{
  "best_metric": 0.15788726458515429,
  "best_model_checkpoint": "checkpoints/checkpoint-8800",
  "epoch": 4.637143519591931,
  "eval_steps": 50,
  "global_step": 10000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.023185717597959656,
      "grad_norm": 0.16052097082138062,
      "learning_rate": 5e-05,
      "loss": 0.6225,
      "step": 50
    },
    {
      "epoch": 0.023185717597959656,
      "eval_loss": 0.1987911110084725,
      "eval_runtime": 63.5433,
      "eval_samples_per_second": 654.451,
      "eval_steps_per_second": 0.645,
      "step": 50
    },
    {
      "epoch": 0.04637143519591931,
      "grad_norm": 0.09532159566879272,
      "learning_rate": 0.0001,
      "loss": 0.1508,
      "step": 100
    },
    {
      "epoch": 0.04637143519591931,
      "eval_loss": 0.18357936446787168,
      "eval_runtime": 60.9844,
      "eval_samples_per_second": 681.912,
      "eval_steps_per_second": 0.672,
      "step": 100
    },
    {
      "epoch": 0.06955715279387897,
      "grad_norm": 0.24056212604045868,
      "learning_rate": 9.999370638369377e-05,
      "loss": 0.1449,
      "step": 150
    },
    {
      "epoch": 0.06955715279387897,
      "eval_loss": 0.17892658896642444,
      "eval_runtime": 60.7834,
      "eval_samples_per_second": 684.167,
      "eval_steps_per_second": 0.675,
      "step": 150
    },
    {
      "epoch": 0.09274287039183862,
      "grad_norm": 0.09350813180208206,
      "learning_rate": 9.997482711915927e-05,
      "loss": 0.1421,
      "step": 200
    },
    {
      "epoch": 0.09274287039183862,
      "eval_loss": 0.17624869869175752,
      "eval_runtime": 60.3826,
      "eval_samples_per_second": 688.708,
      "eval_steps_per_second": 0.679,
      "step": 200
    },
    {
      "epoch": 0.11592858798979828,
      "grad_norm": 0.12230529636144638,
      "learning_rate": 9.99433669591504e-05,
      "loss": 0.141,
      "step": 250
    },
    {
      "epoch": 0.11592858798979828,
      "eval_loss": 0.17641382363047173,
      "eval_runtime": 60.4169,
      "eval_samples_per_second": 688.317,
      "eval_steps_per_second": 0.679,
      "step": 250
    },
    {
      "epoch": 0.13911430558775795,
      "grad_norm": 0.14592748880386353,
      "learning_rate": 9.989933382359422e-05,
      "loss": 0.1397,
      "step": 300
    },
    {
      "epoch": 0.13911430558775795,
      "eval_loss": 0.17552215792639078,
      "eval_runtime": 61.6101,
      "eval_samples_per_second": 674.987,
      "eval_steps_per_second": 0.665,
      "step": 300
    },
    {
      "epoch": 0.1623000231857176,
      "grad_norm": 0.10219988226890564,
      "learning_rate": 9.984273879759713e-05,
      "loss": 0.1393,
      "step": 350
    },
    {
      "epoch": 0.1623000231857176,
      "eval_loss": 0.17414749172793012,
      "eval_runtime": 61.4962,
      "eval_samples_per_second": 676.237,
      "eval_steps_per_second": 0.667,
      "step": 350
    },
    {
      "epoch": 0.18548574078367724,
      "grad_norm": 0.12338168174028397,
      "learning_rate": 9.977359612865423e-05,
      "loss": 0.1388,
      "step": 400
    },
    {
      "epoch": 0.18548574078367724,
      "eval_loss": 0.17378012638412807,
      "eval_runtime": 61.0462,
      "eval_samples_per_second": 681.221,
      "eval_steps_per_second": 0.672,
      "step": 400
    },
    {
      "epoch": 0.20867145838163692,
      "grad_norm": 0.09479879587888718,
      "learning_rate": 9.969192322306271e-05,
      "loss": 0.1394,
      "step": 450
    },
    {
      "epoch": 0.20867145838163692,
      "eval_loss": 0.17252362204688398,
      "eval_runtime": 60.963,
      "eval_samples_per_second": 682.151,
      "eval_steps_per_second": 0.673,
      "step": 450
    },
    {
      "epoch": 0.23185717597959657,
      "grad_norm": 0.1108623668551445,
      "learning_rate": 9.959774064153977e-05,
      "loss": 0.1383,
      "step": 500
    },
    {
      "epoch": 0.23185717597959657,
      "eval_loss": 0.17298176916843877,
      "eval_runtime": 60.6546,
      "eval_samples_per_second": 685.62,
      "eval_steps_per_second": 0.676,
      "step": 500
    },
    {
      "epoch": 0.2550428935775562,
      "grad_norm": 0.0725204199552536,
      "learning_rate": 9.949107209404665e-05,
      "loss": 0.1376,
      "step": 550
    },
    {
      "epoch": 0.2550428935775562,
      "eval_loss": 0.17165218165539878,
      "eval_runtime": 59.4888,
      "eval_samples_per_second": 699.056,
      "eval_steps_per_second": 0.689,
      "step": 550
    },
    {
      "epoch": 0.2782286111755159,
      "grad_norm": 0.0955963134765625,
      "learning_rate": 9.937194443381972e-05,
      "loss": 0.1372,
      "step": 600
    },
    {
      "epoch": 0.2782286111755159,
      "eval_loss": 0.17077083113718278,
      "eval_runtime": 60.6021,
      "eval_samples_per_second": 686.214,
      "eval_steps_per_second": 0.677,
      "step": 600
    },
    {
      "epoch": 0.3014143287734755,
      "grad_norm": 0.18736732006072998,
      "learning_rate": 9.924038765061042e-05,
      "loss": 0.1361,
      "step": 650
    },
    {
      "epoch": 0.3014143287734755,
      "eval_loss": 0.1727738813183492,
      "eval_runtime": 60.3343,
      "eval_samples_per_second": 689.259,
      "eval_steps_per_second": 0.68,
      "step": 650
    },
    {
      "epoch": 0.3246000463714352,
      "grad_norm": 0.09572151303291321,
      "learning_rate": 9.909643486313533e-05,
      "loss": 0.1362,
      "step": 700
    },
    {
      "epoch": 0.3246000463714352,
      "eval_loss": 0.17145407115151273,
      "eval_runtime": 60.2732,
      "eval_samples_per_second": 689.959,
      "eval_steps_per_second": 0.68,
      "step": 700
    },
    {
      "epoch": 0.34778576396939487,
      "grad_norm": 0.07214252650737762,
      "learning_rate": 9.894012231073894e-05,
      "loss": 0.1364,
      "step": 750
    },
    {
      "epoch": 0.34778576396939487,
      "eval_loss": 0.17133199408489355,
      "eval_runtime": 60.0148,
      "eval_samples_per_second": 692.929,
      "eval_steps_per_second": 0.683,
      "step": 750
    },
    {
      "epoch": 0.3709714815673545,
      "grad_norm": 0.18224318325519562,
      "learning_rate": 9.877148934427037e-05,
      "loss": 0.1356,
      "step": 800
    },
    {
      "epoch": 0.3709714815673545,
      "eval_loss": 0.16949569222888886,
      "eval_runtime": 60.1491,
      "eval_samples_per_second": 691.382,
      "eval_steps_per_second": 0.682,
      "step": 800
    },
    {
      "epoch": 0.39415719916531416,
      "grad_norm": 0.06306415796279907,
      "learning_rate": 9.859057841617709e-05,
      "loss": 0.1353,
      "step": 850
    },
    {
      "epoch": 0.39415719916531416,
      "eval_loss": 0.1686952690798172,
      "eval_runtime": 60.6447,
      "eval_samples_per_second": 685.731,
      "eval_steps_per_second": 0.676,
      "step": 850
    },
    {
      "epoch": 0.41734291676327384,
      "grad_norm": 0.10090287029743195,
      "learning_rate": 9.839743506981782e-05,
      "loss": 0.1361,
      "step": 900
    },
    {
      "epoch": 0.41734291676327384,
      "eval_loss": 0.17026100034088926,
      "eval_runtime": 61.5224,
      "eval_samples_per_second": 675.949,
      "eval_steps_per_second": 0.666,
      "step": 900
    },
    {
      "epoch": 0.44052863436123346,
      "grad_norm": 0.10061236470937729,
      "learning_rate": 9.819210792799712e-05,
      "loss": 0.1354,
      "step": 950
    },
    {
      "epoch": 0.44052863436123346,
      "eval_loss": 0.16971544565694113,
      "eval_runtime": 60.488,
      "eval_samples_per_second": 687.508,
      "eval_steps_per_second": 0.678,
      "step": 950
    },
    {
      "epoch": 0.46371435195919314,
      "grad_norm": 0.06525534391403198,
      "learning_rate": 9.797464868072488e-05,
      "loss": 0.1352,
      "step": 1000
    },
    {
      "epoch": 0.46371435195919314,
      "eval_loss": 0.16946903195393553,
      "eval_runtime": 61.3558,
      "eval_samples_per_second": 677.784,
      "eval_steps_per_second": 0.668,
      "step": 1000
    },
    {
      "epoch": 0.4869000695571528,
      "grad_norm": 0.06269507855176926,
      "learning_rate": 9.77451120722037e-05,
      "loss": 0.1335,
      "step": 1050
    },
    {
      "epoch": 0.4869000695571528,
      "eval_loss": 0.16825352947444114,
      "eval_runtime": 60.1971,
      "eval_samples_per_second": 690.831,
      "eval_steps_per_second": 0.681,
      "step": 1050
    },
    {
      "epoch": 0.5100857871551124,
      "grad_norm": 0.08187470585107803,
      "learning_rate": 9.750355588704727e-05,
      "loss": 0.1327,
      "step": 1100
    },
    {
      "epoch": 0.5100857871551124,
      "eval_loss": 0.16861737282523587,
      "eval_runtime": 59.5715,
      "eval_samples_per_second": 698.085,
      "eval_steps_per_second": 0.688,
      "step": 1100
    },
    {
      "epoch": 0.5332715047530721,
      "grad_norm": 0.06607680767774582,
      "learning_rate": 9.725004093573342e-05,
      "loss": 0.1337,
      "step": 1150
    },
    {
      "epoch": 0.5332715047530721,
      "eval_loss": 0.1692070748762034,
      "eval_runtime": 60.0637,
      "eval_samples_per_second": 692.364,
      "eval_steps_per_second": 0.683,
      "step": 1150
    },
    {
      "epoch": 0.5564572223510318,
      "grad_norm": 0.09759815782308578,
      "learning_rate": 9.698463103929542e-05,
      "loss": 0.134,
      "step": 1200
    },
    {
      "epoch": 0.5564572223510318,
      "eval_loss": 0.16649228385381692,
      "eval_runtime": 60.4305,
      "eval_samples_per_second": 688.162,
      "eval_steps_per_second": 0.678,
      "step": 1200
    },
    {
      "epoch": 0.5796429399489914,
      "grad_norm": 0.10353852063417435,
      "learning_rate": 9.670739301325534e-05,
      "loss": 0.1341,
      "step": 1250
    },
    {
      "epoch": 0.5796429399489914,
      "eval_loss": 0.16802514322459206,
      "eval_runtime": 60.0955,
      "eval_samples_per_second": 691.999,
      "eval_steps_per_second": 0.682,
      "step": 1250
    },
    {
      "epoch": 0.602828657546951,
      "grad_norm": 0.11834366619586945,
      "learning_rate": 9.641839665080363e-05,
      "loss": 0.1347,
      "step": 1300
    },
    {
      "epoch": 0.602828657546951,
      "eval_loss": 0.1672302417427292,
      "eval_runtime": 60.3484,
      "eval_samples_per_second": 689.098,
      "eval_steps_per_second": 0.679,
      "step": 1300
    },
    {
      "epoch": 0.6260143751449108,
      "grad_norm": 0.06963012367486954,
      "learning_rate": 9.611771470522908e-05,
      "loss": 0.1335,
      "step": 1350
    },
    {
      "epoch": 0.6260143751449108,
      "eval_loss": 0.16607839684977216,
      "eval_runtime": 60.3308,
      "eval_samples_per_second": 689.3,
      "eval_steps_per_second": 0.68,
      "step": 1350
    },
    {
      "epoch": 0.6492000927428704,
      "grad_norm": 0.06842990219593048,
      "learning_rate": 9.580542287160348e-05,
      "loss": 0.1338,
      "step": 1400
    },
    {
      "epoch": 0.6492000927428704,
      "eval_loss": 0.16628812684035693,
      "eval_runtime": 59.9335,
      "eval_samples_per_second": 693.87,
      "eval_steps_per_second": 0.684,
      "step": 1400
    },
    {
      "epoch": 0.67238581034083,
      "grad_norm": 0.07053674757480621,
      "learning_rate": 9.548159976772592e-05,
      "loss": 0.1335,
      "step": 1450
    },
    {
      "epoch": 0.67238581034083,
      "eval_loss": 0.16696060882262428,
      "eval_runtime": 59.8079,
      "eval_samples_per_second": 695.326,
      "eval_steps_per_second": 0.686,
      "step": 1450
    },
    {
      "epoch": 0.6955715279387897,
      "grad_norm": 0.09175281971693039,
      "learning_rate": 9.514632691433107e-05,
      "loss": 0.1332,
      "step": 1500
    },
    {
      "epoch": 0.6955715279387897,
      "eval_loss": 0.16521949465081834,
      "eval_runtime": 60.1856,
      "eval_samples_per_second": 690.963,
      "eval_steps_per_second": 0.681,
      "step": 1500
    },
    {
      "epoch": 0.7187572455367494,
      "grad_norm": 0.05836635082960129,
      "learning_rate": 9.479968871456679e-05,
      "loss": 0.1336,
      "step": 1550
    },
    {
      "epoch": 0.7187572455367494,
      "eval_loss": 0.16626366255041727,
      "eval_runtime": 60.6256,
      "eval_samples_per_second": 685.948,
      "eval_steps_per_second": 0.676,
      "step": 1550
    },
    {
      "epoch": 0.741942963134709,
      "grad_norm": 0.07249301671981812,
      "learning_rate": 9.444177243274618e-05,
      "loss": 0.133,
      "step": 1600
    },
    {
      "epoch": 0.741942963134709,
      "eval_loss": 0.1655649439629329,
      "eval_runtime": 60.2447,
      "eval_samples_per_second": 690.285,
      "eval_steps_per_second": 0.681,
      "step": 1600
    },
    {
      "epoch": 0.7651286807326687,
      "grad_norm": 0.07509302347898483,
      "learning_rate": 9.407266817237911e-05,
      "loss": 0.1332,
      "step": 1650
    },
    {
      "epoch": 0.7651286807326687,
      "eval_loss": 0.16605371203296967,
      "eval_runtime": 59.8196,
      "eval_samples_per_second": 695.191,
      "eval_steps_per_second": 0.685,
      "step": 1650
    },
    {
      "epoch": 0.7883143983306283,
      "grad_norm": 0.07540406286716461,
      "learning_rate": 9.369246885348926e-05,
      "loss": 0.1327,
      "step": 1700
    },
    {
      "epoch": 0.7883143983306283,
      "eval_loss": 0.16555590021301406,
      "eval_runtime": 60.4119,
      "eval_samples_per_second": 688.374,
      "eval_steps_per_second": 0.679,
      "step": 1700
    },
    {
      "epoch": 0.811500115928588,
      "grad_norm": 0.06061087176203728,
      "learning_rate": 9.330127018922194e-05,
      "loss": 0.1318,
      "step": 1750
    },
    {
      "epoch": 0.811500115928588,
      "eval_loss": 0.16623179673527624,
      "eval_runtime": 59.7807,
      "eval_samples_per_second": 695.643,
      "eval_steps_per_second": 0.686,
      "step": 1750
    },
    {
      "epoch": 0.8346858335265477,
      "grad_norm": 0.05577518790960312,
      "learning_rate": 9.289917066174886e-05,
      "loss": 0.1319,
      "step": 1800
    },
    {
      "epoch": 0.8346858335265477,
      "eval_loss": 0.16519989030959317,
      "eval_runtime": 60.1508,
      "eval_samples_per_second": 691.363,
      "eval_steps_per_second": 0.682,
      "step": 1800
    },
    {
      "epoch": 0.8578715511245073,
      "grad_norm": 0.06929640471935272,
      "learning_rate": 9.248627149747573e-05,
      "loss": 0.1337,
      "step": 1850
    },
    {
      "epoch": 0.8578715511245073,
      "eval_loss": 0.16394849849125304,
      "eval_runtime": 60.1044,
      "eval_samples_per_second": 691.896,
      "eval_steps_per_second": 0.682,
      "step": 1850
    },
    {
      "epoch": 0.8810572687224669,
      "grad_norm": 0.07941466569900513,
      "learning_rate": 9.206267664155907e-05,
      "loss": 0.1324,
      "step": 1900
    },
    {
      "epoch": 0.8810572687224669,
      "eval_loss": 0.1648257591054525,
      "eval_runtime": 59.9818,
      "eval_samples_per_second": 693.31,
      "eval_steps_per_second": 0.684,
      "step": 1900
    },
    {
      "epoch": 0.9042429863204267,
      "grad_norm": 0.09700328856706619,
      "learning_rate": 9.162849273173857e-05,
      "loss": 0.1334,
      "step": 1950
    },
    {
      "epoch": 0.9042429863204267,
      "eval_loss": 0.16508159820082235,
      "eval_runtime": 60.0956,
      "eval_samples_per_second": 691.997,
      "eval_steps_per_second": 0.682,
      "step": 1950
    },
    {
      "epoch": 0.9274287039183863,
      "grad_norm": 0.09397923946380615,
      "learning_rate": 9.118382907149165e-05,
      "loss": 0.1317,
      "step": 2000
    },
    {
      "epoch": 0.9274287039183863,
      "eval_loss": 0.16377645660046958,
      "eval_runtime": 60.471,
      "eval_samples_per_second": 687.701,
      "eval_steps_per_second": 0.678,
      "step": 2000
    },
    {
      "epoch": 0.9506144215163459,
      "grad_norm": 0.08097202330827713,
      "learning_rate": 9.072879760251679e-05,
      "loss": 0.1324,
      "step": 2050
    },
    {
      "epoch": 0.9506144215163459,
      "eval_loss": 0.16491611914973717,
      "eval_runtime": 60.6678,
      "eval_samples_per_second": 685.471,
      "eval_steps_per_second": 0.676,
      "step": 2050
    },
    {
      "epoch": 0.9738001391143056,
      "grad_norm": 0.08455361425876617,
      "learning_rate": 9.026351287655294e-05,
      "loss": 0.1326,
      "step": 2100
    },
    {
      "epoch": 0.9738001391143056,
      "eval_loss": 0.16602741997032858,
      "eval_runtime": 60.5593,
      "eval_samples_per_second": 686.698,
      "eval_steps_per_second": 0.677,
      "step": 2100
    },
    {
      "epoch": 0.9969858567122653,
      "grad_norm": 0.056316621601581573,
      "learning_rate": 8.978809202654162e-05,
      "loss": 0.1326,
      "step": 2150
    },
    {
      "epoch": 0.9969858567122653,
      "eval_loss": 0.1640188218462461,
      "eval_runtime": 60.9228,
      "eval_samples_per_second": 682.602,
      "eval_steps_per_second": 0.673,
      "step": 2150
    },
    {
      "epoch": 1.0201715743102249,
      "grad_norm": 0.06686601787805557,
      "learning_rate": 8.930265473713938e-05,
      "loss": 0.132,
      "step": 2200
    },
    {
      "epoch": 1.0201715743102249,
      "eval_loss": 0.1652621257294944,
      "eval_runtime": 60.883,
      "eval_samples_per_second": 683.048,
      "eval_steps_per_second": 0.673,
      "step": 2200
    },
    {
      "epoch": 1.0433572919081846,
      "grad_norm": 0.040202509611845016,
      "learning_rate": 8.880732321458784e-05,
      "loss": 0.1319,
      "step": 2250
    },
    {
      "epoch": 1.0433572919081846,
      "eval_loss": 0.1655291575008717,
      "eval_runtime": 60.3109,
      "eval_samples_per_second": 689.527,
      "eval_steps_per_second": 0.68,
      "step": 2250
    },
    {
      "epoch": 1.0665430095061441,
      "grad_norm": 0.0656428411602974,
      "learning_rate": 8.83022221559489e-05,
      "loss": 0.1326,
      "step": 2300
    },
    {
      "epoch": 1.0665430095061441,
      "eval_loss": 0.16431572407087935,
      "eval_runtime": 60.1036,
      "eval_samples_per_second": 691.906,
      "eval_steps_per_second": 0.682,
      "step": 2300
    },
    {
      "epoch": 1.0897287271041038,
      "grad_norm": 0.06945247948169708,
      "learning_rate": 8.778747871771292e-05,
      "loss": 0.1321,
      "step": 2350
    },
    {
      "epoch": 1.0897287271041038,
      "eval_loss": 0.16585482329987242,
      "eval_runtime": 60.6263,
      "eval_samples_per_second": 685.94,
      "eval_steps_per_second": 0.676,
      "step": 2350
    },
    {
      "epoch": 1.1129144447020636,
      "grad_norm": 0.0523492731153965,
      "learning_rate": 8.726322248378775e-05,
      "loss": 0.1317,
      "step": 2400
    },
    {
      "epoch": 1.1129144447020636,
      "eval_loss": 0.16438524923736036,
      "eval_runtime": 60.317,
      "eval_samples_per_second": 689.457,
      "eval_steps_per_second": 0.68,
      "step": 2400
    },
    {
      "epoch": 1.136100162300023,
      "grad_norm": 0.07777334004640579,
      "learning_rate": 8.672958543287666e-05,
      "loss": 0.1322,
      "step": 2450
    },
    {
      "epoch": 1.136100162300023,
      "eval_loss": 0.16509696053644565,
      "eval_runtime": 60.26,
      "eval_samples_per_second": 690.109,
      "eval_steps_per_second": 0.68,
      "step": 2450
    },
    {
      "epoch": 1.1592858798979828,
      "grad_norm": 0.06430637836456299,
      "learning_rate": 8.618670190525352e-05,
      "loss": 0.1325,
      "step": 2500
    },
    {
      "epoch": 1.1592858798979828,
      "eval_loss": 0.1639541608445008,
      "eval_runtime": 60.5314,
      "eval_samples_per_second": 687.015,
      "eval_steps_per_second": 0.677,
      "step": 2500
    },
    {
      "epoch": 1.1824715974959426,
      "grad_norm": 0.11194106936454773,
      "learning_rate": 8.563470856894316e-05,
      "loss": 0.1311,
      "step": 2550
    },
    {
      "epoch": 1.1824715974959426,
      "eval_loss": 0.16260699934317355,
      "eval_runtime": 60.3659,
      "eval_samples_per_second": 688.899,
      "eval_steps_per_second": 0.679,
      "step": 2550
    },
    {
      "epoch": 1.205657315093902,
      "grad_norm": 0.06165901944041252,
      "learning_rate": 8.507374438531607e-05,
      "loss": 0.1323,
      "step": 2600
    },
    {
      "epoch": 1.205657315093902,
      "eval_loss": 0.1626319663130242,
      "eval_runtime": 59.9516,
      "eval_samples_per_second": 693.66,
      "eval_steps_per_second": 0.684,
      "step": 2600
    },
    {
      "epoch": 1.2288430326918618,
      "grad_norm": 0.10654885321855545,
      "learning_rate": 8.450395057410561e-05,
      "loss": 0.1316,
      "step": 2650
    },
    {
      "epoch": 1.2288430326918618,
      "eval_loss": 0.16393000041041636,
      "eval_runtime": 59.576,
      "eval_samples_per_second": 698.032,
      "eval_steps_per_second": 0.688,
      "step": 2650
    },
    {
      "epoch": 1.2520287502898215,
      "grad_norm": 0.04848140478134155,
      "learning_rate": 8.392547057785661e-05,
      "loss": 0.1314,
      "step": 2700
    },
    {
      "epoch": 1.2520287502898215,
      "eval_loss": 0.16348152455768114,
      "eval_runtime": 60.098,
      "eval_samples_per_second": 691.97,
      "eval_steps_per_second": 0.682,
      "step": 2700
    },
    {
      "epoch": 1.275214467887781,
      "grad_norm": 0.0573604516685009,
      "learning_rate": 8.333845002581458e-05,
      "loss": 0.1314,
      "step": 2750
    },
    {
      "epoch": 1.275214467887781,
      "eval_loss": 0.16364089140116167,
      "eval_runtime": 60.1364,
      "eval_samples_per_second": 691.528,
      "eval_steps_per_second": 0.682,
      "step": 2750
    },
    {
      "epoch": 1.2984001854857408,
      "grad_norm": 0.053159259259700775,
      "learning_rate": 8.274303669726426e-05,
      "loss": 0.131,
      "step": 2800
    },
    {
      "epoch": 1.2984001854857408,
      "eval_loss": 0.16257415365129801,
      "eval_runtime": 60.0025,
      "eval_samples_per_second": 693.071,
      "eval_steps_per_second": 0.683,
      "step": 2800
    },
    {
      "epoch": 1.3215859030837005,
      "grad_norm": 0.09136148542165756,
      "learning_rate": 8.213938048432697e-05,
      "loss": 0.1313,
      "step": 2850
    },
    {
      "epoch": 1.3215859030837005,
      "eval_loss": 0.16324665471619784,
      "eval_runtime": 59.8429,
      "eval_samples_per_second": 694.92,
      "eval_steps_per_second": 0.685,
      "step": 2850
    },
    {
      "epoch": 1.34477162068166,
      "grad_norm": 0.05825324356555939,
      "learning_rate": 8.152763335422613e-05,
      "loss": 0.1312,
      "step": 2900
    },
    {
      "epoch": 1.34477162068166,
      "eval_loss": 0.16367374608121235,
      "eval_runtime": 60.219,
      "eval_samples_per_second": 690.579,
      "eval_steps_per_second": 0.681,
      "step": 2900
    },
    {
      "epoch": 1.3679573382796197,
      "grad_norm": 0.06379790604114532,
      "learning_rate": 8.090794931103026e-05,
      "loss": 0.1317,
      "step": 2950
    },
    {
      "epoch": 1.3679573382796197,
      "eval_loss": 0.16400733758786312,
      "eval_runtime": 59.9641,
      "eval_samples_per_second": 693.515,
      "eval_steps_per_second": 0.684,
      "step": 2950
    },
    {
      "epoch": 1.3911430558775795,
      "grad_norm": 0.05361103266477585,
      "learning_rate": 8.028048435688333e-05,
      "loss": 0.1311,
      "step": 3000
    },
    {
      "epoch": 1.3911430558775795,
      "eval_loss": 0.16210626991928834,
      "eval_runtime": 59.5858,
      "eval_samples_per_second": 697.919,
      "eval_steps_per_second": 0.688,
      "step": 3000
    },
    {
      "epoch": 1.414328773475539,
      "grad_norm": 0.04593402519822121,
      "learning_rate": 7.964539645273204e-05,
      "loss": 0.1304,
      "step": 3050
    },
    {
      "epoch": 1.414328773475539,
      "eval_loss": 0.163067463275087,
      "eval_runtime": 60.098,
      "eval_samples_per_second": 691.97,
      "eval_steps_per_second": 0.682,
      "step": 3050
    },
    {
      "epoch": 1.4375144910734987,
      "grad_norm": 0.057480327785015106,
      "learning_rate": 7.900284547855991e-05,
      "loss": 0.1307,
      "step": 3100
    },
    {
      "epoch": 1.4375144910734987,
      "eval_loss": 0.16243572043734797,
      "eval_runtime": 59.5674,
      "eval_samples_per_second": 698.133,
      "eval_steps_per_second": 0.688,
      "step": 3100
    },
    {
      "epoch": 1.4607002086714584,
      "grad_norm": 0.08223798871040344,
      "learning_rate": 7.835299319313853e-05,
      "loss": 0.1315,
      "step": 3150
    },
    {
      "epoch": 1.4607002086714584,
      "eval_loss": 0.1641944734489707,
      "eval_runtime": 59.5423,
      "eval_samples_per_second": 698.428,
      "eval_steps_per_second": 0.689,
      "step": 3150
    },
    {
      "epoch": 1.483885926269418,
      "grad_norm": 0.09742949903011322,
      "learning_rate": 7.769600319330552e-05,
      "loss": 0.1303,
      "step": 3200
    },
    {
      "epoch": 1.483885926269418,
      "eval_loss": 0.16355698856626613,
      "eval_runtime": 60.1946,
      "eval_samples_per_second": 690.859,
      "eval_steps_per_second": 0.681,
      "step": 3200
    },
    {
      "epoch": 1.5070716438673777,
      "grad_norm": 0.06401767581701279,
      "learning_rate": 7.703204087277988e-05,
      "loss": 0.1315,
      "step": 3250
    },
    {
      "epoch": 1.5070716438673777,
      "eval_loss": 0.16215006705140952,
      "eval_runtime": 59.7822,
      "eval_samples_per_second": 695.625,
      "eval_steps_per_second": 0.686,
      "step": 3250
    },
    {
      "epoch": 1.5302573614653374,
      "grad_norm": 0.07916898280382156,
      "learning_rate": 7.636127338052512e-05,
      "loss": 0.1315,
      "step": 3300
    },
    {
      "epoch": 1.5302573614653374,
      "eval_loss": 0.16288597734760557,
      "eval_runtime": 59.2757,
      "eval_samples_per_second": 701.57,
      "eval_steps_per_second": 0.692,
      "step": 3300
    },
    {
      "epoch": 1.553443079063297,
      "grad_norm": 0.06549016386270523,
      "learning_rate": 7.568386957867033e-05,
      "loss": 0.1303,
      "step": 3350
    },
    {
      "epoch": 1.553443079063297,
      "eval_loss": 0.16416664097655873,
      "eval_runtime": 59.84,
      "eval_samples_per_second": 694.953,
      "eval_steps_per_second": 0.685,
      "step": 3350
    },
    {
      "epoch": 1.5766287966612567,
      "grad_norm": 0.0709395632147789,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.1309,
      "step": 3400
    },
    {
      "epoch": 1.5766287966612567,
      "eval_loss": 0.16179194486424098,
      "eval_runtime": 59.8634,
      "eval_samples_per_second": 694.682,
      "eval_steps_per_second": 0.685,
      "step": 3400
    },
    {
      "epoch": 1.5998145142592164,
      "grad_norm": 0.05671363323926926,
      "learning_rate": 7.430983680502344e-05,
      "loss": 0.1307,
      "step": 3450
    },
    {
      "epoch": 1.5998145142592164,
      "eval_loss": 0.16309191886303373,
      "eval_runtime": 59.618,
      "eval_samples_per_second": 697.541,
      "eval_steps_per_second": 0.688,
      "step": 3450
    },
    {
      "epoch": 1.623000231857176,
      "grad_norm": 0.04889162629842758,
      "learning_rate": 7.361355373863414e-05,
      "loss": 0.1314,
      "step": 3500
    },
    {
      "epoch": 1.623000231857176,
      "eval_loss": 0.16290782983414598,
      "eval_runtime": 60.3904,
      "eval_samples_per_second": 688.619,
      "eval_steps_per_second": 0.679,
      "step": 3500
    },
    {
      "epoch": 1.6461859494551356,
      "grad_norm": 0.0970933735370636,
      "learning_rate": 7.291132608637052e-05,
      "loss": 0.1314,
      "step": 3550
    },
    {
      "epoch": 1.6461859494551356,
      "eval_loss": 0.16278222993823557,
      "eval_runtime": 59.8666,
      "eval_samples_per_second": 694.644,
      "eval_steps_per_second": 0.685,
      "step": 3550
    },
    {
      "epoch": 1.6693716670530954,
      "grad_norm": 0.056557025760412216,
      "learning_rate": 7.220333063028872e-05,
      "loss": 0.1312,
      "step": 3600
    },
    {
      "epoch": 1.6693716670530954,
      "eval_loss": 0.16313205291311117,
      "eval_runtime": 60.0092,
      "eval_samples_per_second": 692.993,
      "eval_steps_per_second": 0.683,
      "step": 3600
    },
    {
      "epoch": 1.6925573846510549,
      "grad_norm": 0.04870522394776344,
      "learning_rate": 7.148974560445859e-05,
      "loss": 0.1299,
      "step": 3650
    },
    {
      "epoch": 1.6925573846510549,
      "eval_loss": 0.1617941082289122,
      "eval_runtime": 60.1721,
      "eval_samples_per_second": 691.117,
      "eval_steps_per_second": 0.681,
      "step": 3650
    },
    {
      "epoch": 1.7157431022490146,
      "grad_norm": 0.0681833028793335,
      "learning_rate": 7.077075065009433e-05,
      "loss": 0.1304,
      "step": 3700
    },
    {
      "epoch": 1.7157431022490146,
      "eval_loss": 0.16243406602519425,
      "eval_runtime": 59.3626,
      "eval_samples_per_second": 700.542,
      "eval_steps_per_second": 0.691,
      "step": 3700
    },
    {
      "epoch": 1.7389288198469743,
      "grad_norm": 0.06506156921386719,
      "learning_rate": 7.004652677033068e-05,
      "loss": 0.1299,
      "step": 3750
    },
    {
      "epoch": 1.7389288198469743,
      "eval_loss": 0.16324780134312317,
      "eval_runtime": 59.6022,
      "eval_samples_per_second": 697.726,
      "eval_steps_per_second": 0.688,
      "step": 3750
    },
    {
      "epoch": 1.7621145374449338,
      "grad_norm": 0.06188170611858368,
      "learning_rate": 6.931725628465643e-05,
      "loss": 0.1309,
      "step": 3800
    },
    {
      "epoch": 1.7621145374449338,
      "eval_loss": 0.1623115342294882,
      "eval_runtime": 59.7694,
      "eval_samples_per_second": 695.774,
      "eval_steps_per_second": 0.686,
      "step": 3800
    },
    {
      "epoch": 1.7853002550428936,
      "grad_norm": 0.05675831064581871,
      "learning_rate": 6.858312278301637e-05,
      "loss": 0.1303,
      "step": 3850
    },
    {
      "epoch": 1.7853002550428936,
      "eval_loss": 0.1630547638293529,
      "eval_runtime": 59.779,
      "eval_samples_per_second": 695.662,
      "eval_steps_per_second": 0.686,
      "step": 3850
    },
    {
      "epoch": 1.8084859726408533,
      "grad_norm": 0.04727062210440636,
      "learning_rate": 6.784431107959359e-05,
      "loss": 0.1312,
      "step": 3900
    },
    {
      "epoch": 1.8084859726408533,
      "eval_loss": 0.1616409071893626,
      "eval_runtime": 59.6005,
      "eval_samples_per_second": 697.746,
      "eval_steps_per_second": 0.688,
      "step": 3900
    },
    {
      "epoch": 1.8316716902388128,
      "grad_norm": 0.06378892064094543,
      "learning_rate": 6.710100716628344e-05,
      "loss": 0.1303,
      "step": 3950
    },
    {
      "epoch": 1.8316716902388128,
      "eval_loss": 0.1622395658739077,
      "eval_runtime": 60.1499,
      "eval_samples_per_second": 691.373,
      "eval_steps_per_second": 0.682,
      "step": 3950
    },
    {
      "epoch": 1.8548574078367726,
      "grad_norm": 0.05470576509833336,
      "learning_rate": 6.635339816587109e-05,
      "loss": 0.1308,
      "step": 4000
    },
    {
      "epoch": 1.8548574078367726,
      "eval_loss": 0.16317236170181762,
      "eval_runtime": 60.014,
      "eval_samples_per_second": 692.939,
      "eval_steps_per_second": 0.683,
      "step": 4000
    },
    {
      "epoch": 1.8780431254347323,
      "grad_norm": 0.053886763751506805,
      "learning_rate": 6.560167228492436e-05,
      "loss": 0.1297,
      "step": 4050
    },
    {
      "epoch": 1.8780431254347323,
      "eval_loss": 0.16198886262197804,
      "eval_runtime": 60.8262,
      "eval_samples_per_second": 683.685,
      "eval_steps_per_second": 0.674,
      "step": 4050
    },
    {
      "epoch": 1.9012288430326918,
      "grad_norm": 0.054583676159381866,
      "learning_rate": 6.484601876641375e-05,
      "loss": 0.1301,
      "step": 4100
    },
    {
      "epoch": 1.9012288430326918,
      "eval_loss": 0.1616550050294764,
      "eval_runtime": 59.7779,
      "eval_samples_per_second": 695.675,
      "eval_steps_per_second": 0.686,
      "step": 4100
    },
    {
      "epoch": 1.9244145606306515,
      "grad_norm": 0.071171335875988,
      "learning_rate": 6.408662784207149e-05,
      "loss": 0.131,
      "step": 4150
    },
    {
      "epoch": 1.9244145606306515,
      "eval_loss": 0.15968682813566223,
      "eval_runtime": 60.227,
      "eval_samples_per_second": 690.487,
      "eval_steps_per_second": 0.681,
      "step": 4150
    },
    {
      "epoch": 1.9476002782286113,
      "grad_norm": 0.05775531381368637,
      "learning_rate": 6.332369068450174e-05,
      "loss": 0.1296,
      "step": 4200
    },
    {
      "epoch": 1.9476002782286113,
      "eval_loss": 0.16262199212265846,
      "eval_runtime": 60.3405,
      "eval_samples_per_second": 689.189,
      "eval_steps_per_second": 0.679,
      "step": 4200
    },
    {
      "epoch": 1.9707859958265708,
      "grad_norm": 0.06425776332616806,
      "learning_rate": 6.255739935905396e-05,
      "loss": 0.1299,
      "step": 4250
    },
    {
      "epoch": 1.9707859958265708,
      "eval_loss": 0.16324524366491053,
      "eval_runtime": 61.417,
      "eval_samples_per_second": 677.109,
      "eval_steps_per_second": 0.668,
      "step": 4250
    },
    {
      "epoch": 1.9939717134245305,
      "grad_norm": 0.045762140303850174,
      "learning_rate": 6.178794677547137e-05,
      "loss": 0.1299,
      "step": 4300
    },
    {
      "epoch": 1.9939717134245305,
      "eval_loss": 0.16053301797614244,
      "eval_runtime": 61.0801,
      "eval_samples_per_second": 680.844,
      "eval_steps_per_second": 0.671,
      "step": 4300
    },
    {
      "epoch": 2.0171574310224902,
      "grad_norm": 0.07060451060533524,
      "learning_rate": 6.1015526639327035e-05,
      "loss": 0.1296,
      "step": 4350
    },
    {
      "epoch": 2.0171574310224902,
      "eval_loss": 0.1620254674138633,
      "eval_runtime": 61.0829,
      "eval_samples_per_second": 680.812,
      "eval_steps_per_second": 0.671,
      "step": 4350
    },
    {
      "epoch": 2.0403431486204497,
      "grad_norm": 0.059919316321611404,
      "learning_rate": 6.024033340325954e-05,
      "loss": 0.1302,
      "step": 4400
    },
    {
      "epoch": 2.0403431486204497,
      "eval_loss": 0.16284223807997533,
      "eval_runtime": 61.5789,
      "eval_samples_per_second": 675.328,
      "eval_steps_per_second": 0.666,
      "step": 4400
    },
    {
      "epoch": 2.0635288662184093,
      "grad_norm": 0.07983385026454926,
      "learning_rate": 5.946256221802051e-05,
      "loss": 0.13,
      "step": 4450
    },
    {
      "epoch": 2.0635288662184093,
      "eval_loss": 0.16209282393788932,
      "eval_runtime": 61.6584,
      "eval_samples_per_second": 674.458,
      "eval_steps_per_second": 0.665,
      "step": 4450
    },
    {
      "epoch": 2.086714583816369,
      "grad_norm": 0.07582173496484756,
      "learning_rate": 5.868240888334653e-05,
      "loss": 0.1296,
      "step": 4500
    },
    {
      "epoch": 2.086714583816369,
      "eval_loss": 0.16158196377974565,
      "eval_runtime": 61.1826,
      "eval_samples_per_second": 679.703,
      "eval_steps_per_second": 0.67,
      "step": 4500
    },
    {
      "epoch": 2.1099003014143287,
      "grad_norm": 0.06049995869398117,
      "learning_rate": 5.79000697986675e-05,
      "loss": 0.1298,
      "step": 4550
    },
    {
      "epoch": 2.1099003014143287,
      "eval_loss": 0.16130609279963956,
      "eval_runtime": 61.0153,
      "eval_samples_per_second": 681.567,
      "eval_steps_per_second": 0.672,
      "step": 4550
    },
    {
      "epoch": 2.1330860190122882,
      "grad_norm": 0.0440148264169693,
      "learning_rate": 5.7115741913664264e-05,
      "loss": 0.1299,
      "step": 4600
    },
    {
      "epoch": 2.1330860190122882,
      "eval_loss": 0.16027799763638953,
      "eval_runtime": 61.1993,
      "eval_samples_per_second": 679.517,
      "eval_steps_per_second": 0.67,
      "step": 4600
    },
    {
      "epoch": 2.156271736610248,
      "grad_norm": 0.05254065990447998,
      "learning_rate": 5.6329622678687463e-05,
      "loss": 0.1299,
      "step": 4650
    },
    {
      "epoch": 2.156271736610248,
      "eval_loss": 0.16206274484291652,
      "eval_runtime": 61.4415,
      "eval_samples_per_second": 676.839,
      "eval_steps_per_second": 0.667,
      "step": 4650
    },
    {
      "epoch": 2.1794574542082077,
      "grad_norm": 0.06294432282447815,
      "learning_rate": 5.5541909995050554e-05,
      "loss": 0.1306,
      "step": 4700
    },
    {
      "epoch": 2.1794574542082077,
      "eval_loss": 0.16140170723024802,
      "eval_runtime": 60.8861,
      "eval_samples_per_second": 683.013,
      "eval_steps_per_second": 0.673,
      "step": 4700
    },
    {
      "epoch": 2.202643171806167,
      "grad_norm": 0.06710942089557648,
      "learning_rate": 5.475280216520913e-05,
      "loss": 0.1303,
      "step": 4750
    },
    {
      "epoch": 2.202643171806167,
      "eval_loss": 0.16245448075670843,
      "eval_runtime": 61.2839,
      "eval_samples_per_second": 678.58,
      "eval_steps_per_second": 0.669,
      "step": 4750
    },
    {
      "epoch": 2.225828889404127,
      "grad_norm": 0.05298132076859474,
      "learning_rate": 5.396249784283942e-05,
      "loss": 0.13,
      "step": 4800
    },
    {
      "epoch": 2.225828889404127,
      "eval_loss": 0.1623738898660767,
      "eval_runtime": 61.1531,
      "eval_samples_per_second": 680.031,
      "eval_steps_per_second": 0.67,
      "step": 4800
    },
    {
      "epoch": 2.2490146070020867,
      "grad_norm": 0.04066763445734978,
      "learning_rate": 5.317119598282823e-05,
      "loss": 0.1295,
      "step": 4850
    },
    {
      "epoch": 2.2490146070020867,
      "eval_loss": 0.1627438727811327,
      "eval_runtime": 61.0414,
      "eval_samples_per_second": 681.275,
      "eval_steps_per_second": 0.672,
      "step": 4850
    },
    {
      "epoch": 2.272200324600046,
      "grad_norm": 0.061821240931749344,
      "learning_rate": 5.2379095791187124e-05,
      "loss": 0.1299,
      "step": 4900
    },
    {
      "epoch": 2.272200324600046,
      "eval_loss": 0.16086717177928397,
      "eval_runtime": 60.7945,
      "eval_samples_per_second": 684.042,
      "eval_steps_per_second": 0.674,
      "step": 4900
    },
    {
      "epoch": 2.295386042198006,
      "grad_norm": 0.08038394153118134,
      "learning_rate": 5.158639667490339e-05,
      "loss": 0.13,
      "step": 4950
    },
    {
      "epoch": 2.295386042198006,
      "eval_loss": 0.16221664317086187,
      "eval_runtime": 61.6742,
      "eval_samples_per_second": 674.285,
      "eval_steps_per_second": 0.665,
      "step": 4950
    },
    {
      "epoch": 2.3185717597959656,
      "grad_norm": 0.0556926503777504,
      "learning_rate": 5.0793298191740404e-05,
      "loss": 0.1311,
      "step": 5000
    },
    {
      "epoch": 2.3185717597959656,
      "eval_loss": 0.16015339844546791,
      "eval_runtime": 61.3657,
      "eval_samples_per_second": 677.675,
      "eval_steps_per_second": 0.668,
      "step": 5000
    },
    {
      "epoch": 2.3417574773939256,
      "grad_norm": 0.06645477563142776,
      "learning_rate": 5e-05,
      "loss": 0.1284,
      "step": 5050
    },
    {
      "epoch": 2.3417574773939256,
      "eval_loss": 0.16160674186313023,
      "eval_runtime": 61.4737,
      "eval_samples_per_second": 676.484,
      "eval_steps_per_second": 0.667,
      "step": 5050
    },
    {
      "epoch": 2.364943194991885,
      "grad_norm": 0.05365500971674919,
      "learning_rate": 4.92067018082596e-05,
      "loss": 0.13,
      "step": 5100
    },
    {
      "epoch": 2.364943194991885,
      "eval_loss": 0.16016484459556096,
      "eval_runtime": 61.4058,
      "eval_samples_per_second": 677.232,
      "eval_steps_per_second": 0.668,
      "step": 5100
    },
    {
      "epoch": 2.3881289125898446,
      "grad_norm": 0.0499204620718956,
      "learning_rate": 4.841360332509663e-05,
      "loss": 0.129,
      "step": 5150
    },
    {
      "epoch": 2.3881289125898446,
      "eval_loss": 0.16054727464378063,
      "eval_runtime": 61.1539,
      "eval_samples_per_second": 680.023,
      "eval_steps_per_second": 0.67,
      "step": 5150
    },
    {
      "epoch": 2.411314630187804,
      "grad_norm": 0.07284457236528397,
      "learning_rate": 4.762090420881289e-05,
      "loss": 0.129,
      "step": 5200
    },
    {
      "epoch": 2.411314630187804,
      "eval_loss": 0.16057287778830004,
      "eval_runtime": 60.5785,
      "eval_samples_per_second": 686.481,
      "eval_steps_per_second": 0.677,
      "step": 5200
    },
    {
      "epoch": 2.434500347785764,
      "grad_norm": 0.06511891633272171,
      "learning_rate": 4.6828804017171776e-05,
      "loss": 0.1297,
      "step": 5250
    },
    {
      "epoch": 2.434500347785764,
      "eval_loss": 0.16202011190836896,
      "eval_runtime": 61.4053,
      "eval_samples_per_second": 677.238,
      "eval_steps_per_second": 0.668,
      "step": 5250
    },
    {
      "epoch": 2.4576860653837236,
      "grad_norm": 0.05936937406659126,
      "learning_rate": 4.603750215716057e-05,
      "loss": 0.1293,
      "step": 5300
    },
    {
      "epoch": 2.4576860653837236,
      "eval_loss": 0.16067086041480225,
      "eval_runtime": 60.4469,
      "eval_samples_per_second": 687.976,
      "eval_steps_per_second": 0.678,
      "step": 5300
    },
    {
      "epoch": 2.480871782981683,
      "grad_norm": 0.039836496114730835,
      "learning_rate": 4.5247197834790876e-05,
      "loss": 0.1288,
      "step": 5350
    },
    {
      "epoch": 2.480871782981683,
      "eval_loss": 0.1614640227625451,
      "eval_runtime": 60.9513,
      "eval_samples_per_second": 682.283,
      "eval_steps_per_second": 0.673,
      "step": 5350
    },
    {
      "epoch": 2.504057500579643,
      "grad_norm": 0.04305760934948921,
      "learning_rate": 4.445809000494946e-05,
      "loss": 0.1294,
      "step": 5400
    },
    {
      "epoch": 2.504057500579643,
      "eval_loss": 0.16139181990447046,
      "eval_runtime": 60.6766,
      "eval_samples_per_second": 685.371,
      "eval_steps_per_second": 0.676,
      "step": 5400
    },
    {
      "epoch": 2.5272432181776026,
      "grad_norm": 0.06780368089675903,
      "learning_rate": 4.3670377321312535e-05,
      "loss": 0.1285,
      "step": 5450
    },
    {
      "epoch": 2.5272432181776026,
      "eval_loss": 0.1619736397134425,
      "eval_runtime": 60.7281,
      "eval_samples_per_second": 684.79,
      "eval_steps_per_second": 0.675,
      "step": 5450
    },
    {
      "epoch": 2.550428935775562,
      "grad_norm": 0.052273835986852646,
      "learning_rate": 4.288425808633575e-05,
      "loss": 0.1303,
      "step": 5500
    },
    {
      "epoch": 2.550428935775562,
      "eval_loss": 0.16178818674979198,
      "eval_runtime": 60.8875,
      "eval_samples_per_second": 682.997,
      "eval_steps_per_second": 0.673,
      "step": 5500
    },
    {
      "epoch": 2.573614653373522,
      "grad_norm": 0.045574627816677094,
      "learning_rate": 4.20999302013325e-05,
      "loss": 0.1291,
      "step": 5550
    },
    {
      "epoch": 2.573614653373522,
      "eval_loss": 0.16034006952877458,
      "eval_runtime": 60.7378,
      "eval_samples_per_second": 684.681,
      "eval_steps_per_second": 0.675,
      "step": 5550
    },
    {
      "epoch": 2.5968003709714815,
      "grad_norm": 0.044092051684856415,
      "learning_rate": 4.131759111665349e-05,
      "loss": 0.1298,
      "step": 5600
    },
    {
      "epoch": 2.5968003709714815,
      "eval_loss": 0.16090484909780667,
      "eval_runtime": 60.4675,
      "eval_samples_per_second": 687.741,
      "eval_steps_per_second": 0.678,
      "step": 5600
    },
    {
      "epoch": 2.6199860885694415,
      "grad_norm": 0.05473971739411354,
      "learning_rate": 4.0537437781979506e-05,
      "loss": 0.1288,
      "step": 5650
    },
    {
      "epoch": 2.6199860885694415,
      "eval_loss": 0.1604315377337276,
      "eval_runtime": 62.8239,
      "eval_samples_per_second": 661.946,
      "eval_steps_per_second": 0.653,
      "step": 5650
    },
    {
      "epoch": 2.643171806167401,
      "grad_norm": 0.07100555300712585,
      "learning_rate": 3.9759666596740476e-05,
      "loss": 0.129,
      "step": 5700
    },
    {
      "epoch": 2.643171806167401,
      "eval_loss": 0.15997494100305837,
      "eval_runtime": 61.3008,
      "eval_samples_per_second": 678.392,
      "eval_steps_per_second": 0.669,
      "step": 5700
    },
    {
      "epoch": 2.6663575237653605,
      "grad_norm": 0.04020215570926666,
      "learning_rate": 3.898447336067297e-05,
      "loss": 0.1291,
      "step": 5750
    },
    {
      "epoch": 2.6663575237653605,
      "eval_loss": 0.1596748490832133,
      "eval_runtime": 60.6148,
      "eval_samples_per_second": 686.07,
      "eval_steps_per_second": 0.676,
      "step": 5750
    },
    {
      "epoch": 2.68954324136332,
      "grad_norm": 0.05526584014296532,
      "learning_rate": 3.821205322452863e-05,
      "loss": 0.1291,
      "step": 5800
    },
    {
      "epoch": 2.68954324136332,
      "eval_loss": 0.16091962633426782,
      "eval_runtime": 60.1717,
      "eval_samples_per_second": 691.122,
      "eval_steps_per_second": 0.681,
      "step": 5800
    },
    {
      "epoch": 2.71272895896128,
      "grad_norm": 0.052167922258377075,
      "learning_rate": 3.744260064094604e-05,
      "loss": 0.129,
      "step": 5850
    },
    {
      "epoch": 2.71272895896128,
      "eval_loss": 0.16112806362615253,
      "eval_runtime": 60.1273,
      "eval_samples_per_second": 691.633,
      "eval_steps_per_second": 0.682,
      "step": 5850
    },
    {
      "epoch": 2.7359146765592395,
      "grad_norm": 0.054320793598890305,
      "learning_rate": 3.6676309315498256e-05,
      "loss": 0.13,
      "step": 5900
    },
    {
      "epoch": 2.7359146765592395,
      "eval_loss": 0.15996250695505343,
      "eval_runtime": 60.655,
      "eval_samples_per_second": 685.616,
      "eval_steps_per_second": 0.676,
      "step": 5900
    },
    {
      "epoch": 2.7591003941571994,
      "grad_norm": 0.05470626428723335,
      "learning_rate": 3.591337215792852e-05,
      "loss": 0.1296,
      "step": 5950
    },
    {
      "epoch": 2.7591003941571994,
      "eval_loss": 0.16025288890609335,
      "eval_runtime": 60.826,
      "eval_samples_per_second": 683.688,
      "eval_steps_per_second": 0.674,
      "step": 5950
    },
    {
      "epoch": 2.782286111755159,
      "grad_norm": 0.04805810749530792,
      "learning_rate": 3.515398123358627e-05,
      "loss": 0.1294,
      "step": 6000
    },
    {
      "epoch": 2.782286111755159,
      "eval_loss": 0.15918263724182835,
      "eval_runtime": 60.2321,
      "eval_samples_per_second": 690.429,
      "eval_steps_per_second": 0.681,
      "step": 6000
    },
    {
      "epoch": 2.8054718293531185,
      "grad_norm": 0.04185302183032036,
      "learning_rate": 3.439832771507565e-05,
      "loss": 0.1283,
      "step": 6050
    },
    {
      "epoch": 2.8054718293531185,
      "eval_loss": 0.16179385240233157,
      "eval_runtime": 60.9176,
      "eval_samples_per_second": 682.66,
      "eval_steps_per_second": 0.673,
      "step": 6050
    },
    {
      "epoch": 2.828657546951078,
      "grad_norm": 0.04609336704015732,
      "learning_rate": 3.364660183412892e-05,
      "loss": 0.1292,
      "step": 6100
    },
    {
      "epoch": 2.828657546951078,
      "eval_loss": 0.1611929898635588,
      "eval_runtime": 60.5916,
      "eval_samples_per_second": 686.333,
      "eval_steps_per_second": 0.677,
      "step": 6100
    },
    {
      "epoch": 2.851843264549038,
      "grad_norm": 0.05404876172542572,
      "learning_rate": 3.289899283371657e-05,
      "loss": 0.128,
      "step": 6150
    },
    {
      "epoch": 2.851843264549038,
      "eval_loss": 0.16039360794951976,
      "eval_runtime": 60.5961,
      "eval_samples_per_second": 686.282,
      "eval_steps_per_second": 0.677,
      "step": 6150
    },
    {
      "epoch": 2.8750289821469974,
      "grad_norm": 0.06787659227848053,
      "learning_rate": 3.215568892040641e-05,
      "loss": 0.1288,
      "step": 6200
    },
    {
      "epoch": 2.8750289821469974,
      "eval_loss": 0.16113480515361805,
      "eval_runtime": 60.2775,
      "eval_samples_per_second": 689.909,
      "eval_steps_per_second": 0.68,
      "step": 6200
    },
    {
      "epoch": 2.8982146997449574,
      "grad_norm": 0.06937435269355774,
      "learning_rate": 3.141687721698363e-05,
      "loss": 0.1283,
      "step": 6250
    },
    {
      "epoch": 2.8982146997449574,
      "eval_loss": 0.16087572214972407,
      "eval_runtime": 60.6789,
      "eval_samples_per_second": 685.345,
      "eval_steps_per_second": 0.676,
      "step": 6250
    },
    {
      "epoch": 2.921400417342917,
      "grad_norm": 0.08074232190847397,
      "learning_rate": 3.0682743715343564e-05,
      "loss": 0.1292,
      "step": 6300
    },
    {
      "epoch": 2.921400417342917,
      "eval_loss": 0.16049740787316144,
      "eval_runtime": 60.3194,
      "eval_samples_per_second": 689.43,
      "eval_steps_per_second": 0.68,
      "step": 6300
    },
    {
      "epoch": 2.9445861349408764,
      "grad_norm": 0.03976515680551529,
      "learning_rate": 2.9953473229669328e-05,
      "loss": 0.1302,
      "step": 6350
    },
    {
      "epoch": 2.9445861349408764,
      "eval_loss": 0.16023700059761273,
      "eval_runtime": 60.8537,
      "eval_samples_per_second": 683.377,
      "eval_steps_per_second": 0.674,
      "step": 6350
    },
    {
      "epoch": 2.967771852538836,
      "grad_norm": 0.05303976684808731,
      "learning_rate": 2.9229249349905684e-05,
      "loss": 0.1285,
      "step": 6400
    },
    {
      "epoch": 2.967771852538836,
      "eval_loss": 0.1601465398516622,
      "eval_runtime": 60.6472,
      "eval_samples_per_second": 685.703,
      "eval_steps_per_second": 0.676,
      "step": 6400
    },
    {
      "epoch": 2.990957570136796,
      "grad_norm": 0.0519745759665966,
      "learning_rate": 2.851025439554142e-05,
      "loss": 0.1286,
      "step": 6450
    },
    {
      "epoch": 2.990957570136796,
      "eval_loss": 0.16085429229133483,
      "eval_runtime": 60.2507,
      "eval_samples_per_second": 690.216,
      "eval_steps_per_second": 0.68,
      "step": 6450
    },
    {
      "epoch": 3.0141432877347554,
      "grad_norm": 0.050518251955509186,
      "learning_rate": 2.7796669369711294e-05,
      "loss": 0.1301,
      "step": 6500
    },
    {
      "epoch": 3.0141432877347554,
      "eval_loss": 0.16015394660421692,
      "eval_runtime": 60.5015,
      "eval_samples_per_second": 687.355,
      "eval_steps_per_second": 0.678,
      "step": 6500
    },
    {
      "epoch": 3.037329005332715,
      "grad_norm": 0.04253960773348808,
      "learning_rate": 2.708867391362948e-05,
      "loss": 0.1296,
      "step": 6550
    },
    {
      "epoch": 3.037329005332715,
      "eval_loss": 0.1597283595131218,
      "eval_runtime": 60.13,
      "eval_samples_per_second": 691.601,
      "eval_steps_per_second": 0.682,
      "step": 6550
    },
    {
      "epoch": 3.060514722930675,
      "grad_norm": 0.06899340450763702,
      "learning_rate": 2.638644626136587e-05,
      "loss": 0.1291,
      "step": 6600
    },
    {
      "epoch": 3.060514722930675,
      "eval_loss": 0.1604277250117246,
      "eval_runtime": 60.4618,
      "eval_samples_per_second": 687.806,
      "eval_steps_per_second": 0.678,
      "step": 6600
    },
    {
      "epoch": 3.0837004405286343,
      "grad_norm": 0.06556117534637451,
      "learning_rate": 2.5690163194976575e-05,
      "loss": 0.1288,
      "step": 6650
    },
    {
      "epoch": 3.0837004405286343,
      "eval_loss": 0.15953636330193482,
      "eval_runtime": 60.2757,
      "eval_samples_per_second": 689.93,
      "eval_steps_per_second": 0.68,
      "step": 6650
    },
    {
      "epoch": 3.106886158126594,
      "grad_norm": 0.03685734421014786,
      "learning_rate": 2.500000000000001e-05,
      "loss": 0.129,
      "step": 6700
    },
    {
      "epoch": 3.106886158126594,
      "eval_loss": 0.159308270335797,
      "eval_runtime": 60.624,
      "eval_samples_per_second": 685.966,
      "eval_steps_per_second": 0.676,
      "step": 6700
    },
    {
      "epoch": 3.130071875724554,
      "grad_norm": 0.0451020672917366,
      "learning_rate": 2.4316130421329697e-05,
      "loss": 0.1286,
      "step": 6750
    },
    {
      "epoch": 3.130071875724554,
      "eval_loss": 0.15995884031774596,
      "eval_runtime": 60.3654,
      "eval_samples_per_second": 688.905,
      "eval_steps_per_second": 0.679,
      "step": 6750
    },
    {
      "epoch": 3.1532575933225133,
      "grad_norm": 0.0495733842253685,
      "learning_rate": 2.363872661947488e-05,
      "loss": 0.1293,
      "step": 6800
    },
    {
      "epoch": 3.1532575933225133,
      "eval_loss": 0.15987331824692497,
      "eval_runtime": 60.4636,
      "eval_samples_per_second": 687.786,
      "eval_steps_per_second": 0.678,
      "step": 6800
    },
    {
      "epoch": 3.176443310920473,
      "grad_norm": 0.05756652355194092,
      "learning_rate": 2.296795912722014e-05,
      "loss": 0.1289,
      "step": 6850
    },
    {
      "epoch": 3.176443310920473,
      "eval_loss": 0.15986134614331013,
      "eval_runtime": 61.0063,
      "eval_samples_per_second": 681.667,
      "eval_steps_per_second": 0.672,
      "step": 6850
    },
    {
      "epoch": 3.199629028518433,
      "grad_norm": 0.0467820018529892,
      "learning_rate": 2.2303996806694488e-05,
      "loss": 0.1295,
      "step": 6900
    },
    {
      "epoch": 3.199629028518433,
      "eval_loss": 0.16011030076900337,
      "eval_runtime": 60.1041,
      "eval_samples_per_second": 691.9,
      "eval_steps_per_second": 0.682,
      "step": 6900
    },
    {
      "epoch": 3.2228147461163923,
      "grad_norm": 0.04179982468485832,
      "learning_rate": 2.164700680686147e-05,
      "loss": 0.1287,
      "step": 6950
    },
    {
      "epoch": 3.2228147461163923,
      "eval_loss": 0.15917751068552838,
      "eval_runtime": 60.5321,
      "eval_samples_per_second": 687.007,
      "eval_steps_per_second": 0.677,
      "step": 6950
    },
    {
      "epoch": 3.246000463714352,
      "grad_norm": 0.053910572081804276,
      "learning_rate": 2.09971545214401e-05,
      "loss": 0.1286,
      "step": 7000
    },
    {
      "epoch": 3.246000463714352,
      "eval_loss": 0.15998092838627764,
      "eval_runtime": 60.4067,
      "eval_samples_per_second": 688.434,
      "eval_steps_per_second": 0.679,
      "step": 7000
    },
    {
      "epoch": 3.2691861813123118,
      "grad_norm": 0.04404950886964798,
      "learning_rate": 2.0354603547267985e-05,
      "loss": 0.1283,
      "step": 7050
    },
    {
      "epoch": 3.2691861813123118,
      "eval_loss": 0.1597617331551387,
      "eval_runtime": 60.4218,
      "eval_samples_per_second": 688.262,
      "eval_steps_per_second": 0.679,
      "step": 7050
    },
    {
      "epoch": 3.2923718989102713,
      "grad_norm": 0.04763752967119217,
      "learning_rate": 1.9719515643116674e-05,
      "loss": 0.1288,
      "step": 7100
    },
    {
      "epoch": 3.2923718989102713,
      "eval_loss": 0.16116006530852447,
      "eval_runtime": 60.2132,
      "eval_samples_per_second": 690.646,
      "eval_steps_per_second": 0.681,
      "step": 7100
    },
    {
      "epoch": 3.3155576165082308,
      "grad_norm": 0.049567196518182755,
      "learning_rate": 1.9092050688969738e-05,
      "loss": 0.1298,
      "step": 7150
    },
    {
      "epoch": 3.3155576165082308,
      "eval_loss": 0.15965543804361845,
      "eval_runtime": 60.3928,
      "eval_samples_per_second": 688.592,
      "eval_steps_per_second": 0.679,
      "step": 7150
    },
    {
      "epoch": 3.3387433341061907,
      "grad_norm": 0.05488676205277443,
      "learning_rate": 1.847236664577389e-05,
      "loss": 0.1284,
      "step": 7200
    },
    {
      "epoch": 3.3387433341061907,
      "eval_loss": 0.16050384662882064,
      "eval_runtime": 60.121,
      "eval_samples_per_second": 691.705,
      "eval_steps_per_second": 0.682,
      "step": 7200
    },
    {
      "epoch": 3.3619290517041502,
      "grad_norm": 0.04124298691749573,
      "learning_rate": 1.7860619515673033e-05,
      "loss": 0.1289,
      "step": 7250
    },
    {
      "epoch": 3.3619290517041502,
      "eval_loss": 0.16054145931691394,
      "eval_runtime": 60.2046,
      "eval_samples_per_second": 690.745,
      "eval_steps_per_second": 0.681,
      "step": 7250
    },
    {
      "epoch": 3.3851147693021097,
      "grad_norm": 0.04400424286723137,
      "learning_rate": 1.725696330273575e-05,
      "loss": 0.1289,
      "step": 7300
    },
    {
      "epoch": 3.3851147693021097,
      "eval_loss": 0.15999099129576416,
      "eval_runtime": 60.4869,
      "eval_samples_per_second": 687.52,
      "eval_steps_per_second": 0.678,
      "step": 7300
    },
    {
      "epoch": 3.4083004869000697,
      "grad_norm": 0.05488509312272072,
      "learning_rate": 1.6661549974185424e-05,
      "loss": 0.1285,
      "step": 7350
    },
    {
      "epoch": 3.4083004869000697,
      "eval_loss": 0.16051823730892306,
      "eval_runtime": 60.1981,
      "eval_samples_per_second": 690.819,
      "eval_steps_per_second": 0.681,
      "step": 7350
    },
    {
      "epoch": 3.431486204498029,
      "grad_norm": 0.06722457706928253,
      "learning_rate": 1.60745294221434e-05,
      "loss": 0.1286,
      "step": 7400
    },
    {
      "epoch": 3.431486204498029,
      "eval_loss": 0.1610307768591294,
      "eval_runtime": 60.7755,
      "eval_samples_per_second": 684.256,
      "eval_steps_per_second": 0.675,
      "step": 7400
    },
    {
      "epoch": 3.4546719220959887,
      "grad_norm": 0.04814394935965538,
      "learning_rate": 1.549604942589441e-05,
      "loss": 0.1278,
      "step": 7450
    },
    {
      "epoch": 3.4546719220959887,
      "eval_loss": 0.1598065741965525,
      "eval_runtime": 59.9968,
      "eval_samples_per_second": 693.136,
      "eval_steps_per_second": 0.683,
      "step": 7450
    },
    {
      "epoch": 3.4778576396939487,
      "grad_norm": 0.04934167116880417,
      "learning_rate": 1.4926255614683932e-05,
      "loss": 0.1274,
      "step": 7500
    },
    {
      "epoch": 3.4778576396939487,
      "eval_loss": 0.15982454893723042,
      "eval_runtime": 60.201,
      "eval_samples_per_second": 690.786,
      "eval_steps_per_second": 0.681,
      "step": 7500
    },
    {
      "epoch": 3.501043357291908,
      "grad_norm": 0.04529615864157677,
      "learning_rate": 1.4365291431056871e-05,
      "loss": 0.1297,
      "step": 7550
    },
    {
      "epoch": 3.501043357291908,
      "eval_loss": 0.15986133524024926,
      "eval_runtime": 59.95,
      "eval_samples_per_second": 693.678,
      "eval_steps_per_second": 0.684,
      "step": 7550
    },
    {
      "epoch": 3.5242290748898677,
      "grad_norm": 0.0399620421230793,
      "learning_rate": 1.3813298094746491e-05,
      "loss": 0.1288,
      "step": 7600
    },
    {
      "epoch": 3.5242290748898677,
      "eval_loss": 0.15905609221590689,
      "eval_runtime": 61.181,
      "eval_samples_per_second": 679.72,
      "eval_steps_per_second": 0.67,
      "step": 7600
    },
    {
      "epoch": 3.5474147924878277,
      "grad_norm": 0.05973295867443085,
      "learning_rate": 1.327041456712334e-05,
      "loss": 0.1281,
      "step": 7650
    },
    {
      "epoch": 3.5474147924878277,
      "eval_loss": 0.15981091550942805,
      "eval_runtime": 60.5605,
      "eval_samples_per_second": 686.685,
      "eval_steps_per_second": 0.677,
      "step": 7650
    },
    {
      "epoch": 3.570600510085787,
      "grad_norm": 0.04896661266684532,
      "learning_rate": 1.2736777516212266e-05,
      "loss": 0.1288,
      "step": 7700
    },
    {
      "epoch": 3.570600510085787,
      "eval_loss": 0.1599924400443614,
      "eval_runtime": 60.486,
      "eval_samples_per_second": 687.531,
      "eval_steps_per_second": 0.678,
      "step": 7700
    },
    {
      "epoch": 3.5937862276837467,
      "grad_norm": 0.07458525151014328,
      "learning_rate": 1.2212521282287092e-05,
      "loss": 0.128,
      "step": 7750
    },
    {
      "epoch": 3.5937862276837467,
      "eval_loss": 0.15936126278835275,
      "eval_runtime": 60.9341,
      "eval_samples_per_second": 682.475,
      "eval_steps_per_second": 0.673,
      "step": 7750
    },
    {
      "epoch": 3.6169719452817066,
      "grad_norm": 0.04200127348303795,
      "learning_rate": 1.1697777844051105e-05,
      "loss": 0.1287,
      "step": 7800
    },
    {
      "epoch": 3.6169719452817066,
      "eval_loss": 0.1603394617678833,
      "eval_runtime": 60.5155,
      "eval_samples_per_second": 687.195,
      "eval_steps_per_second": 0.678,
      "step": 7800
    },
    {
      "epoch": 3.640157662879666,
      "grad_norm": 0.06712640821933746,
      "learning_rate": 1.1192676785412154e-05,
      "loss": 0.1291,
      "step": 7850
    },
    {
      "epoch": 3.640157662879666,
      "eval_loss": 0.15920067938345067,
      "eval_runtime": 60.0225,
      "eval_samples_per_second": 692.84,
      "eval_steps_per_second": 0.683,
      "step": 7850
    },
    {
      "epoch": 3.6633433804776256,
      "grad_norm": 0.049462996423244476,
      "learning_rate": 1.0697345262860636e-05,
      "loss": 0.1287,
      "step": 7900
    },
    {
      "epoch": 3.6633433804776256,
      "eval_loss": 0.15964593569874527,
      "eval_runtime": 60.1965,
      "eval_samples_per_second": 690.837,
      "eval_steps_per_second": 0.681,
      "step": 7900
    },
    {
      "epoch": 3.6865290980755856,
      "grad_norm": 0.05148932337760925,
      "learning_rate": 1.021190797345839e-05,
      "loss": 0.1283,
      "step": 7950
    },
    {
      "epoch": 3.6865290980755856,
      "eval_loss": 0.15903354419354673,
      "eval_runtime": 60.0507,
      "eval_samples_per_second": 692.515,
      "eval_steps_per_second": 0.683,
      "step": 7950
    },
    {
      "epoch": 3.709714815673545,
      "grad_norm": 0.05164024233818054,
      "learning_rate": 9.73648712344707e-06,
      "loss": 0.128,
      "step": 8000
    },
    {
      "epoch": 3.709714815673545,
      "eval_loss": 0.15835035051131605,
      "eval_runtime": 60.5739,
      "eval_samples_per_second": 686.533,
      "eval_steps_per_second": 0.677,
      "step": 8000
    },
    {
      "epoch": 3.7329005332715046,
      "grad_norm": 0.04926716163754463,
      "learning_rate": 9.271202397483215e-06,
      "loss": 0.1276,
      "step": 8050
    },
    {
      "epoch": 3.7329005332715046,
      "eval_loss": 0.160225615529793,
      "eval_runtime": 60.4555,
      "eval_samples_per_second": 687.878,
      "eval_steps_per_second": 0.678,
      "step": 8050
    },
    {
      "epoch": 3.7560862508694646,
      "grad_norm": 0.04355842247605324,
      "learning_rate": 8.816170928508365e-06,
      "loss": 0.1287,
      "step": 8100
    },
    {
      "epoch": 3.7560862508694646,
      "eval_loss": 0.1601867779420742,
      "eval_runtime": 60.7386,
      "eval_samples_per_second": 684.672,
      "eval_steps_per_second": 0.675,
      "step": 8100
    },
    {
      "epoch": 3.779271968467424,
      "grad_norm": 0.039105553179979324,
      "learning_rate": 8.371507268261437e-06,
      "loss": 0.1306,
      "step": 8150
    },
    {
      "epoch": 3.779271968467424,
      "eval_loss": 0.15946348937187382,
      "eval_runtime": 60.9253,
      "eval_samples_per_second": 682.574,
      "eval_steps_per_second": 0.673,
      "step": 8150
    },
    {
      "epoch": 3.8024576860653836,
      "grad_norm": 0.04452899843454361,
      "learning_rate": 7.937323358440935e-06,
      "loss": 0.1286,
      "step": 8200
    },
    {
      "epoch": 3.8024576860653836,
      "eval_loss": 0.15871429728364056,
      "eval_runtime": 60.2776,
      "eval_samples_per_second": 689.908,
      "eval_steps_per_second": 0.68,
      "step": 8200
    },
    {
      "epoch": 3.8256434036633435,
      "grad_norm": 0.043075498193502426,
      "learning_rate": 7.513728502524286e-06,
      "loss": 0.1292,
      "step": 8250
    },
    {
      "epoch": 3.8256434036633435,
      "eval_loss": 0.1592580359542711,
      "eval_runtime": 60.7244,
      "eval_samples_per_second": 684.832,
      "eval_steps_per_second": 0.675,
      "step": 8250
    },
    {
      "epoch": 3.848829121261303,
      "grad_norm": 0.05848800390958786,
      "learning_rate": 7.100829338251147e-06,
      "loss": 0.1275,
      "step": 8300
    },
    {
      "epoch": 3.848829121261303,
      "eval_loss": 0.15895083163665807,
      "eval_runtime": 60.3677,
      "eval_samples_per_second": 688.878,
      "eval_steps_per_second": 0.679,
      "step": 8300
    },
    {
      "epoch": 3.8720148388592626,
      "grad_norm": 0.04980336129665375,
      "learning_rate": 6.698729810778065e-06,
      "loss": 0.1277,
      "step": 8350
    },
    {
      "epoch": 3.8720148388592626,
      "eval_loss": 0.16002303550437,
      "eval_runtime": 60.2742,
      "eval_samples_per_second": 689.947,
      "eval_steps_per_second": 0.68,
      "step": 8350
    },
    {
      "epoch": 3.8952005564572225,
      "grad_norm": 0.057385146617889404,
      "learning_rate": 6.3075311465107535e-06,
      "loss": 0.129,
      "step": 8400
    },
    {
      "epoch": 3.8952005564572225,
      "eval_loss": 0.1601535826416112,
      "eval_runtime": 60.4053,
      "eval_samples_per_second": 688.45,
      "eval_steps_per_second": 0.679,
      "step": 8400
    },
    {
      "epoch": 3.918386274055182,
      "grad_norm": 0.045788682997226715,
      "learning_rate": 5.927331827620903e-06,
      "loss": 0.1286,
      "step": 8450
    },
    {
      "epoch": 3.918386274055182,
      "eval_loss": 0.15926720973175468,
      "eval_runtime": 60.6783,
      "eval_samples_per_second": 685.352,
      "eval_steps_per_second": 0.676,
      "step": 8450
    },
    {
      "epoch": 3.9415719916531415,
      "grad_norm": 0.045575451105833054,
      "learning_rate": 5.558227567253832e-06,
      "loss": 0.1281,
      "step": 8500
    },
    {
      "epoch": 3.9415719916531415,
      "eval_loss": 0.16032033338606583,
      "eval_runtime": 60.4563,
      "eval_samples_per_second": 687.868,
      "eval_steps_per_second": 0.678,
      "step": 8500
    },
    {
      "epoch": 3.964757709251101,
      "grad_norm": 0.034972067922353745,
      "learning_rate": 5.200311285433213e-06,
      "loss": 0.1285,
      "step": 8550
    },
    {
      "epoch": 3.964757709251101,
      "eval_loss": 0.1590997571686103,
      "eval_runtime": 60.7642,
      "eval_samples_per_second": 684.384,
      "eval_steps_per_second": 0.675,
      "step": 8550
    },
    {
      "epoch": 3.987943426849061,
      "grad_norm": 0.05060684680938721,
      "learning_rate": 4.853673085668947e-06,
      "loss": 0.1293,
      "step": 8600
    },
    {
      "epoch": 3.987943426849061,
      "eval_loss": 0.15924322809570868,
      "eval_runtime": 60.0799,
      "eval_samples_per_second": 692.178,
      "eval_steps_per_second": 0.682,
      "step": 8600
    },
    {
      "epoch": 4.011129144447021,
      "grad_norm": 0.04898017644882202,
      "learning_rate": 4.5184002322740785e-06,
      "loss": 0.1283,
      "step": 8650
    },
    {
      "epoch": 4.011129144447021,
      "eval_loss": 0.1587491140112498,
      "eval_runtime": 60.6393,
      "eval_samples_per_second": 685.793,
      "eval_steps_per_second": 0.676,
      "step": 8650
    },
    {
      "epoch": 4.0343148620449805,
      "grad_norm": 0.058361586183309555,
      "learning_rate": 4.19457712839652e-06,
      "loss": 0.1277,
      "step": 8700
    },
    {
      "epoch": 4.0343148620449805,
      "eval_loss": 0.1597737118597627,
      "eval_runtime": 61.5486,
      "eval_samples_per_second": 675.661,
      "eval_steps_per_second": 0.666,
      "step": 8700
    },
    {
      "epoch": 4.05750057964294,
      "grad_norm": 0.05138258635997772,
      "learning_rate": 3.8822852947709375e-06,
      "loss": 0.1283,
      "step": 8750
    },
    {
      "epoch": 4.05750057964294,
      "eval_loss": 0.15985116115580386,
      "eval_runtime": 60.5634,
      "eval_samples_per_second": 686.652,
      "eval_steps_per_second": 0.677,
      "step": 8750
    },
    {
      "epoch": 4.0806862972408995,
      "grad_norm": 0.0461881086230278,
      "learning_rate": 3.581603349196372e-06,
      "loss": 0.1288,
      "step": 8800
    },
    {
      "epoch": 4.0806862972408995,
      "eval_loss": 0.15788726458515429,
      "eval_runtime": 60.6057,
      "eval_samples_per_second": 686.173,
      "eval_steps_per_second": 0.677,
      "step": 8800
    },
    {
      "epoch": 4.103872014838859,
      "grad_norm": 0.0618111789226532,
      "learning_rate": 3.2926069867446675e-06,
      "loss": 0.1287,
      "step": 8850
    },
    {
      "epoch": 4.103872014838859,
      "eval_loss": 0.15881183094974458,
      "eval_runtime": 60.3747,
      "eval_samples_per_second": 688.799,
      "eval_steps_per_second": 0.679,
      "step": 8850
    },
    {
      "epoch": 4.1270577324368185,
      "grad_norm": 0.04804789274930954,
      "learning_rate": 3.0153689607045845e-06,
      "loss": 0.1294,
      "step": 8900
    },
    {
      "epoch": 4.1270577324368185,
      "eval_loss": 0.1607356553004913,
      "eval_runtime": 60.6979,
      "eval_samples_per_second": 685.131,
      "eval_steps_per_second": 0.675,
      "step": 8900
    },
    {
      "epoch": 4.150243450034779,
      "grad_norm": 0.04835003986954689,
      "learning_rate": 2.7499590642665774e-06,
      "loss": 0.1277,
      "step": 8950
    },
    {
      "epoch": 4.150243450034779,
      "eval_loss": 0.1598761189516689,
      "eval_runtime": 61.2003,
      "eval_samples_per_second": 679.507,
      "eval_steps_per_second": 0.67,
      "step": 8950
    },
    {
      "epoch": 4.173429167632738,
      "grad_norm": 0.05750919133424759,
      "learning_rate": 2.496444112952734e-06,
      "loss": 0.1285,
      "step": 9000
    },
    {
      "epoch": 4.173429167632738,
      "eval_loss": 0.15946166188972705,
      "eval_runtime": 60.6795,
      "eval_samples_per_second": 685.339,
      "eval_steps_per_second": 0.676,
      "step": 9000
    },
    {
      "epoch": 4.196614885230698,
      "grad_norm": 0.06801807135343552,
      "learning_rate": 2.2548879277963064e-06,
      "loss": 0.1289,
      "step": 9050
    },
    {
      "epoch": 4.196614885230698,
      "eval_loss": 0.1609577221237089,
      "eval_runtime": 61.0186,
      "eval_samples_per_second": 681.53,
      "eval_steps_per_second": 0.672,
      "step": 9050
    },
    {
      "epoch": 4.219800602828657,
      "grad_norm": 0.04383298382163048,
      "learning_rate": 2.0253513192751373e-06,
      "loss": 0.1289,
      "step": 9100
    },
    {
      "epoch": 4.219800602828657,
      "eval_loss": 0.1598739506376352,
      "eval_runtime": 60.6256,
      "eval_samples_per_second": 685.948,
      "eval_steps_per_second": 0.676,
      "step": 9100
    },
    {
      "epoch": 4.242986320426617,
      "grad_norm": 0.044339120388031006,
      "learning_rate": 1.807892072002898e-06,
      "loss": 0.1283,
      "step": 9150
    },
    {
      "epoch": 4.242986320426617,
      "eval_loss": 0.158920794598519,
      "eval_runtime": 60.5454,
      "eval_samples_per_second": 686.856,
      "eval_steps_per_second": 0.677,
      "step": 9150
    },
    {
      "epoch": 4.2661720380245765,
      "grad_norm": 0.04090524837374687,
      "learning_rate": 1.6025649301821876e-06,
      "loss": 0.1282,
      "step": 9200
    },
    {
      "epoch": 4.2661720380245765,
      "eval_loss": 0.1596859048948022,
      "eval_runtime": 60.7716,
      "eval_samples_per_second": 684.3,
      "eval_steps_per_second": 0.675,
      "step": 9200
    },
    {
      "epoch": 4.289357755622537,
      "grad_norm": 0.042642634361982346,
      "learning_rate": 1.4094215838229176e-06,
      "loss": 0.1286,
      "step": 9250
    },
    {
      "epoch": 4.289357755622537,
      "eval_loss": 0.16079005979316527,
      "eval_runtime": 60.6239,
      "eval_samples_per_second": 685.967,
      "eval_steps_per_second": 0.676,
      "step": 9250
    },
    {
      "epoch": 4.312543473220496,
      "grad_norm": 0.04924129322171211,
      "learning_rate": 1.2285106557296477e-06,
      "loss": 0.1287,
      "step": 9300
    },
    {
      "epoch": 4.312543473220496,
      "eval_loss": 0.16084020796323667,
      "eval_runtime": 60.2581,
      "eval_samples_per_second": 690.131,
      "eval_steps_per_second": 0.68,
      "step": 9300
    },
    {
      "epoch": 4.335729190818456,
      "grad_norm": 0.04222133755683899,
      "learning_rate": 1.0598776892610685e-06,
      "loss": 0.1287,
      "step": 9350
    },
    {
      "epoch": 4.335729190818456,
      "eval_loss": 0.16020395921618655,
      "eval_runtime": 60.5816,
      "eval_samples_per_second": 686.446,
      "eval_steps_per_second": 0.677,
      "step": 9350
    },
    {
      "epoch": 4.358914908416415,
      "grad_norm": 0.05593874678015709,
      "learning_rate": 9.035651368646648e-07,
      "loss": 0.1286,
      "step": 9400
    },
    {
      "epoch": 4.358914908416415,
      "eval_loss": 0.15957607987630548,
      "eval_runtime": 60.8726,
      "eval_samples_per_second": 683.164,
      "eval_steps_per_second": 0.674,
      "step": 9400
    },
    {
      "epoch": 4.382100626014375,
      "grad_norm": 0.059049129486083984,
      "learning_rate": 7.596123493895991e-07,
      "loss": 0.1289,
      "step": 9450
    },
    {
      "epoch": 4.382100626014375,
      "eval_loss": 0.15975197211451994,
      "eval_runtime": 60.6704,
      "eval_samples_per_second": 685.441,
      "eval_steps_per_second": 0.676,
      "step": 9450
    },
    {
      "epoch": 4.405286343612334,
      "grad_norm": 0.053555767983198166,
      "learning_rate": 6.280555661802856e-07,
      "loss": 0.1286,
      "step": 9500
    },
    {
      "epoch": 4.405286343612334,
      "eval_loss": 0.16117730336557945,
      "eval_runtime": 61.9214,
      "eval_samples_per_second": 671.593,
      "eval_steps_per_second": 0.662,
      "step": 9500
    },
    {
      "epoch": 4.428472061210295,
      "grad_norm": 0.04488294571638107,
      "learning_rate": 5.089279059533658e-07,
      "loss": 0.1281,
      "step": 9550
    },
    {
      "epoch": 4.428472061210295,
      "eval_loss": 0.15896389365558133,
      "eval_runtime": 62.0889,
      "eval_samples_per_second": 669.782,
      "eval_steps_per_second": 0.66,
      "step": 9550
    },
    {
      "epoch": 4.451657778808254,
      "grad_norm": 0.044143371284008026,
      "learning_rate": 4.02259358460233e-07,
      "loss": 0.1276,
      "step": 9600
    },
    {
      "epoch": 4.451657778808254,
      "eval_loss": 0.15880485748262804,
      "eval_runtime": 61.9007,
      "eval_samples_per_second": 671.818,
      "eval_steps_per_second": 0.662,
      "step": 9600
    },
    {
      "epoch": 4.474843496406214,
      "grad_norm": 0.054890409111976624,
      "learning_rate": 3.080767769372939e-07,
      "loss": 0.1289,
      "step": 9650
    },
    {
      "epoch": 4.474843496406214,
      "eval_loss": 0.15899979579394047,
      "eval_runtime": 61.7264,
      "eval_samples_per_second": 673.714,
      "eval_steps_per_second": 0.664,
      "step": 9650
    },
    {
      "epoch": 4.498029214004173,
      "grad_norm": 0.04276006668806076,
      "learning_rate": 2.2640387134577058e-07,
      "loss": 0.1284,
      "step": 9700
    },
    {
      "epoch": 4.498029214004173,
      "eval_loss": 0.1587265635928511,
      "eval_runtime": 61.254,
      "eval_samples_per_second": 678.911,
      "eval_steps_per_second": 0.669,
      "step": 9700
    },
    {
      "epoch": 4.521214931602133,
      "grad_norm": 0.04374442994594574,
      "learning_rate": 1.5726120240288634e-07,
      "loss": 0.1284,
      "step": 9750
    },
    {
      "epoch": 4.521214931602133,
      "eval_loss": 0.1596641951113874,
      "eval_runtime": 61.5999,
      "eval_samples_per_second": 675.099,
      "eval_steps_per_second": 0.666,
      "step": 9750
    },
    {
      "epoch": 4.544400649200092,
      "grad_norm": 0.039518803358078,
      "learning_rate": 1.0066617640578368e-07,
      "loss": 0.1297,
      "step": 9800
    },
    {
      "epoch": 4.544400649200092,
      "eval_loss": 0.15941302591091938,
      "eval_runtime": 61.7763,
      "eval_samples_per_second": 673.17,
      "eval_steps_per_second": 0.664,
      "step": 9800
    },
    {
      "epoch": 4.567586366798053,
      "grad_norm": 0.037454187870025635,
      "learning_rate": 5.663304084960186e-08,
      "loss": 0.1276,
      "step": 9850
    },
    {
      "epoch": 4.567586366798053,
      "eval_loss": 0.15932704533345807,
      "eval_runtime": 61.0673,
      "eval_samples_per_second": 680.987,
      "eval_steps_per_second": 0.671,
      "step": 9850
    },
    {
      "epoch": 4.590772084396012,
      "grad_norm": 0.05642937496304512,
      "learning_rate": 2.5172880840745873e-08,
      "loss": 0.129,
      "step": 9900
    },
    {
      "epoch": 4.590772084396012,
      "eval_loss": 0.15923822751292327,
      "eval_runtime": 61.7949,
      "eval_samples_per_second": 672.968,
      "eval_steps_per_second": 0.663,
      "step": 9900
    },
    {
      "epoch": 4.613957801993972,
      "grad_norm": 0.03662274032831192,
      "learning_rate": 6.293616306246586e-09,
      "loss": 0.1285,
      "step": 9950
    },
    {
      "epoch": 4.613957801993972,
      "eval_loss": 0.160277338388973,
      "eval_runtime": 62.2438,
      "eval_samples_per_second": 668.115,
      "eval_steps_per_second": 0.659,
      "step": 9950
    },
    {
      "epoch": 4.637143519591931,
      "grad_norm": 0.0563049279153347,
      "learning_rate": 0.0,
      "loss": 0.1282,
      "step": 10000
    },
    {
      "epoch": 4.637143519591931,
      "eval_loss": 0.16006394581914293,
      "eval_runtime": 61.4917,
      "eval_samples_per_second": 676.286,
      "eval_steps_per_second": 0.667,
      "step": 10000
    },
    {
      "epoch": 4.637143519591931,
      "step": 10000,
      "total_flos": 2.3231400526217216e+17,
      "train_loss": 0.13326009378433226,
      "train_runtime": 41368.3669,
      "train_samples_per_second": 495.064,
      "train_steps_per_second": 0.242
    }
  ],
  "logging_steps": 50,
  "max_steps": 10000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 50,
  "total_flos": 2.3231400526217216e+17,
  "train_batch_size": 1024,
  "trial_name": null,
  "trial_params": null
}