{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 100.0,
  "eval_steps": 500,
  "global_step": 224800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 17.773962020874023,
      "learning_rate": 9.900311387900357e-06,
      "loss": 1.2276,
      "step": 2248
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5729537606239319,
      "eval_loss": 1.3142218589782715,
      "eval_runtime": 43.9858,
      "eval_samples_per_second": 12.777,
      "eval_steps_per_second": 12.777,
      "step": 2248
    },
    {
      "epoch": 2.0,
      "grad_norm": 33.84526824951172,
      "learning_rate": 9.800355871886121e-06,
      "loss": 1.4493,
      "step": 4496
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5907473564147949,
      "eval_loss": 1.5339363813400269,
      "eval_runtime": 44.2068,
      "eval_samples_per_second": 12.713,
      "eval_steps_per_second": 12.713,
      "step": 4496
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.590007483959198,
      "learning_rate": 9.700400355871887e-06,
      "loss": 1.4328,
      "step": 6744
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6583629846572876,
      "eval_loss": 1.5228670835494995,
      "eval_runtime": 44.2648,
      "eval_samples_per_second": 12.696,
      "eval_steps_per_second": 12.696,
      "step": 6744
    },
    {
      "epoch": 4.0,
      "grad_norm": 27.285722732543945,
      "learning_rate": 9.600533807829182e-06,
      "loss": 1.3839,
      "step": 8992
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6921707987785339,
      "eval_loss": 1.333614706993103,
      "eval_runtime": 44.2465,
      "eval_samples_per_second": 12.702,
      "eval_steps_per_second": 12.702,
      "step": 8992
    },
    {
      "epoch": 5.0,
      "grad_norm": 28.629926681518555,
      "learning_rate": 9.500533807829183e-06,
      "loss": 1.4591,
      "step": 11240
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6761565804481506,
      "eval_loss": 1.4407259225845337,
      "eval_runtime": 44.1453,
      "eval_samples_per_second": 12.731,
      "eval_steps_per_second": 12.731,
      "step": 11240
    },
    {
      "epoch": 6.0,
      "grad_norm": 14.747943878173828,
      "learning_rate": 9.400533807829182e-06,
      "loss": 1.3609,
      "step": 13488
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7117437720298767,
      "eval_loss": 1.1341328620910645,
      "eval_runtime": 44.9034,
      "eval_samples_per_second": 12.516,
      "eval_steps_per_second": 12.516,
      "step": 13488
    },
    {
      "epoch": 7.0,
      "grad_norm": 44.90867233276367,
      "learning_rate": 9.300578291814948e-06,
      "loss": 1.222,
      "step": 15736
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7028470039367676,
      "eval_loss": 1.5203638076782227,
      "eval_runtime": 44.0645,
      "eval_samples_per_second": 12.754,
      "eval_steps_per_second": 12.754,
      "step": 15736
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.5433468818664551,
      "learning_rate": 9.200622775800714e-06,
      "loss": 1.1713,
      "step": 17984
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7473309636116028,
      "eval_loss": 1.3221626281738281,
      "eval_runtime": 44.4952,
      "eval_samples_per_second": 12.631,
      "eval_steps_per_second": 12.631,
      "step": 17984
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.15495184063911438,
      "learning_rate": 9.100622775800713e-06,
      "loss": 1.1414,
      "step": 20232
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.7562277317047119,
      "eval_loss": 1.3233251571655273,
      "eval_runtime": 44.534,
      "eval_samples_per_second": 12.62,
      "eval_steps_per_second": 12.62,
      "step": 20232
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.034351445734500885,
      "learning_rate": 9.000711743772242e-06,
      "loss": 1.0977,
      "step": 22480
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.7384341359138489,
      "eval_loss": 1.3587311506271362,
      "eval_runtime": 44.5493,
      "eval_samples_per_second": 12.615,
      "eval_steps_per_second": 12.615,
      "step": 22480
    },
    {
      "epoch": 11.0,
      "grad_norm": 0.17834441363811493,
      "learning_rate": 8.900711743772243e-06,
      "loss": 0.9768,
      "step": 24728
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.7633451819419861,
      "eval_loss": 1.348164677619934,
      "eval_runtime": 44.2727,
      "eval_samples_per_second": 12.694,
      "eval_steps_per_second": 12.694,
      "step": 24728
    },
    {
      "epoch": 12.0,
      "grad_norm": 0.0854184553027153,
      "learning_rate": 8.800800711743773e-06,
      "loss": 0.9219,
      "step": 26976
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.7580071091651917,
      "eval_loss": 1.3742923736572266,
      "eval_runtime": 44.1709,
      "eval_samples_per_second": 12.723,
      "eval_steps_per_second": 12.723,
      "step": 26976
    },
    {
      "epoch": 13.0,
      "grad_norm": 1.15831458568573,
      "learning_rate": 8.700800711743772e-06,
      "loss": 0.8636,
      "step": 29224
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.8167259693145752,
      "eval_loss": 1.1195218563079834,
      "eval_runtime": 44.3771,
      "eval_samples_per_second": 12.664,
      "eval_steps_per_second": 12.664,
      "step": 29224
    },
    {
      "epoch": 14.0,
      "grad_norm": 658.5977172851562,
      "learning_rate": 8.600845195729538e-06,
      "loss": 0.8543,
      "step": 31472
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.8042704463005066,
      "eval_loss": 1.171613097190857,
      "eval_runtime": 43.8385,
      "eval_samples_per_second": 12.82,
      "eval_steps_per_second": 12.82,
      "step": 31472
    },
    {
      "epoch": 15.0,
      "grad_norm": 1.0300912857055664,
      "learning_rate": 8.500934163701069e-06,
      "loss": 0.8053,
      "step": 33720
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.8113878965377808,
      "eval_loss": 1.2032980918884277,
      "eval_runtime": 44.439,
      "eval_samples_per_second": 12.647,
      "eval_steps_per_second": 12.647,
      "step": 33720
    },
    {
      "epoch": 16.0,
      "grad_norm": 0.1946033537387848,
      "learning_rate": 8.400978647686834e-06,
      "loss": 0.7718,
      "step": 35968
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.7989323735237122,
      "eval_loss": 1.249149203300476,
      "eval_runtime": 46.8737,
      "eval_samples_per_second": 11.99,
      "eval_steps_per_second": 11.99,
      "step": 35968
    },
    {
      "epoch": 17.0,
      "grad_norm": 0.08109813928604126,
      "learning_rate": 8.300978647686834e-06,
      "loss": 0.6904,
      "step": 38216
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.836298942565918,
      "eval_loss": 1.0850577354431152,
      "eval_runtime": 46.9201,
      "eval_samples_per_second": 11.978,
      "eval_steps_per_second": 11.978,
      "step": 38216
    },
    {
      "epoch": 18.0,
      "grad_norm": 0.05036694556474686,
      "learning_rate": 8.2010231316726e-06,
      "loss": 0.6545,
      "step": 40464
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.8007117509841919,
      "eval_loss": 1.2962548732757568,
      "eval_runtime": 44.6234,
      "eval_samples_per_second": 12.594,
      "eval_steps_per_second": 12.594,
      "step": 40464
    },
    {
      "epoch": 19.0,
      "grad_norm": 0.5516142249107361,
      "learning_rate": 8.101023131672599e-06,
      "loss": 0.6858,
      "step": 42712
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.8078292012214661,
      "eval_loss": 1.3231106996536255,
      "eval_runtime": 45.2364,
      "eval_samples_per_second": 12.424,
      "eval_steps_per_second": 12.424,
      "step": 42712
    },
    {
      "epoch": 20.0,
      "grad_norm": 21.766986846923828,
      "learning_rate": 8.001067615658363e-06,
      "loss": 0.6444,
      "step": 44960
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.8238434195518494,
      "eval_loss": 1.1918330192565918,
      "eval_runtime": 45.4969,
      "eval_samples_per_second": 12.352,
      "eval_steps_per_second": 12.352,
      "step": 44960
    },
    {
      "epoch": 21.0,
      "grad_norm": 0.047945111989974976,
      "learning_rate": 7.90111209964413e-06,
      "loss": 0.6166,
      "step": 47208
    },
    {
      "epoch": 21.0,
      "eval_accuracy": 0.8345195651054382,
      "eval_loss": 1.1358013153076172,
      "eval_runtime": 44.2578,
      "eval_samples_per_second": 12.698,
      "eval_steps_per_second": 12.698,
      "step": 47208
    },
    {
      "epoch": 22.0,
      "grad_norm": 0.016365328803658485,
      "learning_rate": 7.801156583629894e-06,
      "loss": 0.5437,
      "step": 49456
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.8256227970123291,
      "eval_loss": 1.2446305751800537,
      "eval_runtime": 44.445,
      "eval_samples_per_second": 12.645,
      "eval_steps_per_second": 12.645,
      "step": 49456
    },
    {
      "epoch": 23.0,
      "grad_norm": 0.040828485041856766,
      "learning_rate": 7.701156583629893e-06,
      "loss": 0.4719,
      "step": 51704
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.8149465918540955,
      "eval_loss": 1.4120286703109741,
      "eval_runtime": 44.091,
      "eval_samples_per_second": 12.746,
      "eval_steps_per_second": 12.746,
      "step": 51704
    },
    {
      "epoch": 24.0,
      "grad_norm": 1.33746337890625,
      "learning_rate": 7.601201067615659e-06,
      "loss": 0.4802,
      "step": 53952
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.8202847242355347,
      "eval_loss": 1.2611161470413208,
      "eval_runtime": 44.1951,
      "eval_samples_per_second": 12.716,
      "eval_steps_per_second": 12.716,
      "step": 53952
    },
    {
      "epoch": 25.0,
      "grad_norm": 95.64888763427734,
      "learning_rate": 7.501290035587189e-06,
      "loss": 0.484,
      "step": 56200
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 0.836298942565918,
      "eval_loss": 1.2840217351913452,
      "eval_runtime": 44.2468,
      "eval_samples_per_second": 12.701,
      "eval_steps_per_second": 12.701,
      "step": 56200
    },
    {
      "epoch": 26.0,
      "grad_norm": 0.019676001742482185,
      "learning_rate": 7.40129003558719e-06,
      "loss": 0.3649,
      "step": 58448
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.8434163928031921,
      "eval_loss": 1.2421414852142334,
      "eval_runtime": 44.4397,
      "eval_samples_per_second": 12.646,
      "eval_steps_per_second": 12.646,
      "step": 58448
    },
    {
      "epoch": 27.0,
      "grad_norm": 0.01975702866911888,
      "learning_rate": 7.301379003558719e-06,
      "loss": 0.4146,
      "step": 60696
    },
    {
      "epoch": 27.0,
      "eval_accuracy": 0.8291814923286438,
      "eval_loss": 1.3464977741241455,
      "eval_runtime": 43.9971,
      "eval_samples_per_second": 12.774,
      "eval_steps_per_second": 12.774,
      "step": 60696
    },
    {
      "epoch": 28.0,
      "grad_norm": 0.0131510766223073,
      "learning_rate": 7.201423487544484e-06,
      "loss": 0.3998,
      "step": 62944
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.8505337834358215,
      "eval_loss": 1.2309142351150513,
      "eval_runtime": 44.0844,
      "eval_samples_per_second": 12.748,
      "eval_steps_per_second": 12.748,
      "step": 62944
    },
    {
      "epoch": 29.0,
      "grad_norm": 0.043094441294670105,
      "learning_rate": 7.101423487544484e-06,
      "loss": 0.4113,
      "step": 65192
    },
    {
      "epoch": 29.0,
      "eval_accuracy": 0.8523131608963013,
      "eval_loss": 1.166311264038086,
      "eval_runtime": 45.2787,
      "eval_samples_per_second": 12.412,
      "eval_steps_per_second": 12.412,
      "step": 65192
    },
    {
      "epoch": 30.0,
      "grad_norm": 0.04462951049208641,
      "learning_rate": 7.001467971530249e-06,
      "loss": 0.3385,
      "step": 67440
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.8469750881195068,
      "eval_loss": 1.2566747665405273,
      "eval_runtime": 43.6729,
      "eval_samples_per_second": 12.868,
      "eval_steps_per_second": 12.868,
      "step": 67440
    },
    {
      "epoch": 31.0,
      "grad_norm": 0.003071392187848687,
      "learning_rate": 6.901512455516015e-06,
      "loss": 0.3188,
      "step": 69688
    },
    {
      "epoch": 31.0,
      "eval_accuracy": 0.8434163928031921,
      "eval_loss": 1.2580705881118774,
      "eval_runtime": 45.193,
      "eval_samples_per_second": 12.436,
      "eval_steps_per_second": 12.436,
      "step": 69688
    },
    {
      "epoch": 32.0,
      "grad_norm": 0.051271870732307434,
      "learning_rate": 6.80155693950178e-06,
      "loss": 0.3203,
      "step": 71936
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.854092538356781,
      "eval_loss": 1.2454197406768799,
      "eval_runtime": 43.9824,
      "eval_samples_per_second": 12.778,
      "eval_steps_per_second": 12.778,
      "step": 71936
    },
    {
      "epoch": 33.0,
      "grad_norm": 0.003354266518726945,
      "learning_rate": 6.701645907473309e-06,
      "loss": 0.2766,
      "step": 74184
    },
    {
      "epoch": 33.0,
      "eval_accuracy": 0.8523131608963013,
      "eval_loss": 1.2542338371276855,
      "eval_runtime": 44.114,
      "eval_samples_per_second": 12.74,
      "eval_steps_per_second": 12.74,
      "step": 74184
    },
    {
      "epoch": 34.0,
      "grad_norm": 0.0016816665884107351,
      "learning_rate": 6.6016459074733095e-06,
      "loss": 0.2505,
      "step": 76432
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.8149465918540955,
      "eval_loss": 1.5897480249404907,
      "eval_runtime": 44.5247,
      "eval_samples_per_second": 12.622,
      "eval_steps_per_second": 12.622,
      "step": 76432
    },
    {
      "epoch": 35.0,
      "grad_norm": 0.0018826358718797565,
      "learning_rate": 6.501690391459075e-06,
      "loss": 0.2777,
      "step": 78680
    },
    {
      "epoch": 35.0,
      "eval_accuracy": 0.836298942565918,
      "eval_loss": 1.348265528678894,
      "eval_runtime": 44.5684,
      "eval_samples_per_second": 12.61,
      "eval_steps_per_second": 12.61,
      "step": 78680
    },
    {
      "epoch": 36.0,
      "grad_norm": 0.0017322949133813381,
      "learning_rate": 6.40173487544484e-06,
      "loss": 0.2816,
      "step": 80928
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.8523131608963013,
      "eval_loss": 1.2510393857955933,
      "eval_runtime": 44.2735,
      "eval_samples_per_second": 12.694,
      "eval_steps_per_second": 12.694,
      "step": 80928
    },
    {
      "epoch": 37.0,
      "grad_norm": 0.002890912117436528,
      "learning_rate": 6.301779359430605e-06,
      "loss": 0.2728,
      "step": 83176
    },
    {
      "epoch": 37.0,
      "eval_accuracy": 0.8327401876449585,
      "eval_loss": 1.4422318935394287,
      "eval_runtime": 45.8456,
      "eval_samples_per_second": 12.259,
      "eval_steps_per_second": 12.259,
      "step": 83176
    },
    {
      "epoch": 38.0,
      "grad_norm": 0.021676644682884216,
      "learning_rate": 6.20182384341637e-06,
      "loss": 0.255,
      "step": 85424
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.8487544655799866,
      "eval_loss": 1.2928466796875,
      "eval_runtime": 44.6873,
      "eval_samples_per_second": 12.576,
      "eval_steps_per_second": 12.576,
      "step": 85424
    },
    {
      "epoch": 39.0,
      "grad_norm": 0.06743878126144409,
      "learning_rate": 6.10182384341637e-06,
      "loss": 0.2172,
      "step": 87672
    },
    {
      "epoch": 39.0,
      "eval_accuracy": 0.8451957106590271,
      "eval_loss": 1.4022176265716553,
      "eval_runtime": 44.9479,
      "eval_samples_per_second": 12.503,
      "eval_steps_per_second": 12.503,
      "step": 87672
    },
    {
      "epoch": 40.0,
      "grad_norm": 0.0027778081130236387,
      "learning_rate": 6.001823843416371e-06,
      "loss": 0.2204,
      "step": 89920
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.8380783200263977,
      "eval_loss": 1.4114270210266113,
      "eval_runtime": 44.0198,
      "eval_samples_per_second": 12.767,
      "eval_steps_per_second": 12.767,
      "step": 89920
    },
    {
      "epoch": 41.0,
      "grad_norm": 0.007720357272773981,
      "learning_rate": 5.901912811387901e-06,
      "loss": 0.2232,
      "step": 92168
    },
    {
      "epoch": 41.0,
      "eval_accuracy": 0.8416370153427124,
      "eval_loss": 1.4324493408203125,
      "eval_runtime": 43.6963,
      "eval_samples_per_second": 12.861,
      "eval_steps_per_second": 12.861,
      "step": 92168
    },
    {
      "epoch": 42.0,
      "grad_norm": 0.007710450328886509,
      "learning_rate": 5.8019572953736655e-06,
      "loss": 0.2301,
      "step": 94416
    },
    {
      "epoch": 42.0,
      "eval_accuracy": 0.8487544655799866,
      "eval_loss": 1.3528090715408325,
      "eval_runtime": 45.0443,
      "eval_samples_per_second": 12.477,
      "eval_steps_per_second": 12.477,
      "step": 94416
    },
    {
      "epoch": 43.0,
      "grad_norm": 0.00038957124343141913,
      "learning_rate": 5.70200177935943e-06,
      "loss": 0.1751,
      "step": 96664
    },
    {
      "epoch": 43.0,
      "eval_accuracy": 0.8434163928031921,
      "eval_loss": 1.4648776054382324,
      "eval_runtime": 44.1086,
      "eval_samples_per_second": 12.741,
      "eval_steps_per_second": 12.741,
      "step": 96664
    },
    {
      "epoch": 44.0,
      "grad_norm": 0.0042132362723350525,
      "learning_rate": 5.602046263345197e-06,
      "loss": 0.1982,
      "step": 98912
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.8754448294639587,
      "eval_loss": 1.221611499786377,
      "eval_runtime": 44.0087,
      "eval_samples_per_second": 12.77,
      "eval_steps_per_second": 12.77,
      "step": 98912
    },
    {
      "epoch": 45.0,
      "grad_norm": 0.011830995790660381,
      "learning_rate": 5.502046263345196e-06,
      "loss": 0.1803,
      "step": 101160
    },
    {
      "epoch": 45.0,
      "eval_accuracy": 0.8451957106590271,
      "eval_loss": 1.4569214582443237,
      "eval_runtime": 44.6868,
      "eval_samples_per_second": 12.576,
      "eval_steps_per_second": 12.576,
      "step": 101160
    },
    {
      "epoch": 46.0,
      "grad_norm": 0.0046292925253510475,
      "learning_rate": 5.402090747330961e-06,
      "loss": 0.1582,
      "step": 103408
    },
    {
      "epoch": 46.0,
      "eval_accuracy": 0.8665480613708496,
      "eval_loss": 1.365021824836731,
      "eval_runtime": 43.9986,
      "eval_samples_per_second": 12.773,
      "eval_steps_per_second": 12.773,
      "step": 103408
    },
    {
      "epoch": 47.0,
      "grad_norm": 0.041185297071933746,
      "learning_rate": 5.302135231316726e-06,
      "loss": 0.1837,
      "step": 105656
    },
    {
      "epoch": 47.0,
      "eval_accuracy": 0.854092538356781,
      "eval_loss": 1.2877007722854614,
      "eval_runtime": 44.0731,
      "eval_samples_per_second": 12.752,
      "eval_steps_per_second": 12.752,
      "step": 105656
    },
    {
      "epoch": 48.0,
      "grad_norm": 0.04080544412136078,
      "learning_rate": 5.202179715302491e-06,
      "loss": 0.1458,
      "step": 107904
    },
    {
      "epoch": 48.0,
      "eval_accuracy": 0.8309608697891235,
      "eval_loss": 1.7388625144958496,
      "eval_runtime": 45.0509,
      "eval_samples_per_second": 12.475,
      "eval_steps_per_second": 12.475,
      "step": 107904
    },
    {
      "epoch": 49.0,
      "grad_norm": 0.003914756700396538,
      "learning_rate": 5.1022241992882574e-06,
      "loss": 0.1664,
      "step": 110152
    },
    {
      "epoch": 49.0,
      "eval_accuracy": 0.854092538356781,
      "eval_loss": 1.4000823497772217,
      "eval_runtime": 44.2826,
      "eval_samples_per_second": 12.691,
      "eval_steps_per_second": 12.691,
      "step": 110152
    },
    {
      "epoch": 50.0,
      "grad_norm": 0.0011822431115433574,
      "learning_rate": 5.002268683274022e-06,
      "loss": 0.1473,
      "step": 112400
    },
    {
      "epoch": 50.0,
      "eval_accuracy": 0.8701067566871643,
      "eval_loss": 1.2979094982147217,
      "eval_runtime": 44.3693,
      "eval_samples_per_second": 12.666,
      "eval_steps_per_second": 12.666,
      "step": 112400
    },
    {
      "epoch": 51.0,
      "grad_norm": 0.0008859778754413128,
      "learning_rate": 4.9022686832740216e-06,
      "loss": 0.1341,
      "step": 114648
    },
    {
      "epoch": 51.0,
      "eval_accuracy": 0.8469750881195068,
      "eval_loss": 1.570462942123413,
      "eval_runtime": 44.0624,
      "eval_samples_per_second": 12.755,
      "eval_steps_per_second": 12.755,
      "step": 114648
    },
    {
      "epoch": 52.0,
      "grad_norm": 0.0008422972168773413,
      "learning_rate": 4.8023131672597865e-06,
      "loss": 0.1603,
      "step": 116896
    },
    {
      "epoch": 52.0,
      "eval_accuracy": 0.8380783200263977,
      "eval_loss": 1.6043403148651123,
      "eval_runtime": 44.715,
      "eval_samples_per_second": 12.568,
      "eval_steps_per_second": 12.568,
      "step": 116896
    },
    {
      "epoch": 53.0,
      "grad_norm": 0.1126694455742836,
      "learning_rate": 4.702357651245552e-06,
      "loss": 0.1133,
      "step": 119144
    },
    {
      "epoch": 53.0,
      "eval_accuracy": 0.8451957106590271,
      "eval_loss": 1.6194249391555786,
      "eval_runtime": 44.3847,
      "eval_samples_per_second": 12.662,
      "eval_steps_per_second": 12.662,
      "step": 119144
    },
    {
      "epoch": 54.0,
      "grad_norm": 0.0018404704751446843,
      "learning_rate": 4.602402135231317e-06,
      "loss": 0.107,
      "step": 121392
    },
    {
      "epoch": 54.0,
      "eval_accuracy": 0.8629893064498901,
      "eval_loss": 1.4172945022583008,
      "eval_runtime": 45.9409,
      "eval_samples_per_second": 12.233,
      "eval_steps_per_second": 12.233,
      "step": 121392
    },
    {
      "epoch": 55.0,
      "grad_norm": 0.0005189430085010827,
      "learning_rate": 4.502402135231317e-06,
      "loss": 0.116,
      "step": 123640
    },
    {
      "epoch": 55.0,
      "eval_accuracy": 0.854092538356781,
      "eval_loss": 1.526777982711792,
      "eval_runtime": 44.5269,
      "eval_samples_per_second": 12.622,
      "eval_steps_per_second": 12.622,
      "step": 123640
    },
    {
      "epoch": 56.0,
      "grad_norm": 0.000902254250831902,
      "learning_rate": 4.402446619217082e-06,
      "loss": 0.0988,
      "step": 125888
    },
    {
      "epoch": 56.0,
      "eval_accuracy": 0.8523131608963013,
      "eval_loss": 1.6091721057891846,
      "eval_runtime": 44.4807,
      "eval_samples_per_second": 12.635,
      "eval_steps_per_second": 12.635,
      "step": 125888
    },
    {
      "epoch": 57.0,
      "grad_norm": 0.04167209565639496,
      "learning_rate": 4.302491103202847e-06,
      "loss": 0.139,
      "step": 128136
    },
    {
      "epoch": 57.0,
      "eval_accuracy": 0.8647686839103699,
      "eval_loss": 1.4311786890029907,
      "eval_runtime": 44.6921,
      "eval_samples_per_second": 12.575,
      "eval_steps_per_second": 12.575,
      "step": 128136
    },
    {
      "epoch": 58.0,
      "grad_norm": 0.0009059112053364515,
      "learning_rate": 4.202535587188613e-06,
      "loss": 0.0798,
      "step": 130384
    },
    {
      "epoch": 58.0,
      "eval_accuracy": 0.8327401876449585,
      "eval_loss": 1.7888070344924927,
      "eval_runtime": 46.1517,
      "eval_samples_per_second": 12.177,
      "eval_steps_per_second": 12.177,
      "step": 130384
    },
    {
      "epoch": 59.0,
      "grad_norm": 0.00021046701294835657,
      "learning_rate": 4.102535587188613e-06,
      "loss": 0.0776,
      "step": 132632
    },
    {
      "epoch": 59.0,
      "eval_accuracy": 0.8665480613708496,
      "eval_loss": 1.5457098484039307,
      "eval_runtime": 44.2018,
      "eval_samples_per_second": 12.714,
      "eval_steps_per_second": 12.714,
      "step": 132632
    },
    {
      "epoch": 60.0,
      "grad_norm": 0.0024244808591902256,
      "learning_rate": 4.002580071174378e-06,
      "loss": 0.1288,
      "step": 134880
    },
    {
      "epoch": 60.0,
      "eval_accuracy": 0.8629893064498901,
      "eval_loss": 1.4553505182266235,
      "eval_runtime": 43.9122,
      "eval_samples_per_second": 12.798,
      "eval_steps_per_second": 12.798,
      "step": 134880
    },
    {
      "epoch": 61.0,
      "grad_norm": 0.0004949413705617189,
      "learning_rate": 3.9026245551601425e-06,
      "loss": 0.0828,
      "step": 137128
    },
    {
      "epoch": 61.0,
      "eval_accuracy": 0.8558719158172607,
      "eval_loss": 1.7077866792678833,
      "eval_runtime": 44.3471,
      "eval_samples_per_second": 12.673,
      "eval_steps_per_second": 12.673,
      "step": 137128
    },
    {
      "epoch": 62.0,
      "grad_norm": 0.00021063751773908734,
      "learning_rate": 3.8026690391459074e-06,
      "loss": 0.0823,
      "step": 139376
    },
    {
      "epoch": 62.0,
      "eval_accuracy": 0.8754448294639587,
      "eval_loss": 1.4733619689941406,
      "eval_runtime": 44.2713,
      "eval_samples_per_second": 12.694,
      "eval_steps_per_second": 12.694,
      "step": 139376
    },
    {
      "epoch": 63.0,
      "grad_norm": 0.0006752557819709182,
      "learning_rate": 3.702713523131673e-06,
      "loss": 0.0803,
      "step": 141624
    },
    {
      "epoch": 63.0,
      "eval_accuracy": 0.8594306111335754,
      "eval_loss": 1.6007416248321533,
      "eval_runtime": 44.5589,
      "eval_samples_per_second": 12.613,
      "eval_steps_per_second": 12.613,
      "step": 141624
    },
    {
      "epoch": 64.0,
      "grad_norm": 0.0015100333839654922,
      "learning_rate": 3.602758007117438e-06,
      "loss": 0.0947,
      "step": 143872
    },
    {
      "epoch": 64.0,
      "eval_accuracy": 0.8701067566871643,
      "eval_loss": 1.4466949701309204,
      "eval_runtime": 44.6568,
      "eval_samples_per_second": 12.585,
      "eval_steps_per_second": 12.585,
      "step": 143872
    },
    {
      "epoch": 65.0,
      "grad_norm": 0.0003055017732549459,
      "learning_rate": 3.5028024911032033e-06,
      "loss": 0.0916,
      "step": 146120
    },
    {
      "epoch": 65.0,
      "eval_accuracy": 0.873665452003479,
      "eval_loss": 1.4410454034805298,
      "eval_runtime": 44.5129,
      "eval_samples_per_second": 12.626,
      "eval_steps_per_second": 12.626,
      "step": 146120
    },
    {
      "epoch": 66.0,
      "grad_norm": 0.0007535702898167074,
      "learning_rate": 3.402846975088968e-06,
      "loss": 0.0814,
      "step": 148368
    },
    {
      "epoch": 66.0,
      "eval_accuracy": 0.8469750881195068,
      "eval_loss": 1.711586356163025,
      "eval_runtime": 44.7221,
      "eval_samples_per_second": 12.566,
      "eval_steps_per_second": 12.566,
      "step": 148368
    },
    {
      "epoch": 67.0,
      "grad_norm": NaN,
      "learning_rate": 3.3028914590747335e-06,
      "loss": 0.0938,
      "step": 150616
    },
    {
      "epoch": 67.0,
      "eval_accuracy": 0.8629893064498901,
      "eval_loss": 1.5838223695755005,
      "eval_runtime": 44.5452,
      "eval_samples_per_second": 12.616,
      "eval_steps_per_second": 12.616,
      "step": 150616
    },
    {
      "epoch": 68.0,
      "grad_norm": 0.00016811703972052783,
      "learning_rate": 3.2028914590747336e-06,
      "loss": 0.066,
      "step": 152864
    },
    {
      "epoch": 68.0,
      "eval_accuracy": 0.8558719158172607,
      "eval_loss": 1.645817518234253,
      "eval_runtime": 44.4578,
      "eval_samples_per_second": 12.641,
      "eval_steps_per_second": 12.641,
      "step": 152864
    },
    {
      "epoch": 69.0,
      "grad_norm": 0.0004475938912946731,
      "learning_rate": 3.102935943060498e-06,
      "loss": 0.096,
      "step": 155112
    },
    {
      "epoch": 69.0,
      "eval_accuracy": 0.8558719158172607,
      "eval_loss": 1.6925562620162964,
      "eval_runtime": 44.7853,
      "eval_samples_per_second": 12.549,
      "eval_steps_per_second": 12.549,
      "step": 155112
    },
    {
      "epoch": 70.0,
      "grad_norm": 0.005046023055911064,
      "learning_rate": 3.0030249110320286e-06,
      "loss": 0.0638,
      "step": 157360
    },
    {
      "epoch": 70.0,
      "eval_accuracy": 0.8629893064498901,
      "eval_loss": 1.5232763290405273,
      "eval_runtime": 44.8809,
      "eval_samples_per_second": 12.522,
      "eval_steps_per_second": 12.522,
      "step": 157360
    },
    {
      "epoch": 71.0,
      "grad_norm": 0.00024611467961221933,
      "learning_rate": 2.9030249110320287e-06,
      "loss": 0.063,
      "step": 159608
    },
    {
      "epoch": 71.0,
      "eval_accuracy": 0.8594306111335754,
      "eval_loss": 1.5641191005706787,
      "eval_runtime": 44.3769,
      "eval_samples_per_second": 12.664,
      "eval_steps_per_second": 12.664,
      "step": 159608
    },
    {
      "epoch": 72.0,
      "grad_norm": 0.001299203373491764,
      "learning_rate": 2.8030249110320284e-06,
      "loss": 0.0758,
      "step": 161856
    },
    {
      "epoch": 72.0,
      "eval_accuracy": 0.8505337834358215,
      "eval_loss": 1.6766804456710815,
      "eval_runtime": 44.9521,
      "eval_samples_per_second": 12.502,
      "eval_steps_per_second": 12.502,
      "step": 161856
    },
    {
      "epoch": 73.0,
      "grad_norm": 0.00021305486734490842,
      "learning_rate": 2.7030693950177937e-06,
      "loss": 0.0579,
      "step": 164104
    },
    {
      "epoch": 73.0,
      "eval_accuracy": 0.8629893064498901,
      "eval_loss": 1.5338364839553833,
      "eval_runtime": 46.2439,
      "eval_samples_per_second": 12.153,
      "eval_steps_per_second": 12.153,
      "step": 164104
    },
    {
      "epoch": 74.0,
      "grad_norm": 0.0005646486533805728,
      "learning_rate": 2.6031138790035586e-06,
      "loss": 0.0379,
      "step": 166352
    },
    {
      "epoch": 74.0,
      "eval_accuracy": 0.8629893064498901,
      "eval_loss": 1.6348403692245483,
      "eval_runtime": 44.783,
      "eval_samples_per_second": 12.549,
      "eval_steps_per_second": 12.549,
      "step": 166352
    },
    {
      "epoch": 75.0,
      "grad_norm": 0.00010759654833236709,
      "learning_rate": 2.5031583629893243e-06,
      "loss": 0.0351,
      "step": 168600
    },
    {
      "epoch": 75.0,
      "eval_accuracy": 0.8558719158172607,
      "eval_loss": 1.70370352268219,
      "eval_runtime": 47.2555,
      "eval_samples_per_second": 11.893,
      "eval_steps_per_second": 11.893,
      "step": 168600
    },
    {
      "epoch": 76.0,
      "grad_norm": 0.000321146595524624,
      "learning_rate": 2.403202846975089e-06,
      "loss": 0.0472,
      "step": 170848
    },
    {
      "epoch": 76.0,
      "eval_accuracy": 0.8754448294639587,
      "eval_loss": 1.5681896209716797,
      "eval_runtime": 44.4779,
      "eval_samples_per_second": 12.635,
      "eval_steps_per_second": 12.635,
      "step": 170848
    },
    {
      "epoch": 77.0,
      "grad_norm": 0.00014101610577199608,
      "learning_rate": 2.3032028469750893e-06,
      "loss": 0.0253,
      "step": 173096
    },
    {
      "epoch": 77.0,
      "eval_accuracy": 0.8558719158172607,
      "eval_loss": 1.7067368030548096,
      "eval_runtime": 44.3009,
      "eval_samples_per_second": 12.686,
      "eval_steps_per_second": 12.686,
      "step": 173096
    },
    {
      "epoch": 78.0,
      "grad_norm": 0.0002290535339852795,
      "learning_rate": 2.203247330960854e-06,
      "loss": 0.073,
      "step": 175344
    },
    {
      "epoch": 78.0,
      "eval_accuracy": 0.8754448294639587,
      "eval_loss": 1.4460408687591553,
      "eval_runtime": 44.9156,
      "eval_samples_per_second": 12.512,
      "eval_steps_per_second": 12.512,
      "step": 175344
    },
    {
      "epoch": 79.0,
      "grad_norm": 0.0008211834938265383,
      "learning_rate": 2.1032918149466195e-06,
      "loss": 0.049,
      "step": 177592
    },
    {
      "epoch": 79.0,
      "eval_accuracy": 0.8594306111335754,
      "eval_loss": 1.5897014141082764,
      "eval_runtime": 44.349,
      "eval_samples_per_second": 12.672,
      "eval_steps_per_second": 12.672,
      "step": 177592
    },
    {
      "epoch": 80.0,
      "grad_norm": 0.0002687973901629448,
      "learning_rate": 2.003291814946619e-06,
      "loss": 0.0503,
      "step": 179840
    },
    {
      "epoch": 80.0,
      "eval_accuracy": 0.8647686839103699,
      "eval_loss": 1.601679801940918,
      "eval_runtime": 44.9732,
      "eval_samples_per_second": 12.496,
      "eval_steps_per_second": 12.496,
      "step": 179840
    },
    {
      "epoch": 81.0,
      "grad_norm": 0.00014994025696069002,
      "learning_rate": 1.9033807829181496e-06,
      "loss": 0.0497,
      "step": 182088
    },
    {
      "epoch": 81.0,
      "eval_accuracy": 0.8683273792266846,
      "eval_loss": 1.5318866968154907,
      "eval_runtime": 44.556,
      "eval_samples_per_second": 12.613,
      "eval_steps_per_second": 12.613,
      "step": 182088
    },
    {
      "epoch": 82.0,
      "grad_norm": 0.00014429475413635373,
      "learning_rate": 1.8034697508896797e-06,
      "loss": 0.0553,
      "step": 184336
    },
    {
      "epoch": 82.0,
      "eval_accuracy": 0.8612099885940552,
      "eval_loss": 1.5479316711425781,
      "eval_runtime": 44.3184,
      "eval_samples_per_second": 12.681,
      "eval_steps_per_second": 12.681,
      "step": 184336
    },
    {
      "epoch": 83.0,
      "grad_norm": 0.0028621815145015717,
      "learning_rate": 1.70346975088968e-06,
      "loss": 0.0416,
      "step": 186584
    },
    {
      "epoch": 83.0,
      "eval_accuracy": 0.8576512336730957,
      "eval_loss": 1.5556381940841675,
      "eval_runtime": 44.9266,
      "eval_samples_per_second": 12.509,
      "eval_steps_per_second": 12.509,
      "step": 186584
    },
    {
      "epoch": 84.0,
      "grad_norm": 0.0002136415132554248,
      "learning_rate": 1.60346975088968e-06,
      "loss": 0.0641,
      "step": 188832
    },
    {
      "epoch": 84.0,
      "eval_accuracy": 0.8594306111335754,
      "eval_loss": 1.5674761533737183,
      "eval_runtime": 44.2532,
      "eval_samples_per_second": 12.7,
      "eval_steps_per_second": 12.7,
      "step": 188832
    },
    {
      "epoch": 85.0,
      "grad_norm": 0.00014172360533848405,
      "learning_rate": 1.503514234875445e-06,
      "loss": 0.0425,
      "step": 191080
    },
    {
      "epoch": 85.0,
      "eval_accuracy": 0.8558719158172607,
      "eval_loss": 1.6853959560394287,
      "eval_runtime": 44.7377,
      "eval_samples_per_second": 12.562,
      "eval_steps_per_second": 12.562,
      "step": 191080
    },
    {
      "epoch": 86.0,
      "grad_norm": 0.0003552982525434345,
      "learning_rate": 1.403514234875445e-06,
      "loss": 0.0311,
      "step": 193328
    },
    {
      "epoch": 86.0,
      "eval_accuracy": 0.873665452003479,
      "eval_loss": 1.4628252983093262,
      "eval_runtime": 44.4922,
      "eval_samples_per_second": 12.631,
      "eval_steps_per_second": 12.631,
      "step": 193328
    },
    {
      "epoch": 87.0,
      "grad_norm": 0.00020055480126757175,
      "learning_rate": 1.3036032028469752e-06,
      "loss": 0.0456,
      "step": 195576
    },
    {
      "epoch": 87.0,
      "eval_accuracy": 0.8701067566871643,
      "eval_loss": 1.5069490671157837,
      "eval_runtime": 44.3942,
      "eval_samples_per_second": 12.659,
      "eval_steps_per_second": 12.659,
      "step": 195576
    },
    {
      "epoch": 88.0,
      "grad_norm": 0.0007724438328295946,
      "learning_rate": 1.2036032028469751e-06,
      "loss": 0.0224,
      "step": 197824
    },
    {
      "epoch": 88.0,
      "eval_accuracy": 0.8665480613708496,
      "eval_loss": 1.612963080406189,
      "eval_runtime": 44.8107,
      "eval_samples_per_second": 12.542,
      "eval_steps_per_second": 12.542,
      "step": 197824
    },
    {
      "epoch": 89.0,
      "grad_norm": 0.00014038340304978192,
      "learning_rate": 1.1036476868327404e-06,
      "loss": 0.0345,
      "step": 200072
    },
    {
      "epoch": 89.0,
      "eval_accuracy": 0.8701067566871643,
      "eval_loss": 1.5750340223312378,
      "eval_runtime": 44.6873,
      "eval_samples_per_second": 12.576,
      "eval_steps_per_second": 12.576,
      "step": 200072
    },
    {
      "epoch": 90.0,
      "grad_norm": 6.56926931696944e-05,
      "learning_rate": 1.0036476868327403e-06,
      "loss": 0.041,
      "step": 202320
    },
    {
      "epoch": 90.0,
      "eval_accuracy": 0.871886134147644,
      "eval_loss": 1.5229840278625488,
      "eval_runtime": 44.4416,
      "eval_samples_per_second": 12.646,
      "eval_steps_per_second": 12.646,
      "step": 202320
    },
    {
      "epoch": 91.0,
      "grad_norm": 0.00015356726362369955,
      "learning_rate": 9.036921708185054e-07,
      "loss": 0.0165,
      "step": 204568
    },
    {
      "epoch": 91.0,
      "eval_accuracy": 0.8594306111335754,
      "eval_loss": 1.6564196348190308,
      "eval_runtime": 44.2201,
      "eval_samples_per_second": 12.709,
      "eval_steps_per_second": 12.709,
      "step": 204568
    },
    {
      "epoch": 92.0,
      "grad_norm": 8.38369523989968e-05,
      "learning_rate": 8.037811387900357e-07,
      "loss": 0.0478,
      "step": 206816
    },
    {
      "epoch": 92.0,
      "eval_accuracy": 0.8629893064498901,
      "eval_loss": 1.5940439701080322,
      "eval_runtime": 44.433,
      "eval_samples_per_second": 12.648,
      "eval_steps_per_second": 12.648,
      "step": 206816
    },
    {
      "epoch": 93.0,
      "grad_norm": 0.0002615041739773005,
      "learning_rate": 7.037811387900356e-07,
      "loss": 0.032,
      "step": 209064
    },
    {
      "epoch": 93.0,
      "eval_accuracy": 0.8807829022407532,
      "eval_loss": 1.4740957021713257,
      "eval_runtime": 44.2628,
      "eval_samples_per_second": 12.697,
      "eval_steps_per_second": 12.697,
      "step": 209064
    },
    {
      "epoch": 94.0,
      "grad_norm": 6.115555286407471,
      "learning_rate": 6.037811387900357e-07,
      "loss": 0.0433,
      "step": 211312
    },
    {
      "epoch": 94.0,
      "eval_accuracy": 0.871886134147644,
      "eval_loss": 1.5333490371704102,
      "eval_runtime": 44.3613,
      "eval_samples_per_second": 12.669,
      "eval_steps_per_second": 12.669,
      "step": 211312
    },
    {
      "epoch": 95.0,
      "grad_norm": 0.003299184376373887,
      "learning_rate": 5.038256227758007e-07,
      "loss": 0.0243,
      "step": 213560
    },
    {
      "epoch": 95.0,
      "eval_accuracy": 0.871886134147644,
      "eval_loss": 1.5164633989334106,
      "eval_runtime": 44.7905,
      "eval_samples_per_second": 12.547,
      "eval_steps_per_second": 12.547,
      "step": 213560
    },
    {
      "epoch": 96.0,
      "grad_norm": 5.206004061619751e-05,
      "learning_rate": 4.038701067615659e-07,
      "loss": 0.0165,
      "step": 215808
    },
    {
      "epoch": 96.0,
      "eval_accuracy": 0.8683273792266846,
      "eval_loss": 1.5774868726730347,
      "eval_runtime": 44.4019,
      "eval_samples_per_second": 12.657,
      "eval_steps_per_second": 12.657,
      "step": 215808
    },
    {
      "epoch": 97.0,
      "grad_norm": 0.0002649713715072721,
      "learning_rate": 3.03914590747331e-07,
      "loss": 0.0177,
      "step": 218056
    },
    {
      "epoch": 97.0,
      "eval_accuracy": 0.8772242069244385,
      "eval_loss": 1.5302330255508423,
      "eval_runtime": 44.581,
      "eval_samples_per_second": 12.606,
      "eval_steps_per_second": 12.606,
      "step": 218056
    },
    {
      "epoch": 98.0,
      "grad_norm": 5.933016291237436e-05,
      "learning_rate": 2.039590747330961e-07,
      "loss": 0.0253,
      "step": 220304
    },
    {
      "epoch": 98.0,
      "eval_accuracy": 0.8754448294639587,
      "eval_loss": 1.5424113273620605,
      "eval_runtime": 44.8846,
      "eval_samples_per_second": 12.521,
      "eval_steps_per_second": 12.521,
      "step": 220304
    },
    {
      "epoch": 99.0,
      "grad_norm": 7.96098611317575e-05,
      "learning_rate": 1.0400355871886121e-07,
      "loss": 0.0224,
      "step": 222552
    },
    {
      "epoch": 99.0,
      "eval_accuracy": 0.871886134147644,
      "eval_loss": 1.5461581945419312,
      "eval_runtime": 44.4094,
      "eval_samples_per_second": 12.655,
      "eval_steps_per_second": 12.655,
      "step": 222552
    },
    {
      "epoch": 100.0,
      "grad_norm": 0.00019997352501377463,
      "learning_rate": 4.0925266903914596e-09,
      "loss": 0.0213,
      "step": 224800
    },
    {
      "epoch": 100.0,
      "eval_accuracy": 0.873665452003479,
      "eval_loss": 1.5357913970947266,
      "eval_runtime": 44.4377,
      "eval_samples_per_second": 12.647,
      "eval_steps_per_second": 12.647,
      "step": 224800
    },
    {
      "epoch": 100.0,
      "step": 224800,
      "total_flos": 2.049428872704e+19,
      "train_loss": 0.33036238875677576,
      "train_runtime": 38665.4857,
      "train_samples_per_second": 5.814,
      "train_steps_per_second": 5.814
    }
  ],
  "logging_steps": 35,
  "max_steps": 224800,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 100,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.049428872704e+19,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}