": 4200
    },
    {
      "epoch": 0.086,
      "grad_norm": 0.17019635438919067,
      "learning_rate": 0.0019929297880451674,
      "loss": 3.8095,
      "step": 4300
    },
    {
      "epoch": 0.088,
      "grad_norm": 0.15243591368198395,
      "learning_rate": 0.0019921229885333023,
      "loss": 3.8003,
      "step": 4400
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.16134986281394958,
      "learning_rate": 0.001991272790347886,
      "loss": 3.814,
      "step": 4500
    },
    {
      "epoch": 0.092,
      "grad_norm": 0.16883811354637146,
      "learning_rate": 0.0019903792306793415,
      "loss": 3.7822,
      "step": 4600
    },
    {
      "epoch": 0.094,
      "grad_norm": 0.15718473494052887,
      "learning_rate": 0.001989442348614863,
      "loss": 3.7902,
      "step": 4700
    },
    {
      "epoch": 0.096,
      "grad_norm": 0.14631035923957825,
      "learning_rate": 0.0019884621851367075,
      "loss": 3.7805,
      "step": 4800
    },
    {
      "epoch": 0.098,
      "grad_norm": 0.14771369099617004,
      "learning_rate": 0.001987438783120401,
      "loss": 3.7778,
      "step": 4900
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.1536298543214798,
      "learning_rate": 0.001986372187332862,
      "loss": 3.7848,
      "step": 5000
    },
    {
      "epoch": 0.1,
      "eval_accuracy": 0.340252446183953,
      "eval_loss": 3.7290549278259277,
      "eval_runtime": 11.9375,
      "eval_samples_per_second": 83.77,
      "eval_steps_per_second": 0.67,
      "step": 5000
    },
    {
      "epoch": 0.102,
      "grad_norm": 0.15110976994037628,
      "learning_rate": 0.0019852624444304467,
      "loss": 3.7502,
      "step": 5100
    },
    {
      "epoch": 0.104,
      "grad_norm": 0.15862922370433807,
      "learning_rate": 0.0019841096029569044,
      "loss": 3.761,
      "step": 5200
    },
    {
      "epoch": 0.106,
      "grad_norm": 0.13948658108711243,
      "learning_rate": 0.0019829137133412556,
      "loss": 3.7591,
      "step": 5300
    },
    {
      "epoch": 0.108,
      "grad_norm": 0.148148313164711,
      "learning_rate": 0.001981674827895587,
      "loss": 3.7385,
      "step": 5400
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.13170774281024933,
      "learning_rate": 0.00198039300081276,
      "loss": 3.7424,
      "step": 5500
    },
    {
      "epoch": 0.112,
      "grad_norm": 0.14352479577064514,
      "learning_rate": 0.0019790682881640448,
      "loss": 3.7423,
      "step": 5600
    },
    {
      "epoch": 0.114,
      "grad_norm": 0.12865835428237915,
      "learning_rate": 0.001977700747896664,
      "loss": 3.7263,
      "step": 5700
    },
    {
      "epoch": 0.116,
      "grad_norm": 0.15470698475837708,
      "learning_rate": 0.001976290439831259,
      "loss": 3.7297,
      "step": 5800
    },
    {
      "epoch": 0.118,
      "grad_norm": 0.1353532075881958,
      "learning_rate": 0.0019748374256592736,
      "loss": 3.7218,
      "step": 5900
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.14982788264751434,
      "learning_rate": 0.0019733417689402543,
      "loss": 3.7164,
      "step": 6000
    },
    {
      "epoch": 0.12,
      "eval_accuracy": 0.34496673189823873,
      "eval_loss": 3.6746184825897217,
      "eval_runtime": 11.9904,
      "eval_samples_per_second": 83.4,
      "eval_steps_per_second": 0.667,
      "step": 6000
    },
    {
      "epoch": 0.122,
      "grad_norm": 0.14931881427764893,
      "learning_rate": 0.0019718035350990712,
      "loss": 3.7169,
      "step": 6100
    },
    {
      "epoch": 0.124,
      "grad_norm": 0.14251381158828735,
      "learning_rate": 0.0019702227914230566,
      "loss": 3.6929,
      "step": 6200
    },
    {
      "epoch": 0.126,
      "grad_norm": 0.14251849055290222,
      "learning_rate": 0.001968599607059059,
      "loss": 3.7018,
      "step": 6300
    },
    {
      "epoch": 0.128,
      "grad_norm": 0.15696577727794647,
      "learning_rate": 0.0019669340530104207,
      "loss": 3.7034,
      "step": 6400
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.14737001061439514,
      "learning_rate": 0.001965226202133872,
      "loss": 3.6889,
      "step": 6500
    },
    {
      "epoch": 0.132,
      "grad_norm": 0.13667502999305725,
      "learning_rate": 0.0019634761291363427,
      "loss": 3.6935,
      "step": 6600
    },
    {
      "epoch": 0.134,
      "grad_norm": 0.13390734791755676,
      "learning_rate": 0.0019616839105716954,
      "loss": 3.7032,
      "step": 6700
    },
    {
      "epoch": 0.136,
      "grad_norm": 0.14834214746952057,
      "learning_rate": 0.0019598496248373755,
      "loss": 3.6657,
      "step": 6800
    },
    {
      "epoch": 0.138,
      "grad_norm": 0.1314767599105835,
      "learning_rate": 0.001957973352170984,
      "loss": 3.673,
      "step": 6900
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.15161661803722382,
      "learning_rate": 0.001956055174646765,
      "loss": 3.6713,
      "step": 7000
    },
    {
      "epoch": 0.14,
      "eval_accuracy": 0.3484422700587084,
      "eval_loss": 3.6349546909332275,
      "eval_runtime": 12.1265,
      "eval_samples_per_second": 82.464,
      "eval_steps_per_second": 0.66,
      "step": 7000
    },
    {
      "epoch": 0.142,
      "grad_norm": 0.12710200250148773,
      "learning_rate": 0.0019540951761720174,
      "loss": 3.67,
      "step": 7100
    },
    {
      "epoch": 0.144,
      "grad_norm": 0.18498213589191437,
      "learning_rate": 0.0019520934424834247,
      "loss": 3.675,
      "step": 7200
    },
    {
      "epoch": 0.146,
      "grad_norm": 0.13697299361228943,
      "learning_rate": 0.0019500500611433025,
      "loss": 3.6497,
      "step": 7300
    },
    {
      "epoch": 0.148,
      "grad_norm": 0.13290269672870636,
      "learning_rate": 0.0019479651215357707,
      "loss": 3.662,
      "step": 7400
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.12744054198265076,
      "learning_rate": 0.0019458387148628417,
      "loss": 3.6645,
      "step": 7500
    },
    {
      "epoch": 0.152,
      "grad_norm": 0.13697493076324463,
      "learning_rate": 0.001943670934140432,
      "loss": 3.6374,
      "step": 7600
    },
    {
      "epoch": 0.154,
      "grad_norm": 0.14067181944847107,
      "learning_rate": 0.0019414618741942936,
      "loss": 3.6453,
      "step": 7700
    },
    {
      "epoch": 0.156,
      "grad_norm": 0.1316055953502655,
      "learning_rate": 0.0019392116316558638,
      "loss": 3.6679,
      "step": 7800
    },
    {
      "epoch": 0.158,
      "grad_norm": 0.1261526495218277,
      "learning_rate": 0.001936920304958042,
      "loss": 3.6343,
      "step": 7900
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.13644501566886902,
      "learning_rate": 0.0019345879943308804,
      "loss": 3.6365,
      "step": 8000
    },
    {
      "epoch": 0.16,
      "eval_accuracy": 0.35229158512720155,
      "eval_loss": 3.5980424880981445,
      "eval_runtime": 17.6098,
      "eval_samples_per_second": 56.787,
      "eval_steps_per_second": 0.454,
      "step": 8000
    },
    {
      "epoch": 0.162,
      "grad_norm": 0.13911907374858856,
      "learning_rate": 0.0019322148017972016,
      "loss": 3.646,
      "step": 8100
    },
    {
      "epoch": 0.164,
      "grad_norm": 0.1297680288553238,
      "learning_rate": 0.001929800831168135,
      "loss": 3.6412,
      "step": 8200
    },
    {
      "epoch": 0.166,
      "grad_norm": 0.14803367853164673,
      "learning_rate": 0.001927346188038576,
      "loss": 3.6298,
      "step": 8300
    },
    {
      "epoch": 0.168,
      "grad_norm": 0.1325535923242569,
      "learning_rate": 0.0019248509797825672,
      "loss": 3.6067,
      "step": 8400
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.16948860883712769,
      "learning_rate": 0.0019223153155486009,
      "loss": 3.6284,
      "step": 8500
    },
    {
      "epoch": 0.172,
      "grad_norm": 0.12973898649215698,
      "learning_rate": 0.0019197393062548454,
      "loss": 3.627,
      "step": 8600
    },
    {
      "epoch": 0.174,
      "grad_norm": 0.1576964110136032,
      "learning_rate": 0.0019171230645842923,
      "loss": 3.6027,
      "step": 8700
    },
    {
      "epoch": 0.176,
      "grad_norm": 0.15394999086856842,
      "learning_rate": 0.0019144667049798272,
      "loss": 3.6098,
      "step": 8800
    },
    {
      "epoch": 0.178,
      "grad_norm": 0.12594962120056152,
      "learning_rate": 0.0019117703436392253,
      "loss": 3.6221,
      "step": 8900
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.1420104056596756,
      "learning_rate": 0.001909034098510066,
      "loss": 3.5993,
      "step": 9000
    },
    {
      "epoch": 0.18,
      "eval_accuracy": 0.3549628180039139,
      "eval_loss": 3.570964813232422,
      "eval_runtime": 12.4303,
      "eval_samples_per_second": 80.449,
      "eval_steps_per_second": 0.644,
      "step": 9000
    },
    {
      "epoch": 0.182,
      "grad_norm": 0.12720872461795807,
      "learning_rate": 0.001906258089284576,
      "loss": 3.5913,
      "step": 9100
    },
    {
      "epoch": 0.184,
      "grad_norm": 0.14890190958976746,
      "learning_rate": 0.0019034424373943915,
      "loss": 3.6113,
      "step": 9200
    },
    {
      "epoch": 0.186,
      "grad_norm": 0.15023267269134521,
      "learning_rate": 0.0019005872660052478,
      "loss": 3.6011,
      "step": 9300
    },
    {
      "epoch": 0.188,
      "grad_norm": 0.1637164056301117,
      "learning_rate": 0.001897692700011591,
      "loss": 3.5932,
      "step": 9400
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.13102982938289642,
      "learning_rate": 0.0018947588660311143,
      "loss": 3.5889,
      "step": 9500
    },
    {
      "epoch": 0.192,
      "grad_norm": 0.16267850995063782,
      "learning_rate": 0.0018917858923992211,
      "loss": 3.5939,
      "step": 9600
    },
    {
      "epoch": 0.194,
      "grad_norm": 0.12798982858657837,
      "learning_rate": 0.0018887739091634085,
      "loss": 3.5896,
      "step": 9700
    },
    {
      "epoch": 0.196,
      "grad_norm": 0.13872383534908295,
      "learning_rate": 0.0018857230480775807,
      "loss": 3.574,
      "step": 9800
    },
    {
      "epoch": 0.198,
      "grad_norm": 0.11993825435638428,
      "learning_rate": 0.0018826334425962855,
      "loss": 3.589,
      "step": 9900
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.13074541091918945,
      "learning_rate": 0.0018795052278688753,
      "loss": 3.5828,
      "step": 10000
    },
    {
      "epoch": 0.2,
      "eval_accuracy": 0.35816046966731896,
      "eval_loss": 3.540745973587036,
      "eval_runtime": 17.8387,
      "eval_samples_per_second": 56.058,
      "eval_steps_per_second": 0.448,
      "step": 10000
    },
    {
      "epoch": 0.202,
      "grad_norm": 0.13834109902381897,
      "learning_rate": 0.0018763385407335963,
      "loss": 3.5627,
      "step": 10100
    },
    {
      "epoch": 0.204,
      "grad_norm": 0.15819378197193146,
      "learning_rate": 0.001873133519711602,
      "loss": 3.5634,
      "step": 10200
    },
    {
      "epoch": 0.206,
      "grad_norm": 0.1618553251028061,
      "learning_rate": 0.0018698903050008956,
      "loss": 3.5768,
      "step": 10300
    },
    {
      "epoch": 0.208,
      "grad_norm": 0.14207501709461212,
      "learning_rate": 0.0018666090384701947,
      "loss": 3.5676,
      "step": 10400
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.13513191044330597,
      "learning_rate": 0.001863289863652727,
      "loss": 3.5653,
      "step": 10500
    },
    {
      "epoch": 0.212,
      "grad_norm": 0.12555710971355438,
      "learning_rate": 0.0018599329257399516,
      "loss": 3.5577,
      "step": 10600
    },
    {
      "epoch": 0.214,
      "grad_norm": 0.129461869597435,
      "learning_rate": 0.0018565383715752083,
      "loss": 3.576,
      "step": 10700
    },
    {
      "epoch": 0.216,
      "grad_norm": 0.11782678216695786,
      "learning_rate": 0.0018531063496472927,
      "loss": 3.5695,
      "step": 10800
    },
    {
      "epoch": 0.218,
      "grad_norm": 0.1391923427581787,
      "learning_rate": 0.0018496370100839622,
      "loss": 3.5478,
      "step": 10900
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.15454861521720886,
      "learning_rate": 0.0018461305046453683,
      "loss": 3.56,
      "step": 11000
    },
    {
      "epoch": 0.22,
      "eval_accuracy": 0.3610430528375734,
      "eval_loss": 3.5181522369384766,
      "eval_runtime": 12.5872,
      "eval_samples_per_second": 79.446,
      "eval_steps_per_second": 0.636,
      "step": 11000
    },
    {
      "epoch": 0.222,
      "grad_norm": 0.14800433814525604,
      "learning_rate": 0.0018425869867174187,
      "loss": 3.5559,
      "step": 11100
    },
    {
      "epoch": 0.224,
      "grad_norm": 0.16677607595920563,
      "learning_rate": 0.0018390066113050665,
      "loss": 3.5393,
      "step": 11200
    },
    {
      "epoch": 0.226,
      "grad_norm": 0.17119687795639038,
      "learning_rate": 0.0018353895350255317,
      "loss": 3.5503,
      "step": 11300
    },
    {
      "epoch": 0.228,
      "grad_norm": 0.16998586058616638,
      "learning_rate": 0.0018317359161014477,
      "loss": 3.5546,
      "step": 11400
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.16900590062141418,
      "learning_rate": 0.001828045914353943,
      "loss": 3.5485,
      "step": 11500
    },
    {
      "epoch": 0.232,
      "grad_norm": 0.13857169449329376,
      "learning_rate": 0.0018243196911956476,
      "loss": 3.5357,
      "step": 11600
    },
    {
      "epoch": 0.234,
      "grad_norm": 0.134566068649292,
      "learning_rate": 0.0018205574096236336,
      "loss": 3.5478,
      "step": 11700
    },
    {
      "epoch": 0.236,
      "grad_norm": 0.13580797612667084,
      "learning_rate": 0.0018167592342122857,
      "loss": 3.5452,
      "step": 11800
    },
    {
      "epoch": 0.238,
      "grad_norm": 0.16239416599273682,
      "learning_rate": 0.0018129253311061002,
      "loss": 3.5256,
      "step": 11900
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.16310077905654907,
      "learning_rate": 0.0018090558680124193,
      "loss": 3.5331,
      "step": 12000
    },
    {
      "epoch": 0.24,
      "eval_accuracy": 0.3630156555772994,
      "eval_loss": 3.4982736110687256,
      "eval_runtime": 12.3123,
      "eval_samples_per_second": 81.22,
      "eval_steps_per_second": 0.65,
      "step": 12000
    },
    {
      "epoch": 0.242,
      "grad_norm": 0.1448379009962082,
      "learning_rate": 0.0018051510141940939,
      "loss": 3.5416,
      "step": 12100
    },
    {
      "epoch": 0.244,
      "grad_norm": 0.13583365082740784,
      "learning_rate": 0.00180121094046208,
      "loss": 3.5343,
      "step": 12200
    },
    {
      "epoch": 0.246,
      "grad_norm": 0.17297838628292084,
      "learning_rate": 0.001797235819167967,
      "loss": 3.521,
      "step": 12300
    },
    {
      "epoch": 0.248,
      "grad_norm": 0.15616372227668762,
      "learning_rate": 0.0017932258241964375,
      "loss": 3.533,
      "step": 12400
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.15943922102451324,
      "learning_rate": 0.0017891811309576622,
      "loss": 3.5177,
      "step": 12500
    },
    {
      "epoch": 0.252,
      "grad_norm": 0.15389597415924072,
      "learning_rate": 0.001785101916379627,
      "loss": 3.5209,
      "step": 12600
    },
    {
      "epoch": 0.254,
      "grad_norm": 0.13968117535114288,
      "learning_rate": 0.0017809883589003912,
      "loss": 3.5137,
      "step": 12700
    },
    {
      "epoch": 0.256,
      "grad_norm": 0.1350937783718109,
      "learning_rate": 0.0017768406384602862,
      "loss": 3.5294,
      "step": 12800
    },
    {
      "epoch": 0.258,
      "grad_norm": 0.13691306114196777,
      "learning_rate": 0.00177265893649404,
      "loss": 3.5228,
      "step": 12900
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.1496269851922989,
      "learning_rate": 0.0017684434359228438,
      "loss": 3.5026,
      "step": 13000
    },
    {
      "epoch": 0.26,
      "eval_accuracy": 0.3645440313111546,
      "eval_loss": 3.480961799621582,
      "eval_runtime": 11.914,
      "eval_samples_per_second": 83.935,
      "eval_steps_per_second": 0.671,
      "step": 13000
    },
    {
      "epoch": 0.262,
      "grad_norm": 0.1472535878419876,
      "learning_rate": 0.0017641943211463488,
      "loss": 3.5107,
      "step": 13100
    },
    {
      "epoch": 0.264,
      "grad_norm": 0.1573815494775772,
      "learning_rate": 0.0017599117780346006,
      "loss": 3.5147,
      "step": 13200
    },
    {
      "epoch": 0.266,
      "grad_norm": 0.15841859579086304,
      "learning_rate": 0.0017555959939199084,
      "loss": 3.501,
      "step": 13300
    },
    {
      "epoch": 0.268,
      "grad_norm": 0.15096184611320496,
      "learning_rate": 0.0017512471575886505,
      "loss": 3.5034,
      "step": 13400
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.17669688165187836,
      "learning_rate": 0.0017468654592730168,
      "loss": 3.5101,
      "step": 13500
    },
    {
      "epoch": 0.272,
      "grad_norm": 0.16740451753139496,
      "learning_rate": 0.0017424510906426853,
      "loss": 3.5021,
      "step": 13600
    },
    {
      "epoch": 0.274,
      "grad_norm": 0.1432366967201233,
      "learning_rate": 0.0017380042447964414,
      "loss": 3.4915,
      "step": 13700
    },
    {
      "epoch": 0.276,
      "grad_norm": 0.14669281244277954,
      "learning_rate": 0.0017335251162537279,
      "loss": 3.4975,
      "step": 13800
    },
    {
      "epoch": 0.278,
      "grad_norm": 0.15456849336624146,
      "learning_rate": 0.0017290139009461377,
      "loss": 3.504,
      "step": 13900
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.163591668009758,
      "learning_rate": 0.0017244707962088424,
      "loss": 3.4899,
      "step": 14000
    },
    {
      "epoch": 0.28,
      "eval_accuracy": 0.366587084148728,
      "eval_loss": 3.4621543884277344,
      "eval_runtime": 13.6803,
      "eval_samples_per_second": 73.098,
      "eval_steps_per_second": 0.585,
      "step": 14000
    },
    {
      "epoch": 0.282,
      "grad_norm": 0.1512901335954666,
      "learning_rate": 0.0017198960007719611,
      "loss": 3.4777,
      "step": 14100
    },
    {
      "epoch": 0.284,
      "grad_norm": 0.13871945440769196,
      "learning_rate": 0.0017152897147518665,
      "loss": 3.5067,
      "step": 14200
    },
    {
      "epoch": 0.286,
      "grad_norm": 0.15246713161468506,
      "learning_rate": 0.001710652139642431,
      "loss": 3.5018,
      "step": 14300
    },
    {
      "epoch": 0.288,
      "grad_norm": 0.13609080016613007,
      "learning_rate": 0.0017059834783062142,
      "loss": 3.4762,
      "step": 14400
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.15287351608276367,
      "learning_rate": 0.0017012839349655868,
      "loss": 3.4872,
      "step": 14500
    },
    {
      "epoch": 0.292,
      "grad_norm": 0.14040616154670715,
      "learning_rate": 0.001696553715193799,
      "loss": 3.5042,
      "step": 14600
    },
    {
      "epoch": 0.294,
      "grad_norm": 0.1571398377418518,
      "learning_rate": 0.0016917930259059879,
      "loss": 3.4744,
      "step": 14700
    },
    {
      "epoch": 0.296,
      "grad_norm": 0.1505391150712967,
      "learning_rate": 0.001687002075350125,
      "loss": 3.4749,
      "step": 14800
    },
    {
      "epoch": 0.298,
      "grad_norm": 0.14757487177848816,
      "learning_rate": 0.001682181073097908,
      "loss": 3.4852,
      "step": 14900
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.15271714329719543,
      "learning_rate": 0.0016773302300355935,
      "loss": 3.486,
      "step": 15000
    },
    {
      "epoch": 0.3,
      "eval_accuracy": 0.36803326810176124,
      "eval_loss": 3.4481563568115234,
      "eval_runtime": 12.0908,
      "eval_samples_per_second": 82.707,
      "eval_steps_per_second": 0.662,
      "step": 15000
    },
    {
      "epoch": 0.302,
      "grad_norm": 0.15327778458595276,
      "learning_rate": 0.0016724497583547715,
      "loss": 3.478,
      "step": 15100
    },
    {
      "epoch": 0.304,
      "grad_norm": 0.16108167171478271,
      "learning_rate": 0.0016675398715430842,
      "loss": 3.4579,
      "step": 15200
    },
    {
      "epoch": 0.306,
      "grad_norm": 0.15231835842132568,
      "learning_rate": 0.001662600784374886,
      "loss": 3.4819,
      "step": 15300
    },
    {
      "epoch": 0.308,
      "grad_norm": 0.1505262553691864,
      "learning_rate": 0.0016576327129018504,
      "loss": 3.4808,
      "step": 15400
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.15296302735805511,
      "learning_rate": 0.0016526358744435184,
      "loss": 3.459,
      "step": 15500
    },
    {
      "epoch": 0.312,
      "grad_norm": 0.2003161907196045,
      "learning_rate": 0.001647610487577791,
      "loss": 3.4751,
      "step": 15600
    },
    {
      "epoch": 0.314,
      "grad_norm": 0.130265474319458,
      "learning_rate": 0.0016425567721313694,
      "loss": 3.4819,
      "step": 15700
    },
    {
      "epoch": 0.316,
      "grad_norm": 0.14476525783538818,
      "learning_rate": 0.0016374749491701401,
      "loss": 3.4514,
      "step": 15800
    },
    {
      "epoch": 0.318,
      "grad_norm": 0.1549970805644989,
      "learning_rate": 0.0016323652409895013,
      "loss": 3.4684,
      "step": 15900
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.14814923703670502,
      "learning_rate": 0.0016272278711046424,
      "loss": 3.4674,
      "step": 16000
    },
    {
      "epoch": 0.32,
      "eval_accuracy": 0.36914090019569473,
      "eval_loss": 3.435351848602295,
      "eval_runtime": 12.375,
      "eval_samples_per_second": 80.808,
      "eval_steps_per_second": 0.646,
      "step": 16000
    },
    {
      "epoch": 0.322,
      "grad_norm": 0.207662433385849,
      "learning_rate": 0.001622063064240765,
      "loss": 3.4661,
      "step": 16100
    },
    {
      "epoch": 0.324,
      "grad_norm": 0.14697088301181793,
      "learning_rate": 0.001616871046323253,
      "loss": 3.4735,
      "step": 16200
    },
    {
      "epoch": 0.326,
      "grad_norm": 0.16137182712554932,
      "learning_rate": 0.0016116520444677894,
      "loss": 3.451,
      "step": 16300
    },
    {
      "epoch": 0.328,
      "grad_norm": 0.1839095950126648,
      "learning_rate": 0.001606406286970423,
      "loss": 3.4646,
      "step": 16400
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.1534615308046341,
      "learning_rate": 0.0016011340032975805,
      "loss": 3.4666,
      "step": 16500
    },
    {
      "epoch": 0.332,
      "grad_norm": 0.15798258781433105,
      "learning_rate": 0.00159583542407603,
      "loss": 3.4418,
      "step": 16600
    },
    {
      "epoch": 0.334,
      "grad_norm": 0.13258863985538483,
      "learning_rate": 0.001590510781082791,
      "loss": 3.4641,
      "step": 16700
    },
    {
      "epoch": 0.336,
      "grad_norm": 0.14669552445411682,
      "learning_rate": 0.001585160307234998,
      "loss": 3.4667,
      "step": 16800
    },
    {
      "epoch": 0.338,
      "grad_norm": 0.15622150897979736,
      "learning_rate": 0.00157978423657971,
      "loss": 3.439,
      "step": 16900
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.18012835085391998,
      "learning_rate": 0.0015743828042836733,
      "loss": 3.4512,
      "step": 17000
    },
    {
      "epoch": 0.34,
      "eval_accuracy": 0.3706555772994129,
      "eval_loss": 3.4220447540283203,
      "eval_runtime": 21.6548,
      "eval_samples_per_second": 46.179,
      "eval_steps_per_second": 0.369,
      "step": 17000
    },
    {
      "epoch": 0.342,
      "grad_norm": 0.143183171749115,
      "learning_rate": 0.0015689562466230352,
      "loss": 3.4531,
      "step": 17100
    },
    {
      "epoch": 0.344,
      "grad_norm": 0.15830625593662262,
      "learning_rate": 0.0015635048009730072,
      "loss": 3.4425,
      "step": 17200
    },
    {
      "epoch": 0.346,
      "grad_norm": 0.1351785659790039,
      "learning_rate": 0.0015580287057974822,
      "loss": 3.4491,
      "step": 17300
    },
    {
      "epoch": 0.348,
      "grad_norm": 0.1572723239660263,
      "learning_rate": 0.0015525282006386032,
      "loss": 3.4329,
      "step": 17400
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.16905032098293304,
      "learning_rate": 0.0015470035261062852,
      "loss": 3.4457,
      "step": 17500
    },
    {
      "epoch": 0.352,
      "grad_norm": 0.1542733907699585,
      "learning_rate": 0.0015414549238676899,
      "loss": 3.4466,
      "step": 17600
    },
    {
      "epoch": 0.354,
      "grad_norm": 0.17319922149181366,
      "learning_rate": 0.0015358826366366532,
      "loss": 3.4311,
      "step": 17700
    },
    {
      "epoch": 0.356,
      "grad_norm": 0.17198143899440765,
      "learning_rate": 0.0015302869081630717,
      "loss": 3.4453,
      "step": 17800
    },
    {
      "epoch": 0.358,
      "grad_norm": 0.15461337566375732,
      "learning_rate": 0.0015246679832222353,
      "loss": 3.4539,
      "step": 17900
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.15917198359966278,
      "learning_rate": 0.0015190261076041245,
      "loss": 3.4188,
      "step": 18000
    },
    {
      "epoch": 0.36,
      "eval_accuracy": 0.3723365949119374,
      "eval_loss": 3.4081637859344482,
      "eval_runtime": 12.6994,
      "eval_samples_per_second": 78.744,
      "eval_steps_per_second": 0.63,
      "step": 18000
    },
    {
      "epoch": 0.362,
      "grad_norm": 0.15682971477508545,
      "learning_rate": 0.0015133615281026555,
      "loss": 3.426,
      "step": 18100
    },
    {
      "epoch": 0.364,
      "grad_norm": 0.15073339641094208,
      "learning_rate": 0.0015076744925048868,
      "loss": 3.4358,
      "step": 18200
    },
    {
      "epoch": 0.366,
      "grad_norm": 0.18040859699249268,
      "learning_rate": 0.0015019652495801786,
      "loss": 3.4261,
      "step": 18300
    },
    {
      "epoch": 0.368,
      "grad_norm": 0.18911828100681305,
      "learning_rate": 0.0014962340490693121,
      "loss": 3.4359,
      "step": 18400
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.15816016495227814,
      "learning_rate": 0.0014904811416735636,
      "loss": 3.4216,
      "step": 18500
    },
    {
      "epoch": 0.372,
      "grad_norm": 0.17654721438884735,
      "learning_rate": 0.0014847067790437396,
      "loss": 3.4347,
      "step": 18600
    },
    {
      "epoch": 0.374,
      "grad_norm": 0.17527590692043304,
      "learning_rate": 0.0014789112137691682,
      "loss": 3.4256,
      "step": 18700
    },
    {
      "epoch": 0.376,
      "grad_norm": 0.15412954986095428,
      "learning_rate": 0.001473094699366649,
      "loss": 3.4074,
      "step": 18800
    },
    {
      "epoch": 0.378,
      "grad_norm": 0.1891617327928543,
      "learning_rate": 0.0014672574902693661,
      "loss": 3.4338,
      "step": 18900
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.16060151159763336,
      "learning_rate": 0.001461399841815754,
      "loss": 3.4372,
      "step": 19000
    },
    {
      "epoch": 0.38,
      "eval_accuracy": 0.3734951076320939,
      "eval_loss": 3.3982865810394287,
      "eval_runtime": 12.1716,
      "eval_samples_per_second": 82.158,
      "eval_steps_per_second": 0.657,
      "step": 19000
    },
    {
      "epoch": 0.382,
      "grad_norm": 0.17226573824882507,
      "learning_rate": 0.0014555220102383326,
      "loss": 3.4123,
      "step": 19100
    },
    {
      "epoch": 0.384,
      "grad_norm": 0.15569022297859192,
      "learning_rate": 0.0014496242526524964,
      "loss": 3.4226,
      "step": 19200
    },
    {
      "epoch": 0.386,
      "grad_norm": 0.16819970309734344,
      "learning_rate": 0.001443706827045269,
      "loss": 3.4294,
      "step": 19300
    },
    {
      "epoch": 0.388,
      "grad_norm": 0.14779409766197205,
      "learning_rate": 0.0014377699922640157,
      "loss": 3.4233,
      "step": 19400
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.1541208177804947,
      "learning_rate": 0.0014318140080051228,
      "loss": 3.4087,
      "step": 19500
    },
    {
      "epoch": 0.392,
      "grad_norm": 0.1646755486726761,
      "learning_rate": 0.0014258391348026362,
      "loss": 3.398,
      "step": 19600
    },
    {
      "epoch": 0.394,
      "grad_norm": 0.18935076892375946,
      "learning_rate": 0.0014198456340168662,
      "loss": 3.4231,
      "step": 19700
    },
    {
      "epoch": 0.396,
      "grad_norm": 0.1867913156747818,
      "learning_rate": 0.0014138337678229528,
      "loss": 3.4129,
      "step": 19800
    },
    {
      "epoch": 0.398,
      "grad_norm": 0.15837714076042175,
      "learning_rate": 0.0014078037991993992,
      "loss": 3.3952,
      "step": 19900
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.15315821766853333,
      "learning_rate": 0.0014017559919165675,
      "loss": 3.4155,
      "step": 20000
    },
    {
      "epoch": 0.4,
      "eval_accuracy": 0.37439138943248534,
      "eval_loss": 3.387746572494507,
      "eval_runtime": 12.7148,
      "eval_samples_per_second": 78.648,
      "eval_steps_per_second": 0.629,
      "step": 20000
    },
    {
      "epoch": 0.402,
      "grad_norm": 0.18359500169754028,
      "learning_rate": 0.0013956906105251406,
      "loss": 3.4094,
      "step": 20100
    },
    {
      "epoch": 0.404,
      "grad_norm": 0.1576140820980072,
      "learning_rate": 0.0013896079203445497,
      "loss": 3.3935,
      "step": 20200
    },
    {
      "epoch": 0.406,
      "grad_norm": 0.22549428045749664,
      "learning_rate": 0.0013835081874513679,
      "loss": 3.4006,
      "step": 20300
    },
    {
      "epoch": 0.408,
      "grad_norm": 0.15331390500068665,
      "learning_rate": 0.001377391678667673,
      "loss": 3.4129,
      "step": 20400
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.16119052469730377,
      "learning_rate": 0.0013712586615493736,
      "loss": 3.4036,
      "step": 20500
    },
    {
      "epoch": 0.412,
      "grad_norm": 0.15083439648151398,
      "learning_rate": 0.0013651094043745067,
      "loss": 3.3857,
      "step": 20600
    },
    {
      "epoch": 0.414,
      "grad_norm": 0.14811307191848755,
      "learning_rate": 0.0013589441761315021,
      "loss": 3.4058,
      "step": 20700
    },
    {
      "epoch": 0.416,
      "grad_norm": 0.15205927193164825,
      "learning_rate": 0.0013527632465074157,
      "loss": 3.4109,
      "step": 20800
    },
    {
      "epoch": 0.418,
      "grad_norm": 0.164772629737854,
      "learning_rate": 0.0013465668858761326,
      "loss": 3.3811,
      "step": 20900
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.15865328907966614,
      "learning_rate": 0.00134035536528654,
      "loss": 3.3878,
      "step": 21000
    },
    {
      "epoch": 0.42,
      "eval_accuracy": 0.37497260273972605,
      "eval_loss": 3.3792359828948975,
      "eval_runtime": 12.1681,
      "eval_samples_per_second": 82.182,
      "eval_steps_per_second": 0.657,
      "step": 21000
    },
    {
      "epoch": 0.422,
      "grad_norm": 0.18703462183475494,
      "learning_rate": 0.0013341289564506717,
      "loss": 3.401,
      "step": 21100
    },
    {
      "epoch": 0.424,
      "grad_norm": 0.15354931354522705,
      "learning_rate": 0.0013278879317318202,
      "loss": 3.3947,
      "step": 21200
    },
    {
      "epoch": 0.426,
      "grad_norm": 0.14713342487812042,
      "learning_rate": 0.0013216325641326255,
      "loss": 3.3774,
      "step": 21300
    },
    {
      "epoch": 0.428,
      "grad_norm": 0.17548827826976776,
      "learning_rate": 0.0013153631272831304,
      "loss": 3.3794,
      "step": 21400
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.1897333562374115,
      "learning_rate": 0.001309079895428813,
      "loss": 3.3955,
      "step": 21500
    },
    {
      "epoch": 0.432,
      "grad_norm": 0.1576170176267624,
      "learning_rate": 0.0013027831434185898,
      "loss": 3.3884,
      "step": 21600
    },
    {
      "epoch": 0.434,
      "grad_norm": 0.18555521965026855,
      "learning_rate": 0.0012964731466927916,
      "loss": 3.3695,
      "step": 21700
    },
    {
      "epoch": 0.436,
      "grad_norm": 0.15987864136695862,
      "learning_rate": 0.0012901501812711176,
      "loss": 3.3891,
      "step": 21800
    },
    {
      "epoch": 0.438,
      "grad_norm": 0.16719584167003632,
      "learning_rate": 0.0012838145237405588,
      "loss": 3.4056,
      "step": 21900
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.19731645286083221,
      "learning_rate": 0.0012774664512432996,
      "loss": 3.3772,
      "step": 22000
    },
    {
      "epoch": 0.44,
      "eval_accuracy": 0.3771624266144814,
      "eval_loss": 3.3640236854553223,
      "eval_runtime": 12.4106,
      "eval_samples_per_second": 80.576,
      "eval_steps_per_second": 0.645,
      "step": 22000
    }
  ],
  "logging_steps": 100,
  "max_steps": 50000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 2000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": null
}