{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 371840,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.05,
      "learning_rate": 3.125e-05,
      "loss": 6.2217,
      "step": 1000
    },
    {
      "epoch": 0.11,
      "learning_rate": 6.25e-05,
      "loss": 5.0108,
      "step": 2000
    },
    {
      "epoch": 0.16,
      "learning_rate": 9.375e-05,
      "loss": 4.6837,
      "step": 3000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000125,
      "loss": 4.4551,
      "step": 4000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00015625,
      "loss": 4.2925,
      "step": 5000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0001875,
      "loss": 4.1709,
      "step": 6000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00021875,
      "loss": 4.0678,
      "step": 7000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00025,
      "loss": 3.9755,
      "step": 8000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00028121875,
      "loss": 3.9054,
      "step": 9000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003124375,
      "loss": 3.8472,
      "step": 10000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00034368749999999997,
      "loss": 3.803,
      "step": 11000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0003749375,
      "loss": 3.763,
      "step": 12000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00040615625,
      "loss": 3.7284,
      "step": 13000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00043740625,
      "loss": 3.6975,
      "step": 14000
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00046865625,
      "loss": 3.6673,
      "step": 15000
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.000499875,
      "loss": 3.6516,
      "step": 16000
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.000531125,
      "loss": 3.6242,
      "step": 17000
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.00056234375,
      "loss": 3.6058,
      "step": 18000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.35818437704124967,
      "eval_loss": 3.8078501224517822,
      "eval_runtime": 152.3807,
      "eval_samples_per_second": 380.094,
      "eval_steps_per_second": 5.939,
      "step": 18592
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0005935625,
      "loss": 3.5725,
      "step": 19000
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0006248125,
      "loss": 3.5468,
      "step": 20000
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0006560625,
      "loss": 3.5414,
      "step": 21000
    },
    {
      "epoch": 1.18,
      "learning_rate": 0.00068728125,
      "loss": 3.529,
      "step": 22000
    },
    {
      "epoch": 1.24,
      "learning_rate": 0.00071853125,
      "loss": 3.5128,
      "step": 23000
    },
    {
      "epoch": 1.29,
      "learning_rate": 0.00074978125,
      "loss": 3.5067,
      "step": 24000
    },
    {
      "epoch": 1.34,
      "learning_rate": 0.000781,
      "loss": 3.4941,
      "step": 25000
    },
    {
      "epoch": 1.4,
      "learning_rate": 0.00081225,
      "loss": 3.4825,
      "step": 26000
    },
    {
      "epoch": 1.45,
      "learning_rate": 0.00084346875,
      "loss": 3.4806,
      "step": 27000
    },
    {
      "epoch": 1.51,
      "learning_rate": 0.00087471875,
      "loss": 3.4683,
      "step": 28000
    },
    {
      "epoch": 1.56,
      "learning_rate": 0.00090596875,
      "loss": 3.4639,
      "step": 29000
    },
    {
      "epoch": 1.61,
      "learning_rate": 0.00093721875,
      "loss": 3.4557,
      "step": 30000
    },
    {
      "epoch": 1.67,
      "learning_rate": 0.00096846875,
      "loss": 3.4415,
      "step": 31000
    },
    {
      "epoch": 1.72,
      "learning_rate": 0.0009996875,
      "loss": 3.4396,
      "step": 32000
    },
    {
      "epoch": 1.77,
      "learning_rate": 0.0009970868644067796,
      "loss": 3.4326,
      "step": 33000
    },
    {
      "epoch": 1.83,
      "learning_rate": 0.000994147245762712,
      "loss": 3.417,
      "step": 34000
    },
    {
      "epoch": 1.88,
      "learning_rate": 0.0009912046845574388,
      "loss": 3.4041,
      "step": 35000
    },
    {
      "epoch": 1.94,
      "learning_rate": 0.0009882680084745763,
      "loss": 3.3963,
      "step": 36000
    },
    {
      "epoch": 1.99,
      "learning_rate": 0.0009853254472693032,
      "loss": 3.3918,
      "step": 37000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.38031336262375814,
      "eval_loss": 3.5864098072052,
      "eval_runtime": 155.1129,
      "eval_samples_per_second": 373.399,
      "eval_steps_per_second": 5.834,
      "step": 37184
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.00098238288606403,
      "loss": 3.3388,
      "step": 38000
    },
    {
      "epoch": 2.1,
      "learning_rate": 0.0009794403248587571,
      "loss": 3.3335,
      "step": 39000
    },
    {
      "epoch": 2.15,
      "learning_rate": 0.000976497763653484,
      "loss": 3.3246,
      "step": 40000
    },
    {
      "epoch": 2.21,
      "learning_rate": 0.0009735610875706215,
      "loss": 3.3228,
      "step": 41000
    },
    {
      "epoch": 2.26,
      "learning_rate": 0.0009706185263653485,
      "loss": 3.3153,
      "step": 42000
    },
    {
      "epoch": 2.31,
      "learning_rate": 0.0009676759651600754,
      "loss": 3.3105,
      "step": 43000
    },
    {
      "epoch": 2.37,
      "learning_rate": 0.0009647334039548023,
      "loss": 3.3092,
      "step": 44000
    },
    {
      "epoch": 2.42,
      "learning_rate": 0.0009617937853107345,
      "loss": 3.3021,
      "step": 45000
    },
    {
      "epoch": 2.47,
      "learning_rate": 0.0009588512241054614,
      "loss": 3.2985,
      "step": 46000
    },
    {
      "epoch": 2.53,
      "learning_rate": 0.0009559086629001883,
      "loss": 3.2923,
      "step": 47000
    },
    {
      "epoch": 2.58,
      "learning_rate": 0.0009529690442561205,
      "loss": 3.2919,
      "step": 48000
    },
    {
      "epoch": 2.64,
      "learning_rate": 0.0009500264830508475,
      "loss": 3.2797,
      "step": 49000
    },
    {
      "epoch": 2.69,
      "learning_rate": 0.0009470868644067797,
      "loss": 3.2811,
      "step": 50000
    },
    {
      "epoch": 2.74,
      "learning_rate": 0.0009441443032015066,
      "loss": 3.2732,
      "step": 51000
    },
    {
      "epoch": 2.8,
      "learning_rate": 0.0009412017419962336,
      "loss": 3.2734,
      "step": 52000
    },
    {
      "epoch": 2.85,
      "learning_rate": 0.0009382591807909605,
      "loss": 3.2672,
      "step": 53000
    },
    {
      "epoch": 2.9,
      "learning_rate": 0.000935322504708098,
      "loss": 3.2628,
      "step": 54000
    },
    {
      "epoch": 2.96,
      "learning_rate": 0.0009323799435028248,
      "loss": 3.264,
      "step": 55000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.39303802983141023,
      "eval_loss": 3.4836783409118652,
      "eval_runtime": 154.3521,
      "eval_samples_per_second": 375.239,
      "eval_steps_per_second": 5.863,
      "step": 55776
    },
    {
      "epoch": 3.01,
      "learning_rate": 0.0009294373822975518,
      "loss": 3.2397,
      "step": 56000
    },
    {
      "epoch": 3.07,
      "learning_rate": 0.0009264948210922787,
      "loss": 3.1993,
      "step": 57000
    },
    {
      "epoch": 3.12,
      "learning_rate": 0.0009235552024482109,
      "loss": 3.1972,
      "step": 58000
    },
    {
      "epoch": 3.17,
      "learning_rate": 0.0009206126412429379,
      "loss": 3.1951,
      "step": 59000
    },
    {
      "epoch": 3.23,
      "learning_rate": 0.0009176700800376649,
      "loss": 3.2013,
      "step": 60000
    },
    {
      "epoch": 3.28,
      "learning_rate": 0.0009147275188323917,
      "loss": 3.1991,
      "step": 61000
    },
    {
      "epoch": 3.33,
      "learning_rate": 0.0009117849576271187,
      "loss": 3.1951,
      "step": 62000
    },
    {
      "epoch": 3.39,
      "learning_rate": 0.0009088453389830509,
      "loss": 3.199,
      "step": 63000
    },
    {
      "epoch": 3.44,
      "learning_rate": 0.0009059027777777777,
      "loss": 3.1984,
      "step": 64000
    },
    {
      "epoch": 3.5,
      "learning_rate": 0.00090296315913371,
      "loss": 3.1939,
      "step": 65000
    },
    {
      "epoch": 3.55,
      "learning_rate": 0.000900020597928437,
      "loss": 3.1955,
      "step": 66000
    },
    {
      "epoch": 3.6,
      "learning_rate": 0.0008970780367231639,
      "loss": 3.1906,
      "step": 67000
    },
    {
      "epoch": 3.66,
      "learning_rate": 0.000894138418079096,
      "loss": 3.1877,
      "step": 68000
    },
    {
      "epoch": 3.71,
      "learning_rate": 0.000891195856873823,
      "loss": 3.1859,
      "step": 69000
    },
    {
      "epoch": 3.77,
      "learning_rate": 0.00088825329566855,
      "loss": 3.1838,
      "step": 70000
    },
    {
      "epoch": 3.82,
      "learning_rate": 0.0008853166195856875,
      "loss": 3.1858,
      "step": 71000
    },
    {
      "epoch": 3.87,
      "learning_rate": 0.0008823740583804143,
      "loss": 3.1813,
      "step": 72000
    },
    {
      "epoch": 3.93,
      "learning_rate": 0.0008794314971751413,
      "loss": 3.183,
      "step": 73000
    },
    {
      "epoch": 3.98,
      "learning_rate": 0.0008764918785310735,
      "loss": 3.1794,
      "step": 74000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.39837004355270605,
      "eval_loss": 3.43005633354187,
      "eval_runtime": 154.4263,
      "eval_samples_per_second": 375.059,
      "eval_steps_per_second": 5.86,
      "step": 74368
    },
    {
      "epoch": 4.03,
      "learning_rate": 0.0008735493173258003,
      "loss": 3.1309,
      "step": 75000
    },
    {
      "epoch": 4.09,
      "learning_rate": 0.0008706067561205273,
      "loss": 3.1197,
      "step": 76000
    },
    {
      "epoch": 4.14,
      "learning_rate": 0.0008676641949152543,
      "loss": 3.1191,
      "step": 77000
    },
    {
      "epoch": 4.2,
      "learning_rate": 0.0008647216337099811,
      "loss": 3.1232,
      "step": 78000
    },
    {
      "epoch": 4.25,
      "learning_rate": 0.0008617820150659133,
      "loss": 3.1241,
      "step": 79000
    },
    {
      "epoch": 4.3,
      "learning_rate": 0.0008588394538606403,
      "loss": 3.1253,
      "step": 80000
    },
    {
      "epoch": 4.36,
      "learning_rate": 0.0008558998352165726,
      "loss": 3.1279,
      "step": 81000
    },
    {
      "epoch": 4.41,
      "learning_rate": 0.0008529572740112994,
      "loss": 3.1314,
      "step": 82000
    },
    {
      "epoch": 4.46,
      "learning_rate": 0.0008500147128060264,
      "loss": 3.13,
      "step": 83000
    },
    {
      "epoch": 4.52,
      "learning_rate": 0.0008470750941619586,
      "loss": 3.1235,
      "step": 84000
    },
    {
      "epoch": 4.57,
      "learning_rate": 0.0008441325329566854,
      "loss": 3.1263,
      "step": 85000
    },
    {
      "epoch": 4.63,
      "learning_rate": 0.0008411899717514124,
      "loss": 3.1314,
      "step": 86000
    },
    {
      "epoch": 4.68,
      "learning_rate": 0.0008382474105461394,
      "loss": 3.1293,
      "step": 87000
    },
    {
      "epoch": 4.73,
      "learning_rate": 0.0008353107344632769,
      "loss": 3.1272,
      "step": 88000
    },
    {
      "epoch": 4.79,
      "learning_rate": 0.0008323681732580038,
      "loss": 3.1307,
      "step": 89000
    },
    {
      "epoch": 4.84,
      "learning_rate": 0.0008294256120527307,
      "loss": 3.1257,
      "step": 90000
    },
    {
      "epoch": 4.89,
      "learning_rate": 0.0008264859934086629,
      "loss": 3.1318,
      "step": 91000
    },
    {
      "epoch": 4.95,
      "learning_rate": 0.0008235463747645952,
      "loss": 3.1239,
      "step": 92000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4023224974795406,
      "eval_loss": 3.3843178749084473,
      "eval_runtime": 154.1303,
      "eval_samples_per_second": 375.779,
      "eval_steps_per_second": 5.872,
      "step": 92960
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0008206038135593221,
      "loss": 3.1168,
      "step": 93000
    },
    {
      "epoch": 5.06,
      "learning_rate": 0.000817661252354049,
      "loss": 3.055,
      "step": 94000
    },
    {
      "epoch": 5.11,
      "learning_rate": 0.000814718691148776,
      "loss": 3.0638,
      "step": 95000
    },
    {
      "epoch": 5.16,
      "learning_rate": 0.0008117761299435028,
      "loss": 3.0624,
      "step": 96000
    },
    {
      "epoch": 5.22,
      "learning_rate": 0.0008088335687382298,
      "loss": 3.0712,
      "step": 97000
    },
    {
      "epoch": 5.27,
      "learning_rate": 0.000805893950094162,
      "loss": 3.0734,
      "step": 98000
    },
    {
      "epoch": 5.32,
      "learning_rate": 0.0008029543314500942,
      "loss": 3.0751,
      "step": 99000
    },
    {
      "epoch": 5.38,
      "learning_rate": 0.000800011770244821,
      "loss": 3.0799,
      "step": 100000
    },
    {
      "epoch": 5.43,
      "learning_rate": 0.0007970721516007533,
      "loss": 3.0833,
      "step": 101000
    },
    {
      "epoch": 5.49,
      "learning_rate": 0.0007941295903954803,
      "loss": 3.0781,
      "step": 102000
    },
    {
      "epoch": 5.54,
      "learning_rate": 0.0007911870291902072,
      "loss": 3.0741,
      "step": 103000
    },
    {
      "epoch": 5.59,
      "learning_rate": 0.0007882474105461394,
      "loss": 3.0841,
      "step": 104000
    },
    {
      "epoch": 5.65,
      "learning_rate": 0.0007853048493408663,
      "loss": 3.0833,
      "step": 105000
    },
    {
      "epoch": 5.7,
      "learning_rate": 0.0007823622881355933,
      "loss": 3.0837,
      "step": 106000
    },
    {
      "epoch": 5.76,
      "learning_rate": 0.0007794226694915254,
      "loss": 3.0815,
      "step": 107000
    },
    {
      "epoch": 5.81,
      "learning_rate": 0.0007764801082862523,
      "loss": 3.0827,
      "step": 108000
    },
    {
      "epoch": 5.86,
      "learning_rate": 0.0007735375470809793,
      "loss": 3.0798,
      "step": 109000
    },
    {
      "epoch": 5.92,
      "learning_rate": 0.0007705949858757061,
      "loss": 3.0776,
      "step": 110000
    },
    {
      "epoch": 5.97,
      "learning_rate": 0.0007676553672316384,
      "loss": 3.0814,
      "step": 111000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.4044816244449262,
      "eval_loss": 3.3625779151916504,
      "eval_runtime": 154.4739,
      "eval_samples_per_second": 374.944,
      "eval_steps_per_second": 5.859,
      "step": 111552
    },
    {
      "epoch": 6.02,
      "learning_rate": 0.0007647157485875706,
      "loss": 3.0465,
      "step": 112000
    },
    {
      "epoch": 6.08,
      "learning_rate": 0.0007617731873822976,
      "loss": 3.0156,
      "step": 113000
    },
    {
      "epoch": 6.13,
      "learning_rate": 0.0007588306261770245,
      "loss": 3.019,
      "step": 114000
    },
    {
      "epoch": 6.19,
      "learning_rate": 0.0007558880649717514,
      "loss": 3.0253,
      "step": 115000
    },
    {
      "epoch": 6.24,
      "learning_rate": 0.0007529484463276837,
      "loss": 3.0266,
      "step": 116000
    },
    {
      "epoch": 6.29,
      "learning_rate": 0.0007500058851224106,
      "loss": 3.0332,
      "step": 117000
    },
    {
      "epoch": 6.35,
      "learning_rate": 0.0007470633239171375,
      "loss": 3.0385,
      "step": 118000
    },
    {
      "epoch": 6.4,
      "learning_rate": 0.0007441237052730697,
      "loss": 3.0375,
      "step": 119000
    },
    {
      "epoch": 6.45,
      "learning_rate": 0.0007411811440677966,
      "loss": 3.04,
      "step": 120000
    },
    {
      "epoch": 6.51,
      "learning_rate": 0.0007382385828625235,
      "loss": 3.036,
      "step": 121000
    },
    {
      "epoch": 6.56,
      "learning_rate": 0.0007352960216572505,
      "loss": 3.04,
      "step": 122000
    },
    {
      "epoch": 6.62,
      "learning_rate": 0.000732359345574388,
      "loss": 3.0452,
      "step": 123000
    },
    {
      "epoch": 6.67,
      "learning_rate": 0.0007294167843691149,
      "loss": 3.0428,
      "step": 124000
    },
    {
      "epoch": 6.72,
      "learning_rate": 0.0007264742231638419,
      "loss": 3.0384,
      "step": 125000
    },
    {
      "epoch": 6.78,
      "learning_rate": 0.0007235316619585688,
      "loss": 3.0463,
      "step": 126000
    },
    {
      "epoch": 6.83,
      "learning_rate": 0.000720592043314501,
      "loss": 3.0427,
      "step": 127000
    },
    {
      "epoch": 6.88,
      "learning_rate": 0.0007176494821092279,
      "loss": 3.0505,
      "step": 128000
    },
    {
      "epoch": 6.94,
      "learning_rate": 0.0007147069209039548,
      "loss": 3.0484,
      "step": 129000
    },
    {
      "epoch": 6.99,
      "learning_rate": 0.0007117643596986817,
      "loss": 3.0416,
      "step": 130000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.4076348271969515,
      "eval_loss": 3.347128391265869,
      "eval_runtime": 154.2375,
      "eval_samples_per_second": 375.518,
      "eval_steps_per_second": 5.868,
      "step": 130144
    },
    {
      "epoch": 7.05,
      "learning_rate": 0.0007088217984934087,
      "loss": 2.9862,
      "step": 131000
    },
    {
      "epoch": 7.1,
      "learning_rate": 0.0007058821798493408,
      "loss": 2.9834,
      "step": 132000
    },
    {
      "epoch": 7.15,
      "learning_rate": 0.0007029396186440678,
      "loss": 2.9881,
      "step": 133000
    },
    {
      "epoch": 7.21,
      "learning_rate": 0.0007,
      "loss": 2.9905,
      "step": 134000
    },
    {
      "epoch": 7.26,
      "learning_rate": 0.000697057438794727,
      "loss": 2.9955,
      "step": 135000
    },
    {
      "epoch": 7.31,
      "learning_rate": 0.0006941178201506591,
      "loss": 2.9995,
      "step": 136000
    },
    {
      "epoch": 7.37,
      "learning_rate": 0.0006911752589453861,
      "loss": 3.0034,
      "step": 137000
    },
    {
      "epoch": 7.42,
      "learning_rate": 0.0006882356403013182,
      "loss": 3.0028,
      "step": 138000
    },
    {
      "epoch": 7.48,
      "learning_rate": 0.0006852930790960453,
      "loss": 3.0054,
      "step": 139000
    },
    {
      "epoch": 7.53,
      "learning_rate": 0.0006823534604519775,
      "loss": 3.0097,
      "step": 140000
    },
    {
      "epoch": 7.58,
      "learning_rate": 0.0006794108992467043,
      "loss": 3.0091,
      "step": 141000
    },
    {
      "epoch": 7.64,
      "learning_rate": 0.0006764683380414313,
      "loss": 3.0086,
      "step": 142000
    },
    {
      "epoch": 7.69,
      "learning_rate": 0.0006735257768361582,
      "loss": 3.007,
      "step": 143000
    },
    {
      "epoch": 7.75,
      "learning_rate": 0.0006705861581920904,
      "loss": 3.0159,
      "step": 144000
    },
    {
      "epoch": 7.8,
      "learning_rate": 0.0006676435969868173,
      "loss": 3.0153,
      "step": 145000
    },
    {
      "epoch": 7.85,
      "learning_rate": 0.0006647010357815442,
      "loss": 3.0126,
      "step": 146000
    },
    {
      "epoch": 7.91,
      "learning_rate": 0.0006617584745762712,
      "loss": 3.0126,
      "step": 147000
    },
    {
      "epoch": 7.96,
      "learning_rate": 0.0006588188559322034,
      "loss": 3.0128,
      "step": 148000
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.40787197577617956,
      "eval_loss": 3.3522324562072754,
      "eval_runtime": 154.1725,
      "eval_samples_per_second": 375.677,
      "eval_steps_per_second": 5.87,
      "step": 148736
    },
    {
      "epoch": 8.01,
      "learning_rate": 0.0006558762947269304,
      "loss": 2.9903,
      "step": 149000
    },
    {
      "epoch": 8.07,
      "learning_rate": 0.0006529366760828626,
      "loss": 2.949,
      "step": 150000
    },
    {
      "epoch": 8.12,
      "learning_rate": 0.0006499941148775894,
      "loss": 2.9556,
      "step": 151000
    },
    {
      "epoch": 8.18,
      "learning_rate": 0.0006470544962335216,
      "loss": 2.9566,
      "step": 152000
    },
    {
      "epoch": 8.23,
      "learning_rate": 0.0006441119350282486,
      "loss": 2.9633,
      "step": 153000
    },
    {
      "epoch": 8.28,
      "learning_rate": 0.0006411693738229755,
      "loss": 2.9658,
      "step": 154000
    },
    {
      "epoch": 8.34,
      "learning_rate": 0.0006382297551789077,
      "loss": 2.9686,
      "step": 155000
    },
    {
      "epoch": 8.39,
      "learning_rate": 0.0006352871939736347,
      "loss": 2.9699,
      "step": 156000
    },
    {
      "epoch": 8.44,
      "learning_rate": 0.0006323475753295669,
      "loss": 2.9767,
      "step": 157000
    },
    {
      "epoch": 8.5,
      "learning_rate": 0.0006294050141242938,
      "loss": 2.9748,
      "step": 158000
    },
    {
      "epoch": 8.55,
      "learning_rate": 0.0006264624529190207,
      "loss": 2.9814,
      "step": 159000
    },
    {
      "epoch": 8.61,
      "learning_rate": 0.0006235198917137477,
      "loss": 2.9856,
      "step": 160000
    },
    {
      "epoch": 8.66,
      "learning_rate": 0.0006205802730696798,
      "loss": 2.9811,
      "step": 161000
    },
    {
      "epoch": 8.71,
      "learning_rate": 0.000617640654425612,
      "loss": 2.9811,
      "step": 162000
    },
    {
      "epoch": 8.77,
      "learning_rate": 0.000614698093220339,
      "loss": 2.9817,
      "step": 163000
    },
    {
      "epoch": 8.82,
      "learning_rate": 0.000611755532015066,
      "loss": 2.9823,
      "step": 164000
    },
    {
      "epoch": 8.87,
      "learning_rate": 0.0006088129708097928,
      "loss": 2.9848,
      "step": 165000
    },
    {
      "epoch": 8.93,
      "learning_rate": 0.000605873352165725,
      "loss": 2.9871,
      "step": 166000
    },
    {
      "epoch": 8.98,
      "learning_rate": 0.000602930790960452,
      "loss": 2.9879,
      "step": 167000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.4086848646738169,
      "eval_loss": 3.349667549133301,
      "eval_runtime": 154.1992,
      "eval_samples_per_second": 375.612,
      "eval_steps_per_second": 5.869,
      "step": 167328
    },
    {
      "epoch": 9.04,
      "learning_rate": 0.0005999882297551789,
      "loss": 2.9388,
      "step": 168000
    },
    {
      "epoch": 9.09,
      "learning_rate": 0.0005970486111111111,
      "loss": 2.9269,
      "step": 169000
    },
    {
      "epoch": 9.14,
      "learning_rate": 0.0005941060499058381,
      "loss": 2.9299,
      "step": 170000
    },
    {
      "epoch": 9.2,
      "learning_rate": 0.0005911664312617703,
      "loss": 2.933,
      "step": 171000
    },
    {
      "epoch": 9.25,
      "learning_rate": 0.0005882238700564971,
      "loss": 2.9403,
      "step": 172000
    },
    {
      "epoch": 9.31,
      "learning_rate": 0.0005852813088512241,
      "loss": 2.9383,
      "step": 173000
    },
    {
      "epoch": 9.36,
      "learning_rate": 0.0005823387476459511,
      "loss": 2.9456,
      "step": 174000
    },
    {
      "epoch": 9.41,
      "learning_rate": 0.0005793991290018833,
      "loss": 2.9448,
      "step": 175000
    },
    {
      "epoch": 9.47,
      "learning_rate": 0.0005764565677966101,
      "loss": 2.9461,
      "step": 176000
    },
    {
      "epoch": 9.52,
      "learning_rate": 0.0005735140065913371,
      "loss": 2.9527,
      "step": 177000
    },
    {
      "epoch": 9.57,
      "learning_rate": 0.0005705743879472694,
      "loss": 2.9517,
      "step": 178000
    },
    {
      "epoch": 9.63,
      "learning_rate": 0.0005676347693032016,
      "loss": 2.9524,
      "step": 179000
    },
    {
      "epoch": 9.68,
      "learning_rate": 0.0005646922080979284,
      "loss": 2.9597,
      "step": 180000
    },
    {
      "epoch": 9.74,
      "learning_rate": 0.0005617496468926554,
      "loss": 2.9558,
      "step": 181000
    },
    {
      "epoch": 9.79,
      "learning_rate": 0.0005588070856873822,
      "loss": 2.9596,
      "step": 182000
    },
    {
      "epoch": 9.84,
      "learning_rate": 0.0005558674670433144,
      "loss": 2.9595,
      "step": 183000
    },
    {
      "epoch": 9.9,
      "learning_rate": 0.0005529249058380414,
      "loss": 2.9628,
      "step": 184000
    },
    {
      "epoch": 9.95,
      "learning_rate": 0.0005499823446327685,
      "loss": 2.9616,
      "step": 185000
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.41226580822016096,
      "eval_loss": 3.3192811012268066,
      "eval_runtime": 154.2887,
      "eval_samples_per_second": 375.394,
      "eval_steps_per_second": 5.866,
      "step": 185920
    },
    {
      "epoch": 10.0,
      "learning_rate": 0.0005470397834274953,
      "loss": 2.9548,
      "step": 186000
    },
    {
      "epoch": 10.06,
      "learning_rate": 0.0005441031073446327,
      "loss": 2.8968,
      "step": 187000
    },
    {
      "epoch": 10.11,
      "learning_rate": 0.0005411605461393597,
      "loss": 2.9016,
      "step": 188000
    },
    {
      "epoch": 10.17,
      "learning_rate": 0.0005382179849340867,
      "loss": 2.9077,
      "step": 189000
    },
    {
      "epoch": 10.22,
      "learning_rate": 0.0005352754237288135,
      "loss": 2.9131,
      "step": 190000
    },
    {
      "epoch": 10.27,
      "learning_rate": 0.0005323328625235405,
      "loss": 2.9148,
      "step": 191000
    },
    {
      "epoch": 10.33,
      "learning_rate": 0.0005293903013182673,
      "loss": 2.9188,
      "step": 192000
    },
    {
      "epoch": 10.38,
      "learning_rate": 0.0005264506826741996,
      "loss": 2.9221,
      "step": 193000
    },
    {
      "epoch": 10.43,
      "learning_rate": 0.0005235081214689266,
      "loss": 2.9241,
      "step": 194000
    },
    {
      "epoch": 10.49,
      "learning_rate": 0.0005205655602636536,
      "loss": 2.9225,
      "step": 195000
    },
    {
      "epoch": 10.54,
      "learning_rate": 0.0005176229990583804,
      "loss": 2.9246,
      "step": 196000
    },
    {
      "epoch": 10.6,
      "learning_rate": 0.0005146833804143126,
      "loss": 2.9304,
      "step": 197000
    },
    {
      "epoch": 10.65,
      "learning_rate": 0.0005117437617702448,
      "loss": 2.9316,
      "step": 198000
    },
    {
      "epoch": 10.7,
      "learning_rate": 0.0005088012005649718,
      "loss": 2.9305,
      "step": 199000
    },
    {
      "epoch": 10.76,
      "learning_rate": 0.0005058586393596986,
      "loss": 2.933,
      "step": 200000
    },
    {
      "epoch": 10.81,
      "learning_rate": 0.0005029190207156309,
      "loss": 2.9344,
      "step": 201000
    },
    {
      "epoch": 10.86,
      "learning_rate": 0.0004999764595103578,
      "loss": 2.9375,
      "step": 202000
    },
    {
      "epoch": 10.92,
      "learning_rate": 0.0004970368408662901,
      "loss": 2.9357,
      "step": 203000
    },
    {
      "epoch": 10.97,
      "learning_rate": 0.0004940942796610169,
      "loss": 2.941,
      "step": 204000
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.4113357558318228,
      "eval_loss": 3.338052988052368,
      "eval_runtime": 154.1151,
      "eval_samples_per_second": 375.817,
      "eval_steps_per_second": 5.872,
      "step": 204512
    },
    {
      "epoch": 11.03,
      "learning_rate": 0.0004911546610169491,
      "loss": 2.9097,
      "step": 205000
    },
    {
      "epoch": 11.08,
      "learning_rate": 0.0004882120998116761,
      "loss": 2.8766,
      "step": 206000
    },
    {
      "epoch": 11.13,
      "learning_rate": 0.00048526953860640305,
      "loss": 2.8791,
      "step": 207000
    },
    {
      "epoch": 11.19,
      "learning_rate": 0.00048232991996233524,
      "loss": 2.8864,
      "step": 208000
    },
    {
      "epoch": 11.24,
      "learning_rate": 0.00047938735875706215,
      "loss": 2.888,
      "step": 209000
    },
    {
      "epoch": 11.3,
      "learning_rate": 0.0004764447975517891,
      "loss": 2.8976,
      "step": 210000
    },
    {
      "epoch": 11.35,
      "learning_rate": 0.0004735081214689266,
      "loss": 2.8996,
      "step": 211000
    },
    {
      "epoch": 11.4,
      "learning_rate": 0.0004705655602636535,
      "loss": 2.9,
      "step": 212000
    },
    {
      "epoch": 11.46,
      "learning_rate": 0.0004676229990583804,
      "loss": 2.9,
      "step": 213000
    },
    {
      "epoch": 11.51,
      "learning_rate": 0.00046468043785310737,
      "loss": 2.9033,
      "step": 214000
    },
    {
      "epoch": 11.56,
      "learning_rate": 0.00046174081920903956,
      "loss": 2.9067,
      "step": 215000
    },
    {
      "epoch": 11.62,
      "learning_rate": 0.0004587982580037665,
      "loss": 2.9069,
      "step": 216000
    },
    {
      "epoch": 11.67,
      "learning_rate": 0.00045585863935969866,
      "loss": 2.9136,
      "step": 217000
    },
    {
      "epoch": 11.73,
      "learning_rate": 0.0004529160781544256,
      "loss": 2.9132,
      "step": 218000
    },
    {
      "epoch": 11.78,
      "learning_rate": 0.0004499764595103578,
      "loss": 2.9136,
      "step": 219000
    },
    {
      "epoch": 11.83,
      "learning_rate": 0.0004470338983050847,
      "loss": 2.91,
      "step": 220000
    },
    {
      "epoch": 11.89,
      "learning_rate": 0.0004440913370998117,
      "loss": 2.9103,
      "step": 221000
    },
    {
      "epoch": 11.94,
      "learning_rate": 0.0004411487758945386,
      "loss": 2.9183,
      "step": 222000
    },
    {
      "epoch": 11.99,
      "learning_rate": 0.0004382062146892655,
      "loss": 2.9156,
      "step": 223000
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.41138298400496653,
      "eval_loss": 3.347946882247925,
      "eval_runtime": 154.2621,
      "eval_samples_per_second": 375.458,
      "eval_steps_per_second": 5.867,
      "step": 223104
    },
    {
      "epoch": 12.05,
      "learning_rate": 0.00043526659604519775,
      "loss": 2.8568,
      "step": 224000
    },
    {
      "epoch": 12.1,
      "learning_rate": 0.00043232403483992466,
      "loss": 2.8592,
      "step": 225000
    },
    {
      "epoch": 12.16,
      "learning_rate": 0.0004293844161958569,
      "loss": 2.8674,
      "step": 226000
    },
    {
      "epoch": 12.21,
      "learning_rate": 0.0004264418549905838,
      "loss": 2.87,
      "step": 227000
    },
    {
      "epoch": 12.26,
      "learning_rate": 0.000423502236346516,
      "loss": 2.8674,
      "step": 228000
    },
    {
      "epoch": 12.32,
      "learning_rate": 0.0004205596751412429,
      "loss": 2.8738,
      "step": 229000
    },
    {
      "epoch": 12.37,
      "learning_rate": 0.0004176171139359698,
      "loss": 2.8776,
      "step": 230000
    },
    {
      "epoch": 12.42,
      "learning_rate": 0.00041467749529190207,
      "loss": 2.8789,
      "step": 231000
    },
    {
      "epoch": 12.48,
      "learning_rate": 0.000411734934086629,
      "loss": 2.8833,
      "step": 232000
    },
    {
      "epoch": 12.53,
      "learning_rate": 0.00040879237288135595,
      "loss": 2.8848,
      "step": 233000
    },
    {
      "epoch": 12.59,
      "learning_rate": 0.00040584981167608286,
      "loss": 2.8865,
      "step": 234000
    },
    {
      "epoch": 12.64,
      "learning_rate": 0.00040290725047080977,
      "loss": 2.8891,
      "step": 235000
    },
    {
      "epoch": 12.69,
      "learning_rate": 0.000399967631826742,
      "loss": 2.8939,
      "step": 236000
    },
    {
      "epoch": 12.75,
      "learning_rate": 0.0003970250706214689,
      "loss": 2.889,
      "step": 237000
    },
    {
      "epoch": 12.8,
      "learning_rate": 0.00039408545197740116,
      "loss": 2.8899,
      "step": 238000
    },
    {
      "epoch": 12.85,
      "learning_rate": 0.0003911428907721281,
      "loss": 2.8905,
      "step": 239000
    },
    {
      "epoch": 12.91,
      "learning_rate": 0.00038820327212806026,
      "loss": 2.8949,
      "step": 240000
    },
    {
      "epoch": 12.96,
      "learning_rate": 0.00038526365348399245,
      "loss": 2.8946,
      "step": 241000
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.4130300581457413,
      "eval_loss": 3.3280186653137207,
      "eval_runtime": 154.4059,
      "eval_samples_per_second": 375.109,
      "eval_steps_per_second": 5.861,
      "step": 241696
    },
    {
      "epoch": 13.02,
      "learning_rate": 0.0003823210922787194,
      "loss": 2.8749,
      "step": 242000
    },
    {
      "epoch": 13.07,
      "learning_rate": 0.0003793785310734463,
      "loss": 2.8369,
      "step": 243000
    },
    {
      "epoch": 13.12,
      "learning_rate": 0.00037643596986817324,
      "loss": 2.8469,
      "step": 244000
    },
    {
      "epoch": 13.18,
      "learning_rate": 0.0003734934086629002,
      "loss": 2.8488,
      "step": 245000
    },
    {
      "epoch": 13.23,
      "learning_rate": 0.0003705537900188324,
      "loss": 2.8528,
      "step": 246000
    },
    {
      "epoch": 13.29,
      "learning_rate": 0.00036761122881355936,
      "loss": 2.8522,
      "step": 247000
    },
    {
      "epoch": 13.34,
      "learning_rate": 0.00036467161016949154,
      "loss": 2.8577,
      "step": 248000
    },
    {
      "epoch": 13.39,
      "learning_rate": 0.0003617290489642185,
      "loss": 2.8556,
      "step": 249000
    },
    {
      "epoch": 13.45,
      "learning_rate": 0.0003587864877589454,
      "loss": 2.8629,
      "step": 250000
    },
    {
      "epoch": 13.5,
      "learning_rate": 0.00035584686911487755,
      "loss": 2.8615,
      "step": 251000
    },
    {
      "epoch": 13.55,
      "learning_rate": 0.0003529043079096045,
      "loss": 2.8621,
      "step": 252000
    },
    {
      "epoch": 13.61,
      "learning_rate": 0.00034996174670433143,
      "loss": 2.8699,
      "step": 253000
    },
    {
      "epoch": 13.66,
      "learning_rate": 0.00034701918549905834,
      "loss": 2.8663,
      "step": 254000
    },
    {
      "epoch": 13.72,
      "learning_rate": 0.0003440795668549906,
      "loss": 2.8708,
      "step": 255000
    },
    {
      "epoch": 13.77,
      "learning_rate": 0.0003411370056497175,
      "loss": 2.8699,
      "step": 256000
    },
    {
      "epoch": 13.82,
      "learning_rate": 0.00033819738700564974,
      "loss": 2.8675,
      "step": 257000
    },
    {
      "epoch": 13.88,
      "learning_rate": 0.00033525482580037665,
      "loss": 2.8698,
      "step": 258000
    },
    {
      "epoch": 13.93,
      "learning_rate": 0.00033231520715630884,
      "loss": 2.8735,
      "step": 259000
    },
    {
      "epoch": 13.98,
      "learning_rate": 0.0003293726459510358,
      "loss": 2.8744,
      "step": 260000
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.4122856937667478,
      "eval_loss": 3.3444502353668213,
      "eval_runtime": 154.5486,
      "eval_samples_per_second": 374.762,
      "eval_steps_per_second": 5.856,
      "step": 260288
    },
    {
      "epoch": 14.04,
      "learning_rate": 0.00032643008474576277,
      "loss": 2.8318,
      "step": 261000
    },
    {
      "epoch": 14.09,
      "learning_rate": 0.0003234875235404897,
      "loss": 2.8257,
      "step": 262000
    },
    {
      "epoch": 14.15,
      "learning_rate": 0.0003205479048964218,
      "loss": 2.8257,
      "step": 263000
    },
    {
      "epoch": 14.2,
      "learning_rate": 0.0003176053436911488,
      "loss": 2.834,
      "step": 264000
    },
    {
      "epoch": 14.25,
      "learning_rate": 0.00031466572504708096,
      "loss": 2.8316,
      "step": 265000
    },
    {
      "epoch": 14.31,
      "learning_rate": 0.00031172316384180793,
      "loss": 2.8371,
      "step": 266000
    },
    {
      "epoch": 14.36,
      "learning_rate": 0.00030878060263653484,
      "loss": 2.8336,
      "step": 267000
    },
    {
      "epoch": 14.41,
      "learning_rate": 0.0003058409839924671,
      "loss": 2.8406,
      "step": 268000
    },
    {
      "epoch": 14.47,
      "learning_rate": 0.000302898422787194,
      "loss": 2.842,
      "step": 269000
    },
    {
      "epoch": 14.52,
      "learning_rate": 0.0002999558615819209,
      "loss": 2.8453,
      "step": 270000
    },
    {
      "epoch": 14.58,
      "learning_rate": 0.00029701624293785315,
      "loss": 2.8445,
      "step": 271000
    },
    {
      "epoch": 14.63,
      "learning_rate": 0.00029407368173258006,
      "loss": 2.8511,
      "step": 272000
    },
    {
      "epoch": 14.68,
      "learning_rate": 0.00029113112052730697,
      "loss": 2.8493,
      "step": 273000
    },
    {
      "epoch": 14.74,
      "learning_rate": 0.00028819150188323916,
      "loss": 2.8474,
      "step": 274000
    },
    {
      "epoch": 14.79,
      "learning_rate": 0.0002852518832391714,
      "loss": 2.8523,
      "step": 275000
    },
    {
      "epoch": 14.85,
      "learning_rate": 0.0002823093220338983,
      "loss": 2.8517,
      "step": 276000
    },
    {
      "epoch": 14.9,
      "learning_rate": 0.0002793667608286252,
      "loss": 2.8556,
      "step": 277000
    },
    {
      "epoch": 14.95,
      "learning_rate": 0.00027642714218455746,
      "loss": 2.8532,
      "step": 278000
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.41189805997010587,
      "eval_loss": 3.3571324348449707,
      "eval_runtime": 154.1251,
      "eval_samples_per_second": 375.792,
      "eval_steps_per_second": 5.872,
      "step": 278880
    },
    {
      "epoch": 15.01,
      "learning_rate": 0.0002734845809792844,
      "loss": 2.8476,
      "step": 279000
    },
    {
      "epoch": 15.06,
      "learning_rate": 0.00027054201977401134,
      "loss": 2.8033,
      "step": 280000
    },
    {
      "epoch": 15.11,
      "learning_rate": 0.00026759945856873825,
      "loss": 2.8064,
      "step": 281000
    },
    {
      "epoch": 15.17,
      "learning_rate": 0.00026465689736346516,
      "loss": 2.8148,
      "step": 282000
    },
    {
      "epoch": 15.22,
      "learning_rate": 0.0002617172787193974,
      "loss": 2.815,
      "step": 283000
    },
    {
      "epoch": 15.28,
      "learning_rate": 0.0002587747175141243,
      "loss": 2.815,
      "step": 284000
    },
    {
      "epoch": 15.33,
      "learning_rate": 0.0002558321563088512,
      "loss": 2.821,
      "step": 285000
    },
    {
      "epoch": 15.38,
      "learning_rate": 0.0002528895951035782,
      "loss": 2.8189,
      "step": 286000
    },
    {
      "epoch": 15.44,
      "learning_rate": 0.0002499499764595104,
      "loss": 2.8231,
      "step": 287000
    },
    {
      "epoch": 15.49,
      "learning_rate": 0.00024701035781544257,
      "loss": 2.8291,
      "step": 288000
    },
    {
      "epoch": 15.54,
      "learning_rate": 0.0002440677966101695,
      "loss": 2.828,
      "step": 289000
    },
    {
      "epoch": 15.6,
      "learning_rate": 0.0002411281779661017,
      "loss": 2.827,
      "step": 290000
    },
    {
      "epoch": 15.65,
      "learning_rate": 0.0002381885593220339,
      "loss": 2.8286,
      "step": 291000
    },
    {
      "epoch": 15.71,
      "learning_rate": 0.00023524599811676085,
      "loss": 2.8287,
      "step": 292000
    },
    {
      "epoch": 15.76,
      "learning_rate": 0.00023230343691148776,
      "loss": 2.8319,
      "step": 293000
    },
    {
      "epoch": 15.81,
      "learning_rate": 0.00022936381826741997,
      "loss": 2.8359,
      "step": 294000
    },
    {
      "epoch": 15.87,
      "learning_rate": 0.00022642125706214688,
      "loss": 2.8349,
      "step": 295000
    },
    {
      "epoch": 15.92,
      "learning_rate": 0.00022347869585687382,
      "loss": 2.8312,
      "step": 296000
    },
    {
      "epoch": 15.97,
      "learning_rate": 0.00022053613465160073,
      "loss": 2.831,
      "step": 297000
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.412203329982574,
      "eval_loss": 3.362921953201294,
      "eval_runtime": 154.0833,
      "eval_samples_per_second": 375.894,
      "eval_steps_per_second": 5.873,
      "step": 297472
    },
    {
      "epoch": 16.03,
      "learning_rate": 0.00021759651600753298,
      "loss": 2.8063,
      "step": 298000
    },
    {
      "epoch": 16.08,
      "learning_rate": 0.00021465395480225989,
      "loss": 2.7925,
      "step": 299000
    },
    {
      "epoch": 16.14,
      "learning_rate": 0.00021171139359698682,
      "loss": 2.7942,
      "step": 300000
    },
    {
      "epoch": 16.19,
      "learning_rate": 0.000208771774952919,
      "loss": 2.7952,
      "step": 301000
    },
    {
      "epoch": 16.24,
      "learning_rate": 0.00020582921374764595,
      "loss": 2.8007,
      "step": 302000
    },
    {
      "epoch": 16.3,
      "learning_rate": 0.00020288959510357817,
      "loss": 2.8039,
      "step": 303000
    },
    {
      "epoch": 16.35,
      "learning_rate": 0.0001999470338983051,
      "loss": 2.8051,
      "step": 304000
    },
    {
      "epoch": 16.4,
      "learning_rate": 0.00019700447269303201,
      "loss": 2.806,
      "step": 305000
    },
    {
      "epoch": 16.46,
      "learning_rate": 0.00019406191148775895,
      "loss": 2.8044,
      "step": 306000
    },
    {
      "epoch": 16.51,
      "learning_rate": 0.00019112523540489642,
      "loss": 2.806,
      "step": 307000
    },
    {
      "epoch": 16.57,
      "learning_rate": 0.00018818267419962336,
      "loss": 2.8064,
      "step": 308000
    },
    {
      "epoch": 16.62,
      "learning_rate": 0.0001852401129943503,
      "loss": 2.8072,
      "step": 309000
    },
    {
      "epoch": 16.67,
      "learning_rate": 0.00018229755178907723,
      "loss": 2.8126,
      "step": 310000
    },
    {
      "epoch": 16.73,
      "learning_rate": 0.00017935793314500942,
      "loss": 2.8149,
      "step": 311000
    },
    {
      "epoch": 16.78,
      "learning_rate": 0.00017641537193973636,
      "loss": 2.8102,
      "step": 312000
    },
    {
      "epoch": 16.84,
      "learning_rate": 0.00017347281073446327,
      "loss": 2.8137,
      "step": 313000
    },
    {
      "epoch": 16.89,
      "learning_rate": 0.00017053319209039549,
      "loss": 2.8149,
      "step": 314000
    },
    {
      "epoch": 16.94,
      "learning_rate": 0.00016759063088512242,
      "loss": 2.8155,
      "step": 315000
    },
    {
      "epoch": 17.0,
      "learning_rate": 0.00016464806967984936,
      "loss": 2.8168,
      "step": 316000
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.4121428671720059,
      "eval_loss": 3.3628885746002197,
      "eval_runtime": 154.2564,
      "eval_samples_per_second": 375.472,
      "eval_steps_per_second": 5.867,
      "step": 316064
    },
    {
      "epoch": 17.05,
      "learning_rate": 0.00016170845103578155,
      "loss": 2.7786,
      "step": 317000
    },
    {
      "epoch": 17.1,
      "learning_rate": 0.00015876883239171374,
      "loss": 2.7814,
      "step": 318000
    },
    {
      "epoch": 17.16,
      "learning_rate": 0.00015582627118644068,
      "loss": 2.785,
      "step": 319000
    },
    {
      "epoch": 17.21,
      "learning_rate": 0.0001528837099811676,
      "loss": 2.7861,
      "step": 320000
    },
    {
      "epoch": 17.27,
      "learning_rate": 0.00014994114877589455,
      "loss": 2.7827,
      "step": 321000
    },
    {
      "epoch": 17.32,
      "learning_rate": 0.00014700153013182674,
      "loss": 2.7892,
      "step": 322000
    },
    {
      "epoch": 17.37,
      "learning_rate": 0.00014405896892655368,
      "loss": 2.7889,
      "step": 323000
    },
    {
      "epoch": 17.43,
      "learning_rate": 0.00014111935028248587,
      "loss": 2.7877,
      "step": 324000
    },
    {
      "epoch": 17.48,
      "learning_rate": 0.0001381767890772128,
      "loss": 2.7861,
      "step": 325000
    },
    {
      "epoch": 17.53,
      "learning_rate": 0.00013523422787193974,
      "loss": 2.7889,
      "step": 326000
    },
    {
      "epoch": 17.59,
      "learning_rate": 0.00013229166666666668,
      "loss": 2.7912,
      "step": 327000
    },
    {
      "epoch": 17.64,
      "learning_rate": 0.00012934910546139362,
      "loss": 2.7903,
      "step": 328000
    },
    {
      "epoch": 17.7,
      "learning_rate": 0.00012640654425612053,
      "loss": 2.7922,
      "step": 329000
    },
    {
      "epoch": 17.75,
      "learning_rate": 0.00012346692561205274,
      "loss": 2.7947,
      "step": 330000
    },
    {
      "epoch": 17.8,
      "learning_rate": 0.00012052730696798493,
      "loss": 2.7981,
      "step": 331000
    },
    {
      "epoch": 17.86,
      "learning_rate": 0.00011758474576271187,
      "loss": 2.7932,
      "step": 332000
    },
    {
      "epoch": 17.91,
      "learning_rate": 0.0001146421845574388,
      "loss": 2.7986,
      "step": 333000
    },
    {
      "epoch": 17.96,
      "learning_rate": 0.00011169962335216573,
      "loss": 2.7943,
      "step": 334000
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.4118951040104781,
      "eval_loss": 3.3743107318878174,
      "eval_runtime": 153.9729,
      "eval_samples_per_second": 376.164,
      "eval_steps_per_second": 5.878,
      "step": 334656
    },
    {
      "epoch": 18.02,
      "learning_rate": 0.00010875706214689266,
      "loss": 2.7857,
      "step": 335000
    },
    {
      "epoch": 18.07,
      "learning_rate": 0.00010581744350282486,
      "loss": 2.7674,
      "step": 336000
    },
    {
      "epoch": 18.13,
      "learning_rate": 0.0001028748822975518,
      "loss": 2.7639,
      "step": 337000
    },
    {
      "epoch": 18.18,
      "learning_rate": 9.993232109227872e-05,
      "loss": 2.7699,
      "step": 338000
    },
    {
      "epoch": 18.23,
      "learning_rate": 9.699270244821092e-05,
      "loss": 2.772,
      "step": 339000
    },
    {
      "epoch": 18.29,
      "learning_rate": 9.405014124293786e-05,
      "loss": 2.7725,
      "step": 340000
    },
    {
      "epoch": 18.34,
      "learning_rate": 9.111052259887005e-05,
      "loss": 2.7695,
      "step": 341000
    },
    {
      "epoch": 18.4,
      "learning_rate": 8.817090395480227e-05,
      "loss": 2.772,
      "step": 342000
    },
    {
      "epoch": 18.45,
      "learning_rate": 8.522834274952919e-05,
      "loss": 2.7731,
      "step": 343000
    },
    {
      "epoch": 18.5,
      "learning_rate": 8.228578154425611e-05,
      "loss": 2.7791,
      "step": 344000
    },
    {
      "epoch": 18.56,
      "learning_rate": 7.934322033898306e-05,
      "loss": 2.7715,
      "step": 345000
    },
    {
      "epoch": 18.61,
      "learning_rate": 7.640065913370999e-05,
      "loss": 2.7788,
      "step": 346000
    },
    {
      "epoch": 18.66,
      "learning_rate": 7.346104048964218e-05,
      "loss": 2.7768,
      "step": 347000
    },
    {
      "epoch": 18.72,
      "learning_rate": 7.051847928436913e-05,
      "loss": 2.7737,
      "step": 348000
    },
    {
      "epoch": 18.77,
      "learning_rate": 6.757591807909605e-05,
      "loss": 2.7762,
      "step": 349000
    },
    {
      "epoch": 18.83,
      "learning_rate": 6.463629943502824e-05,
      "loss": 2.776,
      "step": 350000
    },
    {
      "epoch": 18.88,
      "learning_rate": 6.169373822975518e-05,
      "loss": 2.7788,
      "step": 351000
    },
    {
      "epoch": 18.93,
      "learning_rate": 5.875411958568738e-05,
      "loss": 2.7776,
      "step": 352000
    },
    {
      "epoch": 18.99,
      "learning_rate": 5.581155838041432e-05,
      "loss": 2.7777,
      "step": 353000
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.412103499164236,
      "eval_loss": 3.378060817718506,
      "eval_runtime": 154.0655,
      "eval_samples_per_second": 375.937,
      "eval_steps_per_second": 5.874,
      "step": 353248
    },
    {
      "epoch": 19.04,
      "learning_rate": 5.286899717514124e-05,
      "loss": 2.7602,
      "step": 354000
    },
    {
      "epoch": 19.09,
      "learning_rate": 4.9926435969868174e-05,
      "loss": 2.757,
      "step": 355000
    },
    {
      "epoch": 19.15,
      "learning_rate": 4.698975988700565e-05,
      "loss": 2.7585,
      "step": 356000
    },
    {
      "epoch": 19.2,
      "learning_rate": 4.4047198681732586e-05,
      "loss": 2.7613,
      "step": 357000
    },
    {
      "epoch": 19.26,
      "learning_rate": 4.110463747645951e-05,
      "loss": 2.7583,
      "step": 358000
    },
    {
      "epoch": 19.31,
      "learning_rate": 3.816796139359699e-05,
      "loss": 2.758,
      "step": 359000
    },
    {
      "epoch": 19.36,
      "learning_rate": 3.5225400188323915e-05,
      "loss": 2.7611,
      "step": 360000
    },
    {
      "epoch": 19.42,
      "learning_rate": 3.228283898305085e-05,
      "loss": 2.7623,
      "step": 361000
    },
    {
      "epoch": 19.47,
      "learning_rate": 2.934027777777778e-05,
      "loss": 2.7581,
      "step": 362000
    },
    {
      "epoch": 19.52,
      "learning_rate": 2.639771657250471e-05,
      "loss": 2.7582,
      "step": 363000
    },
    {
      "epoch": 19.58,
      "learning_rate": 2.3458097928436913e-05,
      "loss": 2.7596,
      "step": 364000
    },
    {
      "epoch": 19.63,
      "learning_rate": 2.0515536723163844e-05,
      "loss": 2.7617,
      "step": 365000
    },
    {
      "epoch": 19.69,
      "learning_rate": 1.7575918079096047e-05,
      "loss": 2.7599,
      "step": 366000
    },
    {
      "epoch": 19.74,
      "learning_rate": 1.4633356873822976e-05,
      "loss": 2.7614,
      "step": 367000
    },
    {
      "epoch": 19.79,
      "learning_rate": 1.169373822975518e-05,
      "loss": 2.7569,
      "step": 368000
    },
    {
      "epoch": 19.85,
      "learning_rate": 8.75117702448211e-06,
      "loss": 2.7589,
      "step": 369000
    },
    {
      "epoch": 19.9,
      "learning_rate": 5.80861581920904e-06,
      "loss": 2.7593,
      "step": 370000
    },
    {
      "epoch": 19.95,
      "learning_rate": 2.868997175141243e-06,
      "loss": 2.7631,
      "step": 371000
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.4119714215135951,
      "eval_loss": 3.3829431533813477,
      "eval_runtime": 154.4564,
      "eval_samples_per_second": 374.986,
      "eval_steps_per_second": 5.859,
      "step": 371840
    },
    {
      "epoch": 20.0,
      "step": 371840,
      "total_flos": 1.56667295384064e+18,
      "train_loss": 3.0290212811783546,
      "train_runtime": 81078.4615,
      "train_samples_per_second": 146.756,
      "train_steps_per_second": 4.586
    }
  ],
  "logging_steps": 1000,
  "max_steps": 371840,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 5000,
  "total_flos": 1.56667295384064e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}