vit-cifar10 / trainer_state.json
jamescalam's picture
first model
b363d25
{
"best_metric": 0.07883724570274353,
"best_model_checkpoint": "./cifar/checkpoint-6200",
"epoch": 4.0,
"global_step": 6252,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.01,
"learning_rate": 0.0001996801023672425,
"loss": 2.0965,
"step": 10
},
{
"epoch": 0.01,
"learning_rate": 0.00019936020473448497,
"loss": 1.4515,
"step": 20
},
{
"epoch": 0.02,
"learning_rate": 0.00019904030710172745,
"loss": 0.9389,
"step": 30
},
{
"epoch": 0.03,
"learning_rate": 0.00019872040946896996,
"loss": 0.689,
"step": 40
},
{
"epoch": 0.03,
"learning_rate": 0.0001984005118362124,
"loss": 0.4917,
"step": 50
},
{
"epoch": 0.04,
"learning_rate": 0.0001980806142034549,
"loss": 0.4551,
"step": 60
},
{
"epoch": 0.04,
"learning_rate": 0.0001977607165706974,
"loss": 0.4832,
"step": 70
},
{
"epoch": 0.05,
"learning_rate": 0.00019744081893793988,
"loss": 0.4072,
"step": 80
},
{
"epoch": 0.06,
"learning_rate": 0.00019712092130518236,
"loss": 0.3861,
"step": 90
},
{
"epoch": 0.06,
"learning_rate": 0.00019680102367242483,
"loss": 0.3212,
"step": 100
},
{
"epoch": 0.06,
"eval_accuracy": 0.8785,
"eval_loss": 0.47495898604393005,
"eval_runtime": 47.4814,
"eval_samples_per_second": 210.609,
"eval_steps_per_second": 13.163,
"step": 100
},
{
"epoch": 0.07,
"learning_rate": 0.00019648112603966731,
"loss": 0.3596,
"step": 110
},
{
"epoch": 0.08,
"learning_rate": 0.0001961612284069098,
"loss": 0.3605,
"step": 120
},
{
"epoch": 0.08,
"learning_rate": 0.00019584133077415227,
"loss": 0.3013,
"step": 130
},
{
"epoch": 0.09,
"learning_rate": 0.00019552143314139478,
"loss": 0.3512,
"step": 140
},
{
"epoch": 0.1,
"learning_rate": 0.00019520153550863723,
"loss": 0.3202,
"step": 150
},
{
"epoch": 0.1,
"learning_rate": 0.00019488163787587971,
"loss": 0.2308,
"step": 160
},
{
"epoch": 0.11,
"learning_rate": 0.00019456174024312222,
"loss": 0.4124,
"step": 170
},
{
"epoch": 0.12,
"learning_rate": 0.0001942418426103647,
"loss": 0.2991,
"step": 180
},
{
"epoch": 0.12,
"learning_rate": 0.00019392194497760718,
"loss": 0.2891,
"step": 190
},
{
"epoch": 0.13,
"learning_rate": 0.00019360204734484966,
"loss": 0.3145,
"step": 200
},
{
"epoch": 0.13,
"eval_accuracy": 0.9347,
"eval_loss": 0.2536068856716156,
"eval_runtime": 48.1227,
"eval_samples_per_second": 207.802,
"eval_steps_per_second": 12.988,
"step": 200
},
{
"epoch": 0.13,
"learning_rate": 0.00019328214971209214,
"loss": 0.1861,
"step": 210
},
{
"epoch": 0.14,
"learning_rate": 0.00019296225207933462,
"loss": 0.3056,
"step": 220
},
{
"epoch": 0.15,
"learning_rate": 0.0001926423544465771,
"loss": 0.3428,
"step": 230
},
{
"epoch": 0.15,
"learning_rate": 0.0001923224568138196,
"loss": 0.1819,
"step": 240
},
{
"epoch": 0.16,
"learning_rate": 0.00019200255918106206,
"loss": 0.2286,
"step": 250
},
{
"epoch": 0.17,
"learning_rate": 0.00019168266154830454,
"loss": 0.1687,
"step": 260
},
{
"epoch": 0.17,
"learning_rate": 0.00019136276391554705,
"loss": 0.2,
"step": 270
},
{
"epoch": 0.18,
"learning_rate": 0.0001910428662827895,
"loss": 0.2408,
"step": 280
},
{
"epoch": 0.19,
"learning_rate": 0.000190722968650032,
"loss": 0.2047,
"step": 290
},
{
"epoch": 0.19,
"learning_rate": 0.00019040307101727449,
"loss": 0.3278,
"step": 300
},
{
"epoch": 0.19,
"eval_accuracy": 0.9314,
"eval_loss": 0.2535996735095978,
"eval_runtime": 47.8164,
"eval_samples_per_second": 209.133,
"eval_steps_per_second": 13.071,
"step": 300
},
{
"epoch": 0.2,
"learning_rate": 0.00019008317338451697,
"loss": 0.2192,
"step": 310
},
{
"epoch": 0.2,
"learning_rate": 0.00018976327575175944,
"loss": 0.244,
"step": 320
},
{
"epoch": 0.21,
"learning_rate": 0.00018944337811900192,
"loss": 0.2785,
"step": 330
},
{
"epoch": 0.22,
"learning_rate": 0.00018912348048624443,
"loss": 0.1992,
"step": 340
},
{
"epoch": 0.22,
"learning_rate": 0.00018880358285348688,
"loss": 0.2039,
"step": 350
},
{
"epoch": 0.23,
"learning_rate": 0.00018848368522072936,
"loss": 0.2919,
"step": 360
},
{
"epoch": 0.24,
"learning_rate": 0.00018816378758797187,
"loss": 0.2075,
"step": 370
},
{
"epoch": 0.24,
"learning_rate": 0.00018784388995521432,
"loss": 0.2507,
"step": 380
},
{
"epoch": 0.25,
"learning_rate": 0.00018752399232245683,
"loss": 0.1471,
"step": 390
},
{
"epoch": 0.26,
"learning_rate": 0.0001872040946896993,
"loss": 0.2429,
"step": 400
},
{
"epoch": 0.26,
"eval_accuracy": 0.9265,
"eval_loss": 0.2561478912830353,
"eval_runtime": 48.1994,
"eval_samples_per_second": 207.472,
"eval_steps_per_second": 12.967,
"step": 400
},
{
"epoch": 0.26,
"learning_rate": 0.0001868841970569418,
"loss": 0.3039,
"step": 410
},
{
"epoch": 0.27,
"learning_rate": 0.00018656429942418427,
"loss": 0.3438,
"step": 420
},
{
"epoch": 0.28,
"learning_rate": 0.00018624440179142675,
"loss": 0.208,
"step": 430
},
{
"epoch": 0.28,
"learning_rate": 0.00018592450415866926,
"loss": 0.1909,
"step": 440
},
{
"epoch": 0.29,
"learning_rate": 0.0001856046065259117,
"loss": 0.2136,
"step": 450
},
{
"epoch": 0.29,
"learning_rate": 0.0001852847088931542,
"loss": 0.2753,
"step": 460
},
{
"epoch": 0.3,
"learning_rate": 0.0001849648112603967,
"loss": 0.2285,
"step": 470
},
{
"epoch": 0.31,
"learning_rate": 0.00018464491362763915,
"loss": 0.2331,
"step": 480
},
{
"epoch": 0.31,
"learning_rate": 0.00018432501599488166,
"loss": 0.2163,
"step": 490
},
{
"epoch": 0.32,
"learning_rate": 0.00018400511836212414,
"loss": 0.2622,
"step": 500
},
{
"epoch": 0.32,
"eval_accuracy": 0.9447,
"eval_loss": 0.19128669798374176,
"eval_runtime": 48.5832,
"eval_samples_per_second": 205.832,
"eval_steps_per_second": 12.865,
"step": 500
},
{
"epoch": 0.33,
"learning_rate": 0.00018368522072936662,
"loss": 0.1883,
"step": 510
},
{
"epoch": 0.33,
"learning_rate": 0.0001833653230966091,
"loss": 0.2193,
"step": 520
},
{
"epoch": 0.34,
"learning_rate": 0.00018304542546385158,
"loss": 0.1819,
"step": 530
},
{
"epoch": 0.35,
"learning_rate": 0.00018272552783109406,
"loss": 0.2348,
"step": 540
},
{
"epoch": 0.35,
"learning_rate": 0.00018240563019833653,
"loss": 0.1955,
"step": 550
},
{
"epoch": 0.36,
"learning_rate": 0.00018208573256557901,
"loss": 0.1497,
"step": 560
},
{
"epoch": 0.36,
"learning_rate": 0.00018176583493282152,
"loss": 0.1614,
"step": 570
},
{
"epoch": 0.37,
"learning_rate": 0.00018144593730006397,
"loss": 0.1747,
"step": 580
},
{
"epoch": 0.38,
"learning_rate": 0.00018112603966730648,
"loss": 0.1957,
"step": 590
},
{
"epoch": 0.38,
"learning_rate": 0.00018080614203454896,
"loss": 0.1714,
"step": 600
},
{
"epoch": 0.38,
"eval_accuracy": 0.939,
"eval_loss": 0.21139121055603027,
"eval_runtime": 48.6064,
"eval_samples_per_second": 205.734,
"eval_steps_per_second": 12.858,
"step": 600
},
{
"epoch": 0.39,
"learning_rate": 0.00018048624440179141,
"loss": 0.2064,
"step": 610
},
{
"epoch": 0.4,
"learning_rate": 0.00018016634676903392,
"loss": 0.2239,
"step": 620
},
{
"epoch": 0.4,
"learning_rate": 0.0001798464491362764,
"loss": 0.1753,
"step": 630
},
{
"epoch": 0.41,
"learning_rate": 0.00017952655150351888,
"loss": 0.2009,
"step": 640
},
{
"epoch": 0.42,
"learning_rate": 0.00017920665387076136,
"loss": 0.1754,
"step": 650
},
{
"epoch": 0.42,
"learning_rate": 0.00017888675623800384,
"loss": 0.1365,
"step": 660
},
{
"epoch": 0.43,
"learning_rate": 0.00017856685860524635,
"loss": 0.142,
"step": 670
},
{
"epoch": 0.44,
"learning_rate": 0.0001782469609724888,
"loss": 0.2181,
"step": 680
},
{
"epoch": 0.44,
"learning_rate": 0.0001779270633397313,
"loss": 0.1755,
"step": 690
},
{
"epoch": 0.45,
"learning_rate": 0.00017760716570697379,
"loss": 0.1966,
"step": 700
},
{
"epoch": 0.45,
"eval_accuracy": 0.9348,
"eval_loss": 0.21518036723136902,
"eval_runtime": 48.7309,
"eval_samples_per_second": 205.209,
"eval_steps_per_second": 12.826,
"step": 700
},
{
"epoch": 0.45,
"learning_rate": 0.00017728726807421624,
"loss": 0.1925,
"step": 710
},
{
"epoch": 0.46,
"learning_rate": 0.00017696737044145875,
"loss": 0.2256,
"step": 720
},
{
"epoch": 0.47,
"learning_rate": 0.00017664747280870123,
"loss": 0.1714,
"step": 730
},
{
"epoch": 0.47,
"learning_rate": 0.0001763275751759437,
"loss": 0.2855,
"step": 740
},
{
"epoch": 0.48,
"learning_rate": 0.00017600767754318619,
"loss": 0.2259,
"step": 750
},
{
"epoch": 0.49,
"learning_rate": 0.00017568777991042867,
"loss": 0.1718,
"step": 760
},
{
"epoch": 0.49,
"learning_rate": 0.00017536788227767115,
"loss": 0.1441,
"step": 770
},
{
"epoch": 0.5,
"learning_rate": 0.00017504798464491362,
"loss": 0.3405,
"step": 780
},
{
"epoch": 0.51,
"learning_rate": 0.00017472808701215613,
"loss": 0.1524,
"step": 790
},
{
"epoch": 0.51,
"learning_rate": 0.0001744081893793986,
"loss": 0.1395,
"step": 800
},
{
"epoch": 0.51,
"eval_accuracy": 0.9453,
"eval_loss": 0.18169045448303223,
"eval_runtime": 47.7658,
"eval_samples_per_second": 209.355,
"eval_steps_per_second": 13.085,
"step": 800
},
{
"epoch": 0.52,
"learning_rate": 0.00017408829174664106,
"loss": 0.134,
"step": 810
},
{
"epoch": 0.52,
"learning_rate": 0.00017376839411388357,
"loss": 0.1626,
"step": 820
},
{
"epoch": 0.53,
"learning_rate": 0.00017344849648112605,
"loss": 0.2228,
"step": 830
},
{
"epoch": 0.54,
"learning_rate": 0.00017312859884836853,
"loss": 0.204,
"step": 840
},
{
"epoch": 0.54,
"learning_rate": 0.000172808701215611,
"loss": 0.0642,
"step": 850
},
{
"epoch": 0.55,
"learning_rate": 0.0001724888035828535,
"loss": 0.215,
"step": 860
},
{
"epoch": 0.56,
"learning_rate": 0.00017216890595009597,
"loss": 0.1824,
"step": 870
},
{
"epoch": 0.56,
"learning_rate": 0.00017184900831733845,
"loss": 0.1739,
"step": 880
},
{
"epoch": 0.57,
"learning_rate": 0.00017152911068458096,
"loss": 0.1884,
"step": 890
},
{
"epoch": 0.58,
"learning_rate": 0.00017120921305182344,
"loss": 0.1323,
"step": 900
},
{
"epoch": 0.58,
"eval_accuracy": 0.9324,
"eval_loss": 0.2200065404176712,
"eval_runtime": 47.5971,
"eval_samples_per_second": 210.097,
"eval_steps_per_second": 13.131,
"step": 900
},
{
"epoch": 0.58,
"learning_rate": 0.0001708893154190659,
"loss": 0.1711,
"step": 910
},
{
"epoch": 0.59,
"learning_rate": 0.0001705694177863084,
"loss": 0.1293,
"step": 920
},
{
"epoch": 0.6,
"learning_rate": 0.00017024952015355088,
"loss": 0.2275,
"step": 930
},
{
"epoch": 0.6,
"learning_rate": 0.00016992962252079336,
"loss": 0.1531,
"step": 940
},
{
"epoch": 0.61,
"learning_rate": 0.00016960972488803584,
"loss": 0.1187,
"step": 950
},
{
"epoch": 0.61,
"learning_rate": 0.00016928982725527832,
"loss": 0.158,
"step": 960
},
{
"epoch": 0.62,
"learning_rate": 0.0001689699296225208,
"loss": 0.1857,
"step": 970
},
{
"epoch": 0.63,
"learning_rate": 0.00016865003198976328,
"loss": 0.142,
"step": 980
},
{
"epoch": 0.63,
"learning_rate": 0.00016833013435700578,
"loss": 0.2363,
"step": 990
},
{
"epoch": 0.64,
"learning_rate": 0.00016801023672424826,
"loss": 0.1707,
"step": 1000
},
{
"epoch": 0.64,
"eval_accuracy": 0.945,
"eval_loss": 0.18686933815479279,
"eval_runtime": 47.6282,
"eval_samples_per_second": 209.96,
"eval_steps_per_second": 13.122,
"step": 1000
},
{
"epoch": 0.65,
"learning_rate": 0.00016769033909149071,
"loss": 0.2039,
"step": 1010
},
{
"epoch": 0.65,
"learning_rate": 0.00016737044145873322,
"loss": 0.146,
"step": 1020
},
{
"epoch": 0.66,
"learning_rate": 0.0001670505438259757,
"loss": 0.1763,
"step": 1030
},
{
"epoch": 0.67,
"learning_rate": 0.00016673064619321818,
"loss": 0.1645,
"step": 1040
},
{
"epoch": 0.67,
"learning_rate": 0.00016641074856046066,
"loss": 0.1368,
"step": 1050
},
{
"epoch": 0.68,
"learning_rate": 0.00016609085092770314,
"loss": 0.1442,
"step": 1060
},
{
"epoch": 0.68,
"learning_rate": 0.00016577095329494562,
"loss": 0.1986,
"step": 1070
},
{
"epoch": 0.69,
"learning_rate": 0.0001654510556621881,
"loss": 0.1744,
"step": 1080
},
{
"epoch": 0.7,
"learning_rate": 0.0001651311580294306,
"loss": 0.1307,
"step": 1090
},
{
"epoch": 0.7,
"learning_rate": 0.00016481126039667306,
"loss": 0.1799,
"step": 1100
},
{
"epoch": 0.7,
"eval_accuracy": 0.9447,
"eval_loss": 0.18500135838985443,
"eval_runtime": 47.9607,
"eval_samples_per_second": 208.504,
"eval_steps_per_second": 13.032,
"step": 1100
},
{
"epoch": 0.71,
"learning_rate": 0.00016449136276391554,
"loss": 0.1933,
"step": 1110
},
{
"epoch": 0.72,
"learning_rate": 0.00016417146513115805,
"loss": 0.1567,
"step": 1120
},
{
"epoch": 0.72,
"learning_rate": 0.00016385156749840053,
"loss": 0.2155,
"step": 1130
},
{
"epoch": 0.73,
"learning_rate": 0.000163531669865643,
"loss": 0.2179,
"step": 1140
},
{
"epoch": 0.74,
"learning_rate": 0.0001632117722328855,
"loss": 0.2175,
"step": 1150
},
{
"epoch": 0.74,
"learning_rate": 0.00016289187460012797,
"loss": 0.1873,
"step": 1160
},
{
"epoch": 0.75,
"learning_rate": 0.00016257197696737045,
"loss": 0.1778,
"step": 1170
},
{
"epoch": 0.75,
"learning_rate": 0.00016225207933461293,
"loss": 0.2057,
"step": 1180
},
{
"epoch": 0.76,
"learning_rate": 0.00016193218170185543,
"loss": 0.1469,
"step": 1190
},
{
"epoch": 0.77,
"learning_rate": 0.00016161228406909789,
"loss": 0.1104,
"step": 1200
},
{
"epoch": 0.77,
"eval_accuracy": 0.9376,
"eval_loss": 0.20156049728393555,
"eval_runtime": 48.4092,
"eval_samples_per_second": 206.572,
"eval_steps_per_second": 12.911,
"step": 1200
},
{
"epoch": 0.77,
"learning_rate": 0.00016129238643634037,
"loss": 0.22,
"step": 1210
},
{
"epoch": 0.78,
"learning_rate": 0.00016097248880358287,
"loss": 0.1224,
"step": 1220
},
{
"epoch": 0.79,
"learning_rate": 0.00016065259117082535,
"loss": 0.1591,
"step": 1230
},
{
"epoch": 0.79,
"learning_rate": 0.00016033269353806783,
"loss": 0.1492,
"step": 1240
},
{
"epoch": 0.8,
"learning_rate": 0.0001600127959053103,
"loss": 0.1605,
"step": 1250
},
{
"epoch": 0.81,
"learning_rate": 0.0001596928982725528,
"loss": 0.1457,
"step": 1260
},
{
"epoch": 0.81,
"learning_rate": 0.00015937300063979527,
"loss": 0.0612,
"step": 1270
},
{
"epoch": 0.82,
"learning_rate": 0.00015905310300703775,
"loss": 0.082,
"step": 1280
},
{
"epoch": 0.83,
"learning_rate": 0.00015873320537428026,
"loss": 0.0897,
"step": 1290
},
{
"epoch": 0.83,
"learning_rate": 0.0001584133077415227,
"loss": 0.1298,
"step": 1300
},
{
"epoch": 0.83,
"eval_accuracy": 0.9577,
"eval_loss": 0.13800279796123505,
"eval_runtime": 62.6571,
"eval_samples_per_second": 159.599,
"eval_steps_per_second": 9.975,
"step": 1300
},
{
"epoch": 0.84,
"learning_rate": 0.0001580934101087652,
"loss": 0.1759,
"step": 1310
},
{
"epoch": 0.84,
"learning_rate": 0.0001577735124760077,
"loss": 0.1883,
"step": 1320
},
{
"epoch": 0.85,
"learning_rate": 0.00015745361484325015,
"loss": 0.1002,
"step": 1330
},
{
"epoch": 0.86,
"learning_rate": 0.00015713371721049266,
"loss": 0.1287,
"step": 1340
},
{
"epoch": 0.86,
"learning_rate": 0.00015681381957773514,
"loss": 0.1834,
"step": 1350
},
{
"epoch": 0.87,
"learning_rate": 0.00015649392194497762,
"loss": 0.1503,
"step": 1360
},
{
"epoch": 0.88,
"learning_rate": 0.0001561740243122201,
"loss": 0.2471,
"step": 1370
},
{
"epoch": 0.88,
"learning_rate": 0.00015585412667946258,
"loss": 0.164,
"step": 1380
},
{
"epoch": 0.89,
"learning_rate": 0.00015553422904670508,
"loss": 0.1632,
"step": 1390
},
{
"epoch": 0.9,
"learning_rate": 0.00015521433141394754,
"loss": 0.1124,
"step": 1400
},
{
"epoch": 0.9,
"eval_accuracy": 0.957,
"eval_loss": 0.14739342033863068,
"eval_runtime": 60.3393,
"eval_samples_per_second": 165.729,
"eval_steps_per_second": 10.358,
"step": 1400
},
{
"epoch": 0.9,
"learning_rate": 0.00015489443378119002,
"loss": 0.1396,
"step": 1410
},
{
"epoch": 0.91,
"learning_rate": 0.00015457453614843252,
"loss": 0.1766,
"step": 1420
},
{
"epoch": 0.91,
"learning_rate": 0.00015425463851567498,
"loss": 0.1866,
"step": 1430
},
{
"epoch": 0.92,
"learning_rate": 0.00015393474088291748,
"loss": 0.1443,
"step": 1440
},
{
"epoch": 0.93,
"learning_rate": 0.00015361484325015996,
"loss": 0.0976,
"step": 1450
},
{
"epoch": 0.93,
"learning_rate": 0.00015329494561740244,
"loss": 0.1618,
"step": 1460
},
{
"epoch": 0.94,
"learning_rate": 0.00015297504798464492,
"loss": 0.095,
"step": 1470
},
{
"epoch": 0.95,
"learning_rate": 0.0001526551503518874,
"loss": 0.1946,
"step": 1480
},
{
"epoch": 0.95,
"learning_rate": 0.00015233525271912988,
"loss": 0.1362,
"step": 1490
},
{
"epoch": 0.96,
"learning_rate": 0.00015201535508637236,
"loss": 0.2286,
"step": 1500
},
{
"epoch": 0.96,
"eval_accuracy": 0.9426,
"eval_loss": 0.18567727506160736,
"eval_runtime": 60.124,
"eval_samples_per_second": 166.323,
"eval_steps_per_second": 10.395,
"step": 1500
},
{
"epoch": 0.97,
"learning_rate": 0.00015169545745361484,
"loss": 0.168,
"step": 1510
},
{
"epoch": 0.97,
"learning_rate": 0.00015137555982085735,
"loss": 0.085,
"step": 1520
},
{
"epoch": 0.98,
"learning_rate": 0.0001510556621880998,
"loss": 0.1409,
"step": 1530
},
{
"epoch": 0.99,
"learning_rate": 0.0001507357645553423,
"loss": 0.2012,
"step": 1540
},
{
"epoch": 0.99,
"learning_rate": 0.0001504158669225848,
"loss": 0.1522,
"step": 1550
},
{
"epoch": 1.0,
"learning_rate": 0.00015009596928982727,
"loss": 0.0599,
"step": 1560
},
{
"epoch": 1.0,
"learning_rate": 0.00014977607165706975,
"loss": 0.0955,
"step": 1570
},
{
"epoch": 1.01,
"learning_rate": 0.00014945617402431223,
"loss": 0.0773,
"step": 1580
},
{
"epoch": 1.02,
"learning_rate": 0.0001491362763915547,
"loss": 0.0747,
"step": 1590
},
{
"epoch": 1.02,
"learning_rate": 0.0001488163787587972,
"loss": 0.0489,
"step": 1600
},
{
"epoch": 1.02,
"eval_accuracy": 0.9683,
"eval_loss": 0.10664957016706467,
"eval_runtime": 53.6606,
"eval_samples_per_second": 186.356,
"eval_steps_per_second": 11.647,
"step": 1600
},
{
"epoch": 1.03,
"learning_rate": 0.00014849648112603967,
"loss": 0.0523,
"step": 1610
},
{
"epoch": 1.04,
"learning_rate": 0.00014817658349328217,
"loss": 0.0387,
"step": 1620
},
{
"epoch": 1.04,
"learning_rate": 0.00014785668586052463,
"loss": 0.0738,
"step": 1630
},
{
"epoch": 1.05,
"learning_rate": 0.00014753678822776713,
"loss": 0.0868,
"step": 1640
},
{
"epoch": 1.06,
"learning_rate": 0.0001472168905950096,
"loss": 0.0614,
"step": 1650
},
{
"epoch": 1.06,
"learning_rate": 0.00014689699296225207,
"loss": 0.0782,
"step": 1660
},
{
"epoch": 1.07,
"learning_rate": 0.00014657709532949457,
"loss": 0.1417,
"step": 1670
},
{
"epoch": 1.07,
"learning_rate": 0.00014625719769673705,
"loss": 0.07,
"step": 1680
},
{
"epoch": 1.08,
"learning_rate": 0.00014593730006397953,
"loss": 0.0856,
"step": 1690
},
{
"epoch": 1.09,
"learning_rate": 0.000145617402431222,
"loss": 0.0829,
"step": 1700
},
{
"epoch": 1.09,
"eval_accuracy": 0.9584,
"eval_loss": 0.14192388951778412,
"eval_runtime": 47.2202,
"eval_samples_per_second": 211.774,
"eval_steps_per_second": 13.236,
"step": 1700
},
{
"epoch": 1.09,
"learning_rate": 0.0001452975047984645,
"loss": 0.0773,
"step": 1710
},
{
"epoch": 1.1,
"learning_rate": 0.000144977607165707,
"loss": 0.1056,
"step": 1720
},
{
"epoch": 1.11,
"learning_rate": 0.00014465770953294945,
"loss": 0.1197,
"step": 1730
},
{
"epoch": 1.11,
"learning_rate": 0.00014433781190019196,
"loss": 0.1031,
"step": 1740
},
{
"epoch": 1.12,
"learning_rate": 0.00014401791426743444,
"loss": 0.0558,
"step": 1750
},
{
"epoch": 1.13,
"learning_rate": 0.0001436980166346769,
"loss": 0.0797,
"step": 1760
},
{
"epoch": 1.13,
"learning_rate": 0.0001433781190019194,
"loss": 0.12,
"step": 1770
},
{
"epoch": 1.14,
"learning_rate": 0.00014305822136916188,
"loss": 0.1093,
"step": 1780
},
{
"epoch": 1.15,
"learning_rate": 0.00014273832373640436,
"loss": 0.0932,
"step": 1790
},
{
"epoch": 1.15,
"learning_rate": 0.00014241842610364684,
"loss": 0.1269,
"step": 1800
},
{
"epoch": 1.15,
"eval_accuracy": 0.9597,
"eval_loss": 0.13903552293777466,
"eval_runtime": 47.4785,
"eval_samples_per_second": 210.622,
"eval_steps_per_second": 13.164,
"step": 1800
},
{
"epoch": 1.16,
"learning_rate": 0.00014209852847088932,
"loss": 0.0779,
"step": 1810
},
{
"epoch": 1.16,
"learning_rate": 0.0001417786308381318,
"loss": 0.1292,
"step": 1820
},
{
"epoch": 1.17,
"learning_rate": 0.00014145873320537428,
"loss": 0.0975,
"step": 1830
},
{
"epoch": 1.18,
"learning_rate": 0.00014113883557261678,
"loss": 0.126,
"step": 1840
},
{
"epoch": 1.18,
"learning_rate": 0.00014081893793985926,
"loss": 0.0781,
"step": 1850
},
{
"epoch": 1.19,
"learning_rate": 0.00014049904030710172,
"loss": 0.0943,
"step": 1860
},
{
"epoch": 1.2,
"learning_rate": 0.00014017914267434422,
"loss": 0.0936,
"step": 1870
},
{
"epoch": 1.2,
"learning_rate": 0.0001398592450415867,
"loss": 0.086,
"step": 1880
},
{
"epoch": 1.21,
"learning_rate": 0.00013953934740882918,
"loss": 0.1549,
"step": 1890
},
{
"epoch": 1.22,
"learning_rate": 0.00013921944977607166,
"loss": 0.0635,
"step": 1900
},
{
"epoch": 1.22,
"eval_accuracy": 0.961,
"eval_loss": 0.1290891319513321,
"eval_runtime": 47.9741,
"eval_samples_per_second": 208.446,
"eval_steps_per_second": 13.028,
"step": 1900
},
{
"epoch": 1.22,
"learning_rate": 0.00013889955214331414,
"loss": 0.0465,
"step": 1910
},
{
"epoch": 1.23,
"learning_rate": 0.00013857965451055662,
"loss": 0.0874,
"step": 1920
},
{
"epoch": 1.23,
"learning_rate": 0.0001382597568777991,
"loss": 0.0738,
"step": 1930
},
{
"epoch": 1.24,
"learning_rate": 0.0001379398592450416,
"loss": 0.0777,
"step": 1940
},
{
"epoch": 1.25,
"learning_rate": 0.0001376199616122841,
"loss": 0.0866,
"step": 1950
},
{
"epoch": 1.25,
"learning_rate": 0.00013730006397952654,
"loss": 0.0895,
"step": 1960
},
{
"epoch": 1.26,
"learning_rate": 0.00013698016634676905,
"loss": 0.0758,
"step": 1970
},
{
"epoch": 1.27,
"learning_rate": 0.00013666026871401153,
"loss": 0.0421,
"step": 1980
},
{
"epoch": 1.27,
"learning_rate": 0.000136340371081254,
"loss": 0.0586,
"step": 1990
},
{
"epoch": 1.28,
"learning_rate": 0.0001360204734484965,
"loss": 0.1098,
"step": 2000
},
{
"epoch": 1.28,
"eval_accuracy": 0.9582,
"eval_loss": 0.1469649374485016,
"eval_runtime": 48.3383,
"eval_samples_per_second": 206.875,
"eval_steps_per_second": 12.93,
"step": 2000
},
{
"epoch": 1.29,
"learning_rate": 0.00013570057581573897,
"loss": 0.0549,
"step": 2010
},
{
"epoch": 1.29,
"learning_rate": 0.00013538067818298145,
"loss": 0.0379,
"step": 2020
},
{
"epoch": 1.3,
"learning_rate": 0.00013506078055022393,
"loss": 0.1064,
"step": 2030
},
{
"epoch": 1.31,
"learning_rate": 0.00013474088291746643,
"loss": 0.0964,
"step": 2040
},
{
"epoch": 1.31,
"learning_rate": 0.0001344209852847089,
"loss": 0.0705,
"step": 2050
},
{
"epoch": 1.32,
"learning_rate": 0.00013410108765195137,
"loss": 0.0691,
"step": 2060
},
{
"epoch": 1.32,
"learning_rate": 0.00013378119001919387,
"loss": 0.0611,
"step": 2070
},
{
"epoch": 1.33,
"learning_rate": 0.00013346129238643635,
"loss": 0.0718,
"step": 2080
},
{
"epoch": 1.34,
"learning_rate": 0.00013314139475367883,
"loss": 0.0535,
"step": 2090
},
{
"epoch": 1.34,
"learning_rate": 0.0001328214971209213,
"loss": 0.1065,
"step": 2100
},
{
"epoch": 1.34,
"eval_accuracy": 0.9598,
"eval_loss": 0.13903473317623138,
"eval_runtime": 47.7783,
"eval_samples_per_second": 209.3,
"eval_steps_per_second": 13.081,
"step": 2100
},
{
"epoch": 1.35,
"learning_rate": 0.0001325015994881638,
"loss": 0.0715,
"step": 2110
},
{
"epoch": 1.36,
"learning_rate": 0.00013218170185540627,
"loss": 0.0849,
"step": 2120
},
{
"epoch": 1.36,
"learning_rate": 0.00013186180422264875,
"loss": 0.0926,
"step": 2130
},
{
"epoch": 1.37,
"learning_rate": 0.00013154190658989126,
"loss": 0.1035,
"step": 2140
},
{
"epoch": 1.38,
"learning_rate": 0.0001312220089571337,
"loss": 0.0495,
"step": 2150
},
{
"epoch": 1.38,
"learning_rate": 0.0001309021113243762,
"loss": 0.0785,
"step": 2160
},
{
"epoch": 1.39,
"learning_rate": 0.0001305822136916187,
"loss": 0.053,
"step": 2170
},
{
"epoch": 1.39,
"learning_rate": 0.00013026231605886118,
"loss": 0.0576,
"step": 2180
},
{
"epoch": 1.4,
"learning_rate": 0.00012994241842610366,
"loss": 0.0728,
"step": 2190
},
{
"epoch": 1.41,
"learning_rate": 0.00012962252079334614,
"loss": 0.0753,
"step": 2200
},
{
"epoch": 1.41,
"eval_accuracy": 0.9651,
"eval_loss": 0.12330373376607895,
"eval_runtime": 49.6152,
"eval_samples_per_second": 201.551,
"eval_steps_per_second": 12.597,
"step": 2200
},
{
"epoch": 1.41,
"learning_rate": 0.00012930262316058862,
"loss": 0.0441,
"step": 2210
},
{
"epoch": 1.42,
"learning_rate": 0.0001289827255278311,
"loss": 0.0526,
"step": 2220
},
{
"epoch": 1.43,
"learning_rate": 0.00012866282789507358,
"loss": 0.082,
"step": 2230
},
{
"epoch": 1.43,
"learning_rate": 0.00012834293026231608,
"loss": 0.0734,
"step": 2240
},
{
"epoch": 1.44,
"learning_rate": 0.00012802303262955854,
"loss": 0.0465,
"step": 2250
},
{
"epoch": 1.45,
"learning_rate": 0.00012770313499680102,
"loss": 0.0841,
"step": 2260
},
{
"epoch": 1.45,
"learning_rate": 0.00012738323736404352,
"loss": 0.1249,
"step": 2270
},
{
"epoch": 1.46,
"learning_rate": 0.000127063339731286,
"loss": 0.0852,
"step": 2280
},
{
"epoch": 1.47,
"learning_rate": 0.00012674344209852848,
"loss": 0.1211,
"step": 2290
},
{
"epoch": 1.47,
"learning_rate": 0.00012642354446577096,
"loss": 0.0855,
"step": 2300
},
{
"epoch": 1.47,
"eval_accuracy": 0.958,
"eval_loss": 0.14163978397846222,
"eval_runtime": 48.1054,
"eval_samples_per_second": 207.877,
"eval_steps_per_second": 12.992,
"step": 2300
},
{
"epoch": 1.48,
"learning_rate": 0.00012610364683301344,
"loss": 0.0525,
"step": 2310
},
{
"epoch": 1.48,
"learning_rate": 0.00012578374920025592,
"loss": 0.1068,
"step": 2320
},
{
"epoch": 1.49,
"learning_rate": 0.0001254638515674984,
"loss": 0.0847,
"step": 2330
},
{
"epoch": 1.5,
"learning_rate": 0.0001251439539347409,
"loss": 0.0945,
"step": 2340
},
{
"epoch": 1.5,
"learning_rate": 0.00012482405630198336,
"loss": 0.108,
"step": 2350
},
{
"epoch": 1.51,
"learning_rate": 0.00012450415866922584,
"loss": 0.0529,
"step": 2360
},
{
"epoch": 1.52,
"learning_rate": 0.00012418426103646835,
"loss": 0.0679,
"step": 2370
},
{
"epoch": 1.52,
"learning_rate": 0.0001238643634037108,
"loss": 0.1067,
"step": 2380
},
{
"epoch": 1.53,
"learning_rate": 0.0001235444657709533,
"loss": 0.0752,
"step": 2390
},
{
"epoch": 1.54,
"learning_rate": 0.0001232245681381958,
"loss": 0.0723,
"step": 2400
},
{
"epoch": 1.54,
"eval_accuracy": 0.9606,
"eval_loss": 0.13063974678516388,
"eval_runtime": 47.4226,
"eval_samples_per_second": 210.87,
"eval_steps_per_second": 13.179,
"step": 2400
},
{
"epoch": 1.54,
"learning_rate": 0.00012290467050543827,
"loss": 0.0722,
"step": 2410
},
{
"epoch": 1.55,
"learning_rate": 0.00012258477287268075,
"loss": 0.1103,
"step": 2420
},
{
"epoch": 1.55,
"learning_rate": 0.00012226487523992323,
"loss": 0.0846,
"step": 2430
},
{
"epoch": 1.56,
"learning_rate": 0.00012194497760716572,
"loss": 0.1057,
"step": 2440
},
{
"epoch": 1.57,
"learning_rate": 0.0001216250799744082,
"loss": 0.0528,
"step": 2450
},
{
"epoch": 1.57,
"learning_rate": 0.00012130518234165067,
"loss": 0.0776,
"step": 2460
},
{
"epoch": 1.58,
"learning_rate": 0.00012098528470889316,
"loss": 0.0833,
"step": 2470
},
{
"epoch": 1.59,
"learning_rate": 0.00012066538707613564,
"loss": 0.0716,
"step": 2480
},
{
"epoch": 1.59,
"learning_rate": 0.00012034548944337813,
"loss": 0.0524,
"step": 2490
},
{
"epoch": 1.6,
"learning_rate": 0.00012002559181062061,
"loss": 0.0515,
"step": 2500
},
{
"epoch": 1.6,
"eval_accuracy": 0.9652,
"eval_loss": 0.11919395625591278,
"eval_runtime": 47.5032,
"eval_samples_per_second": 210.512,
"eval_steps_per_second": 13.157,
"step": 2500
},
{
"epoch": 1.61,
"learning_rate": 0.00011970569417786308,
"loss": 0.0542,
"step": 2510
},
{
"epoch": 1.61,
"learning_rate": 0.00011938579654510557,
"loss": 0.0625,
"step": 2520
},
{
"epoch": 1.62,
"learning_rate": 0.00011906589891234805,
"loss": 0.0598,
"step": 2530
},
{
"epoch": 1.63,
"learning_rate": 0.00011874600127959055,
"loss": 0.0668,
"step": 2540
},
{
"epoch": 1.63,
"learning_rate": 0.00011842610364683303,
"loss": 0.0539,
"step": 2550
},
{
"epoch": 1.64,
"learning_rate": 0.00011810620601407549,
"loss": 0.0635,
"step": 2560
},
{
"epoch": 1.64,
"learning_rate": 0.00011778630838131799,
"loss": 0.0501,
"step": 2570
},
{
"epoch": 1.65,
"learning_rate": 0.00011746641074856047,
"loss": 0.0573,
"step": 2580
},
{
"epoch": 1.66,
"learning_rate": 0.00011714651311580296,
"loss": 0.0435,
"step": 2590
},
{
"epoch": 1.66,
"learning_rate": 0.00011682661548304543,
"loss": 0.0997,
"step": 2600
},
{
"epoch": 1.66,
"eval_accuracy": 0.9689,
"eval_loss": 0.11294878274202347,
"eval_runtime": 48.242,
"eval_samples_per_second": 207.288,
"eval_steps_per_second": 12.956,
"step": 2600
},
{
"epoch": 1.67,
"learning_rate": 0.0001165067178502879,
"loss": 0.1147,
"step": 2610
},
{
"epoch": 1.68,
"learning_rate": 0.0001161868202175304,
"loss": 0.0868,
"step": 2620
},
{
"epoch": 1.68,
"learning_rate": 0.00011586692258477288,
"loss": 0.1187,
"step": 2630
},
{
"epoch": 1.69,
"learning_rate": 0.00011554702495201537,
"loss": 0.0814,
"step": 2640
},
{
"epoch": 1.7,
"learning_rate": 0.00011522712731925784,
"loss": 0.0916,
"step": 2650
},
{
"epoch": 1.7,
"learning_rate": 0.00011490722968650032,
"loss": 0.0995,
"step": 2660
},
{
"epoch": 1.71,
"learning_rate": 0.00011458733205374281,
"loss": 0.0913,
"step": 2670
},
{
"epoch": 1.71,
"learning_rate": 0.00011426743442098529,
"loss": 0.0493,
"step": 2680
},
{
"epoch": 1.72,
"learning_rate": 0.00011394753678822778,
"loss": 0.0965,
"step": 2690
},
{
"epoch": 1.73,
"learning_rate": 0.00011362763915547025,
"loss": 0.0958,
"step": 2700
},
{
"epoch": 1.73,
"eval_accuracy": 0.9622,
"eval_loss": 0.12688662111759186,
"eval_runtime": 47.1481,
"eval_samples_per_second": 212.097,
"eval_steps_per_second": 13.256,
"step": 2700
},
{
"epoch": 1.73,
"learning_rate": 0.00011330774152271273,
"loss": 0.0561,
"step": 2710
},
{
"epoch": 1.74,
"learning_rate": 0.00011298784388995522,
"loss": 0.0533,
"step": 2720
},
{
"epoch": 1.75,
"learning_rate": 0.0001126679462571977,
"loss": 0.0738,
"step": 2730
},
{
"epoch": 1.75,
"learning_rate": 0.0001123480486244402,
"loss": 0.0765,
"step": 2740
},
{
"epoch": 1.76,
"learning_rate": 0.00011202815099168266,
"loss": 0.087,
"step": 2750
},
{
"epoch": 1.77,
"learning_rate": 0.00011170825335892514,
"loss": 0.036,
"step": 2760
},
{
"epoch": 1.77,
"learning_rate": 0.00011138835572616764,
"loss": 0.0699,
"step": 2770
},
{
"epoch": 1.78,
"learning_rate": 0.00011106845809341012,
"loss": 0.0565,
"step": 2780
},
{
"epoch": 1.79,
"learning_rate": 0.00011074856046065261,
"loss": 0.0718,
"step": 2790
},
{
"epoch": 1.79,
"learning_rate": 0.00011042866282789508,
"loss": 0.0502,
"step": 2800
},
{
"epoch": 1.79,
"eval_accuracy": 0.9641,
"eval_loss": 0.12282514572143555,
"eval_runtime": 48.2449,
"eval_samples_per_second": 207.276,
"eval_steps_per_second": 12.955,
"step": 2800
},
{
"epoch": 1.8,
"learning_rate": 0.00011010876519513756,
"loss": 0.0535,
"step": 2810
},
{
"epoch": 1.8,
"learning_rate": 0.00010978886756238005,
"loss": 0.0484,
"step": 2820
},
{
"epoch": 1.81,
"learning_rate": 0.00010946896992962253,
"loss": 0.0597,
"step": 2830
},
{
"epoch": 1.82,
"learning_rate": 0.00010914907229686502,
"loss": 0.0604,
"step": 2840
},
{
"epoch": 1.82,
"learning_rate": 0.00010882917466410749,
"loss": 0.0644,
"step": 2850
},
{
"epoch": 1.83,
"learning_rate": 0.00010850927703134997,
"loss": 0.0727,
"step": 2860
},
{
"epoch": 1.84,
"learning_rate": 0.00010818937939859246,
"loss": 0.0565,
"step": 2870
},
{
"epoch": 1.84,
"learning_rate": 0.00010786948176583493,
"loss": 0.0717,
"step": 2880
},
{
"epoch": 1.85,
"learning_rate": 0.00010754958413307744,
"loss": 0.0892,
"step": 2890
},
{
"epoch": 1.86,
"learning_rate": 0.0001072296865003199,
"loss": 0.0379,
"step": 2900
},
{
"epoch": 1.86,
"eval_accuracy": 0.9639,
"eval_loss": 0.12743479013442993,
"eval_runtime": 47.9702,
"eval_samples_per_second": 208.463,
"eval_steps_per_second": 13.029,
"step": 2900
},
{
"epoch": 1.86,
"learning_rate": 0.00010690978886756238,
"loss": 0.065,
"step": 2910
},
{
"epoch": 1.87,
"learning_rate": 0.00010658989123480487,
"loss": 0.0668,
"step": 2920
},
{
"epoch": 1.87,
"learning_rate": 0.00010626999360204734,
"loss": 0.0243,
"step": 2930
},
{
"epoch": 1.88,
"learning_rate": 0.00010595009596928985,
"loss": 0.0314,
"step": 2940
},
{
"epoch": 1.89,
"learning_rate": 0.00010563019833653231,
"loss": 0.0629,
"step": 2950
},
{
"epoch": 1.89,
"learning_rate": 0.0001053103007037748,
"loss": 0.0629,
"step": 2960
},
{
"epoch": 1.9,
"learning_rate": 0.00010499040307101729,
"loss": 0.0671,
"step": 2970
},
{
"epoch": 1.91,
"learning_rate": 0.00010467050543825975,
"loss": 0.0863,
"step": 2980
},
{
"epoch": 1.91,
"learning_rate": 0.00010435060780550226,
"loss": 0.1094,
"step": 2990
},
{
"epoch": 1.92,
"learning_rate": 0.00010403071017274473,
"loss": 0.0814,
"step": 3000
},
{
"epoch": 1.92,
"eval_accuracy": 0.963,
"eval_loss": 0.12582720816135406,
"eval_runtime": 47.6406,
"eval_samples_per_second": 209.905,
"eval_steps_per_second": 13.119,
"step": 3000
},
{
"epoch": 1.93,
"learning_rate": 0.0001037108125399872,
"loss": 0.091,
"step": 3010
},
{
"epoch": 1.93,
"learning_rate": 0.0001033909149072297,
"loss": 0.0237,
"step": 3020
},
{
"epoch": 1.94,
"learning_rate": 0.00010307101727447217,
"loss": 0.0732,
"step": 3030
},
{
"epoch": 1.94,
"learning_rate": 0.00010275111964171466,
"loss": 0.0283,
"step": 3040
},
{
"epoch": 1.95,
"learning_rate": 0.00010243122200895714,
"loss": 0.0972,
"step": 3050
},
{
"epoch": 1.96,
"learning_rate": 0.00010211132437619962,
"loss": 0.0706,
"step": 3060
},
{
"epoch": 1.96,
"learning_rate": 0.00010179142674344211,
"loss": 0.029,
"step": 3070
},
{
"epoch": 1.97,
"learning_rate": 0.00010147152911068458,
"loss": 0.0587,
"step": 3080
},
{
"epoch": 1.98,
"learning_rate": 0.00010115163147792707,
"loss": 0.1159,
"step": 3090
},
{
"epoch": 1.98,
"learning_rate": 0.00010083173384516955,
"loss": 0.0892,
"step": 3100
},
{
"epoch": 1.98,
"eval_accuracy": 0.9652,
"eval_loss": 0.11734022200107574,
"eval_runtime": 47.9972,
"eval_samples_per_second": 208.346,
"eval_steps_per_second": 13.022,
"step": 3100
},
{
"epoch": 1.99,
"learning_rate": 0.00010051183621241203,
"loss": 0.0566,
"step": 3110
},
{
"epoch": 2.0,
"learning_rate": 0.00010019193857965453,
"loss": 0.0434,
"step": 3120
},
{
"epoch": 2.0,
"learning_rate": 9.987204094689699e-05,
"loss": 0.0197,
"step": 3130
},
{
"epoch": 2.01,
"learning_rate": 9.955214331413948e-05,
"loss": 0.0411,
"step": 3140
},
{
"epoch": 2.02,
"learning_rate": 9.923224568138196e-05,
"loss": 0.0278,
"step": 3150
},
{
"epoch": 2.02,
"learning_rate": 9.891234804862444e-05,
"loss": 0.0242,
"step": 3160
},
{
"epoch": 2.03,
"learning_rate": 9.859245041586694e-05,
"loss": 0.0324,
"step": 3170
},
{
"epoch": 2.03,
"learning_rate": 9.82725527831094e-05,
"loss": 0.0368,
"step": 3180
},
{
"epoch": 2.04,
"learning_rate": 9.79526551503519e-05,
"loss": 0.0174,
"step": 3190
},
{
"epoch": 2.05,
"learning_rate": 9.763275751759438e-05,
"loss": 0.0126,
"step": 3200
},
{
"epoch": 2.05,
"eval_accuracy": 0.9662,
"eval_loss": 0.11732649058103561,
"eval_runtime": 48.648,
"eval_samples_per_second": 205.558,
"eval_steps_per_second": 12.847,
"step": 3200
},
{
"epoch": 2.05,
"learning_rate": 9.731285988483686e-05,
"loss": 0.0398,
"step": 3210
},
{
"epoch": 2.06,
"learning_rate": 9.699296225207935e-05,
"loss": 0.0233,
"step": 3220
},
{
"epoch": 2.07,
"learning_rate": 9.667306461932182e-05,
"loss": 0.0243,
"step": 3230
},
{
"epoch": 2.07,
"learning_rate": 9.63531669865643e-05,
"loss": 0.0247,
"step": 3240
},
{
"epoch": 2.08,
"learning_rate": 9.603326935380679e-05,
"loss": 0.0153,
"step": 3250
},
{
"epoch": 2.09,
"learning_rate": 9.571337172104927e-05,
"loss": 0.0199,
"step": 3260
},
{
"epoch": 2.09,
"learning_rate": 9.539347408829176e-05,
"loss": 0.0257,
"step": 3270
},
{
"epoch": 2.1,
"learning_rate": 9.507357645553423e-05,
"loss": 0.0221,
"step": 3280
},
{
"epoch": 2.1,
"learning_rate": 9.475367882277671e-05,
"loss": 0.0118,
"step": 3290
},
{
"epoch": 2.11,
"learning_rate": 9.44337811900192e-05,
"loss": 0.04,
"step": 3300
},
{
"epoch": 2.11,
"eval_accuracy": 0.972,
"eval_loss": 0.10201091319322586,
"eval_runtime": 47.4962,
"eval_samples_per_second": 210.543,
"eval_steps_per_second": 13.159,
"step": 3300
},
{
"epoch": 2.12,
"learning_rate": 9.411388355726168e-05,
"loss": 0.0027,
"step": 3310
},
{
"epoch": 2.12,
"learning_rate": 9.379398592450416e-05,
"loss": 0.0182,
"step": 3320
},
{
"epoch": 2.13,
"learning_rate": 9.347408829174664e-05,
"loss": 0.011,
"step": 3330
},
{
"epoch": 2.14,
"learning_rate": 9.315419065898912e-05,
"loss": 0.0502,
"step": 3340
},
{
"epoch": 2.14,
"learning_rate": 9.283429302623161e-05,
"loss": 0.0529,
"step": 3350
},
{
"epoch": 2.15,
"learning_rate": 9.25143953934741e-05,
"loss": 0.0202,
"step": 3360
},
{
"epoch": 2.16,
"learning_rate": 9.219449776071657e-05,
"loss": 0.0138,
"step": 3370
},
{
"epoch": 2.16,
"learning_rate": 9.187460012795905e-05,
"loss": 0.0037,
"step": 3380
},
{
"epoch": 2.17,
"learning_rate": 9.155470249520153e-05,
"loss": 0.0033,
"step": 3390
},
{
"epoch": 2.18,
"learning_rate": 9.123480486244403e-05,
"loss": 0.0098,
"step": 3400
},
{
"epoch": 2.18,
"eval_accuracy": 0.9731,
"eval_loss": 0.10650923103094101,
"eval_runtime": 47.4681,
"eval_samples_per_second": 210.668,
"eval_steps_per_second": 13.167,
"step": 3400
},
{
"epoch": 2.18,
"learning_rate": 9.091490722968651e-05,
"loss": 0.0056,
"step": 3410
},
{
"epoch": 2.19,
"learning_rate": 9.059500959692899e-05,
"loss": 0.0289,
"step": 3420
},
{
"epoch": 2.19,
"learning_rate": 9.027511196417147e-05,
"loss": 0.0659,
"step": 3430
},
{
"epoch": 2.2,
"learning_rate": 8.995521433141395e-05,
"loss": 0.0388,
"step": 3440
},
{
"epoch": 2.21,
"learning_rate": 8.963531669865644e-05,
"loss": 0.0206,
"step": 3450
},
{
"epoch": 2.21,
"learning_rate": 8.931541906589892e-05,
"loss": 0.0202,
"step": 3460
},
{
"epoch": 2.22,
"learning_rate": 8.89955214331414e-05,
"loss": 0.0128,
"step": 3470
},
{
"epoch": 2.23,
"learning_rate": 8.867562380038388e-05,
"loss": 0.0367,
"step": 3480
},
{
"epoch": 2.23,
"learning_rate": 8.835572616762636e-05,
"loss": 0.0132,
"step": 3490
},
{
"epoch": 2.24,
"learning_rate": 8.803582853486885e-05,
"loss": 0.0159,
"step": 3500
},
{
"epoch": 2.24,
"eval_accuracy": 0.9741,
"eval_loss": 0.1040922999382019,
"eval_runtime": 50.3231,
"eval_samples_per_second": 198.716,
"eval_steps_per_second": 12.42,
"step": 3500
},
{
"epoch": 2.25,
"learning_rate": 8.771593090211133e-05,
"loss": 0.0254,
"step": 3510
},
{
"epoch": 2.25,
"learning_rate": 8.739603326935381e-05,
"loss": 0.0333,
"step": 3520
},
{
"epoch": 2.26,
"learning_rate": 8.707613563659629e-05,
"loss": 0.0099,
"step": 3530
},
{
"epoch": 2.26,
"learning_rate": 8.675623800383877e-05,
"loss": 0.0211,
"step": 3540
},
{
"epoch": 2.27,
"learning_rate": 8.643634037108127e-05,
"loss": 0.0201,
"step": 3550
},
{
"epoch": 2.28,
"learning_rate": 8.611644273832375e-05,
"loss": 0.0223,
"step": 3560
},
{
"epoch": 2.28,
"learning_rate": 8.579654510556623e-05,
"loss": 0.0263,
"step": 3570
},
{
"epoch": 2.29,
"learning_rate": 8.54766474728087e-05,
"loss": 0.0067,
"step": 3580
},
{
"epoch": 2.3,
"learning_rate": 8.515674984005118e-05,
"loss": 0.008,
"step": 3590
},
{
"epoch": 2.3,
"learning_rate": 8.483685220729366e-05,
"loss": 0.0247,
"step": 3600
},
{
"epoch": 2.3,
"eval_accuracy": 0.9716,
"eval_loss": 0.10976359993219376,
"eval_runtime": 47.3179,
"eval_samples_per_second": 211.337,
"eval_steps_per_second": 13.209,
"step": 3600
},
{
"epoch": 2.31,
"learning_rate": 8.451695457453616e-05,
"loss": 0.0691,
"step": 3610
},
{
"epoch": 2.32,
"learning_rate": 8.419705694177864e-05,
"loss": 0.0153,
"step": 3620
},
{
"epoch": 2.32,
"learning_rate": 8.387715930902112e-05,
"loss": 0.0115,
"step": 3630
},
{
"epoch": 2.33,
"learning_rate": 8.35572616762636e-05,
"loss": 0.0267,
"step": 3640
},
{
"epoch": 2.34,
"learning_rate": 8.323736404350608e-05,
"loss": 0.0102,
"step": 3650
},
{
"epoch": 2.34,
"learning_rate": 8.291746641074857e-05,
"loss": 0.0121,
"step": 3660
},
{
"epoch": 2.35,
"learning_rate": 8.259756877799105e-05,
"loss": 0.0255,
"step": 3670
},
{
"epoch": 2.35,
"learning_rate": 8.227767114523353e-05,
"loss": 0.0246,
"step": 3680
},
{
"epoch": 2.36,
"learning_rate": 8.195777351247601e-05,
"loss": 0.0574,
"step": 3690
},
{
"epoch": 2.37,
"learning_rate": 8.163787587971849e-05,
"loss": 0.0457,
"step": 3700
},
{
"epoch": 2.37,
"eval_accuracy": 0.972,
"eval_loss": 0.10533589124679565,
"eval_runtime": 48.4465,
"eval_samples_per_second": 206.413,
"eval_steps_per_second": 12.901,
"step": 3700
},
{
"epoch": 2.37,
"learning_rate": 8.131797824696098e-05,
"loss": 0.0185,
"step": 3710
},
{
"epoch": 2.38,
"learning_rate": 8.099808061420346e-05,
"loss": 0.0066,
"step": 3720
},
{
"epoch": 2.39,
"learning_rate": 8.067818298144594e-05,
"loss": 0.0036,
"step": 3730
},
{
"epoch": 2.39,
"learning_rate": 8.035828534868842e-05,
"loss": 0.012,
"step": 3740
},
{
"epoch": 2.4,
"learning_rate": 8.00383877159309e-05,
"loss": 0.009,
"step": 3750
},
{
"epoch": 2.41,
"learning_rate": 7.97184900831734e-05,
"loss": 0.0113,
"step": 3760
},
{
"epoch": 2.41,
"learning_rate": 7.939859245041588e-05,
"loss": 0.0248,
"step": 3770
},
{
"epoch": 2.42,
"learning_rate": 7.907869481765836e-05,
"loss": 0.0051,
"step": 3780
},
{
"epoch": 2.42,
"learning_rate": 7.875879718490084e-05,
"loss": 0.0392,
"step": 3790
},
{
"epoch": 2.43,
"learning_rate": 7.843889955214332e-05,
"loss": 0.0621,
"step": 3800
},
{
"epoch": 2.43,
"eval_accuracy": 0.9761,
"eval_loss": 0.09739825129508972,
"eval_runtime": 47.7784,
"eval_samples_per_second": 209.3,
"eval_steps_per_second": 13.081,
"step": 3800
},
{
"epoch": 2.44,
"learning_rate": 7.811900191938581e-05,
"loss": 0.0091,
"step": 3810
},
{
"epoch": 2.44,
"learning_rate": 7.779910428662829e-05,
"loss": 0.0228,
"step": 3820
},
{
"epoch": 2.45,
"learning_rate": 7.747920665387077e-05,
"loss": 0.0359,
"step": 3830
},
{
"epoch": 2.46,
"learning_rate": 7.715930902111325e-05,
"loss": 0.0575,
"step": 3840
},
{
"epoch": 2.46,
"learning_rate": 7.683941138835573e-05,
"loss": 0.0705,
"step": 3850
},
{
"epoch": 2.47,
"learning_rate": 7.651951375559822e-05,
"loss": 0.0266,
"step": 3860
},
{
"epoch": 2.48,
"learning_rate": 7.61996161228407e-05,
"loss": 0.025,
"step": 3870
},
{
"epoch": 2.48,
"learning_rate": 7.587971849008317e-05,
"loss": 0.0691,
"step": 3880
},
{
"epoch": 2.49,
"learning_rate": 7.555982085732566e-05,
"loss": 0.0026,
"step": 3890
},
{
"epoch": 2.5,
"learning_rate": 7.523992322456814e-05,
"loss": 0.0591,
"step": 3900
},
{
"epoch": 2.5,
"eval_accuracy": 0.9713,
"eval_loss": 0.10392673313617706,
"eval_runtime": 47.9322,
"eval_samples_per_second": 208.628,
"eval_steps_per_second": 13.039,
"step": 3900
},
{
"epoch": 2.5,
"learning_rate": 7.492002559181063e-05,
"loss": 0.0062,
"step": 3910
},
{
"epoch": 2.51,
"learning_rate": 7.46001279590531e-05,
"loss": 0.0496,
"step": 3920
},
{
"epoch": 2.51,
"learning_rate": 7.428023032629558e-05,
"loss": 0.0039,
"step": 3930
},
{
"epoch": 2.52,
"learning_rate": 7.396033269353807e-05,
"loss": 0.0136,
"step": 3940
},
{
"epoch": 2.53,
"learning_rate": 7.364043506078055e-05,
"loss": 0.0372,
"step": 3950
},
{
"epoch": 2.53,
"learning_rate": 7.332053742802303e-05,
"loss": 0.0134,
"step": 3960
},
{
"epoch": 2.54,
"learning_rate": 7.300063979526551e-05,
"loss": 0.0085,
"step": 3970
},
{
"epoch": 2.55,
"learning_rate": 7.268074216250799e-05,
"loss": 0.0161,
"step": 3980
},
{
"epoch": 2.55,
"learning_rate": 7.236084452975049e-05,
"loss": 0.0126,
"step": 3990
},
{
"epoch": 2.56,
"learning_rate": 7.204094689699297e-05,
"loss": 0.0624,
"step": 4000
},
{
"epoch": 2.56,
"eval_accuracy": 0.9729,
"eval_loss": 0.11228275299072266,
"eval_runtime": 48.1823,
"eval_samples_per_second": 207.545,
"eval_steps_per_second": 12.972,
"step": 4000
},
{
"epoch": 2.57,
"learning_rate": 7.172104926423545e-05,
"loss": 0.0075,
"step": 4010
},
{
"epoch": 2.57,
"learning_rate": 7.140115163147793e-05,
"loss": 0.0191,
"step": 4020
},
{
"epoch": 2.58,
"learning_rate": 7.10812539987204e-05,
"loss": 0.0201,
"step": 4030
},
{
"epoch": 2.58,
"learning_rate": 7.07613563659629e-05,
"loss": 0.0092,
"step": 4040
},
{
"epoch": 2.59,
"learning_rate": 7.044145873320538e-05,
"loss": 0.0299,
"step": 4050
},
{
"epoch": 2.6,
"learning_rate": 7.012156110044786e-05,
"loss": 0.0233,
"step": 4060
},
{
"epoch": 2.6,
"learning_rate": 6.980166346769034e-05,
"loss": 0.0054,
"step": 4070
},
{
"epoch": 2.61,
"learning_rate": 6.948176583493282e-05,
"loss": 0.0243,
"step": 4080
},
{
"epoch": 2.62,
"learning_rate": 6.916186820217531e-05,
"loss": 0.0286,
"step": 4090
},
{
"epoch": 2.62,
"learning_rate": 6.884197056941779e-05,
"loss": 0.0309,
"step": 4100
},
{
"epoch": 2.62,
"eval_accuracy": 0.9751,
"eval_loss": 0.09711159765720367,
"eval_runtime": 46.9031,
"eval_samples_per_second": 213.206,
"eval_steps_per_second": 13.325,
"step": 4100
},
{
"epoch": 2.63,
"learning_rate": 6.852207293666027e-05,
"loss": 0.0211,
"step": 4110
},
{
"epoch": 2.64,
"learning_rate": 6.820217530390275e-05,
"loss": 0.0168,
"step": 4120
},
{
"epoch": 2.64,
"learning_rate": 6.788227767114523e-05,
"loss": 0.0195,
"step": 4130
},
{
"epoch": 2.65,
"learning_rate": 6.756238003838772e-05,
"loss": 0.0072,
"step": 4140
},
{
"epoch": 2.66,
"learning_rate": 6.72424824056302e-05,
"loss": 0.0243,
"step": 4150
},
{
"epoch": 2.66,
"learning_rate": 6.692258477287268e-05,
"loss": 0.027,
"step": 4160
},
{
"epoch": 2.67,
"learning_rate": 6.660268714011516e-05,
"loss": 0.039,
"step": 4170
},
{
"epoch": 2.67,
"learning_rate": 6.628278950735764e-05,
"loss": 0.017,
"step": 4180
},
{
"epoch": 2.68,
"learning_rate": 6.596289187460014e-05,
"loss": 0.008,
"step": 4190
},
{
"epoch": 2.69,
"learning_rate": 6.564299424184262e-05,
"loss": 0.015,
"step": 4200
},
{
"epoch": 2.69,
"eval_accuracy": 0.9738,
"eval_loss": 0.104751355946064,
"eval_runtime": 48.1137,
"eval_samples_per_second": 207.841,
"eval_steps_per_second": 12.99,
"step": 4200
},
{
"epoch": 2.69,
"learning_rate": 6.53230966090851e-05,
"loss": 0.0041,
"step": 4210
},
{
"epoch": 2.7,
"learning_rate": 6.500319897632758e-05,
"loss": 0.004,
"step": 4220
},
{
"epoch": 2.71,
"learning_rate": 6.468330134357006e-05,
"loss": 0.0206,
"step": 4230
},
{
"epoch": 2.71,
"learning_rate": 6.436340371081254e-05,
"loss": 0.012,
"step": 4240
},
{
"epoch": 2.72,
"learning_rate": 6.404350607805503e-05,
"loss": 0.0303,
"step": 4250
},
{
"epoch": 2.73,
"learning_rate": 6.372360844529751e-05,
"loss": 0.039,
"step": 4260
},
{
"epoch": 2.73,
"learning_rate": 6.340371081253999e-05,
"loss": 0.0225,
"step": 4270
},
{
"epoch": 2.74,
"learning_rate": 6.308381317978247e-05,
"loss": 0.019,
"step": 4280
},
{
"epoch": 2.74,
"learning_rate": 6.276391554702495e-05,
"loss": 0.0014,
"step": 4290
},
{
"epoch": 2.75,
"learning_rate": 6.244401791426744e-05,
"loss": 0.0155,
"step": 4300
},
{
"epoch": 2.75,
"eval_accuracy": 0.9749,
"eval_loss": 0.0932016596198082,
"eval_runtime": 61.5553,
"eval_samples_per_second": 162.456,
"eval_steps_per_second": 10.153,
"step": 4300
},
{
"epoch": 2.76,
"learning_rate": 6.212412028150992e-05,
"loss": 0.0027,
"step": 4310
},
{
"epoch": 2.76,
"learning_rate": 6.18042226487524e-05,
"loss": 0.0181,
"step": 4320
},
{
"epoch": 2.77,
"learning_rate": 6.148432501599488e-05,
"loss": 0.0063,
"step": 4330
},
{
"epoch": 2.78,
"learning_rate": 6.116442738323736e-05,
"loss": 0.0188,
"step": 4340
},
{
"epoch": 2.78,
"learning_rate": 6.084452975047985e-05,
"loss": 0.0195,
"step": 4350
},
{
"epoch": 2.79,
"learning_rate": 6.0524632117722334e-05,
"loss": 0.0057,
"step": 4360
},
{
"epoch": 2.8,
"learning_rate": 6.020473448496481e-05,
"loss": 0.0059,
"step": 4370
},
{
"epoch": 2.8,
"learning_rate": 5.9884836852207293e-05,
"loss": 0.0339,
"step": 4380
},
{
"epoch": 2.81,
"learning_rate": 5.956493921944978e-05,
"loss": 0.0381,
"step": 4390
},
{
"epoch": 2.82,
"learning_rate": 5.924504158669226e-05,
"loss": 0.0269,
"step": 4400
},
{
"epoch": 2.82,
"eval_accuracy": 0.9702,
"eval_loss": 0.11733580380678177,
"eval_runtime": 61.9528,
"eval_samples_per_second": 161.413,
"eval_steps_per_second": 10.088,
"step": 4400
},
{
"epoch": 2.82,
"learning_rate": 5.8925143953934746e-05,
"loss": 0.0158,
"step": 4410
},
{
"epoch": 2.83,
"learning_rate": 5.860524632117722e-05,
"loss": 0.0214,
"step": 4420
},
{
"epoch": 2.83,
"learning_rate": 5.8285348688419706e-05,
"loss": 0.0484,
"step": 4430
},
{
"epoch": 2.84,
"learning_rate": 5.796545105566219e-05,
"loss": 0.0152,
"step": 4440
},
{
"epoch": 2.85,
"learning_rate": 5.764555342290467e-05,
"loss": 0.1147,
"step": 4450
},
{
"epoch": 2.85,
"learning_rate": 5.732565579014716e-05,
"loss": 0.0292,
"step": 4460
},
{
"epoch": 2.86,
"learning_rate": 5.700575815738963e-05,
"loss": 0.052,
"step": 4470
},
{
"epoch": 2.87,
"learning_rate": 5.668586052463212e-05,
"loss": 0.0055,
"step": 4480
},
{
"epoch": 2.87,
"learning_rate": 5.63659628918746e-05,
"loss": 0.0053,
"step": 4490
},
{
"epoch": 2.88,
"learning_rate": 5.6046065259117085e-05,
"loss": 0.0021,
"step": 4500
},
{
"epoch": 2.88,
"eval_accuracy": 0.974,
"eval_loss": 0.0957212820649147,
"eval_runtime": 62.9349,
"eval_samples_per_second": 158.894,
"eval_steps_per_second": 9.931,
"step": 4500
},
{
"epoch": 2.89,
"learning_rate": 5.572616762635957e-05,
"loss": 0.043,
"step": 4510
},
{
"epoch": 2.89,
"learning_rate": 5.5406269993602045e-05,
"loss": 0.0207,
"step": 4520
},
{
"epoch": 2.9,
"learning_rate": 5.508637236084453e-05,
"loss": 0.006,
"step": 4530
},
{
"epoch": 2.9,
"learning_rate": 5.476647472808701e-05,
"loss": 0.0043,
"step": 4540
},
{
"epoch": 2.91,
"learning_rate": 5.44465770953295e-05,
"loss": 0.0257,
"step": 4550
},
{
"epoch": 2.92,
"learning_rate": 5.4126679462571984e-05,
"loss": 0.0241,
"step": 4560
},
{
"epoch": 2.92,
"learning_rate": 5.380678182981446e-05,
"loss": 0.019,
"step": 4570
},
{
"epoch": 2.93,
"learning_rate": 5.3486884197056944e-05,
"loss": 0.0204,
"step": 4580
},
{
"epoch": 2.94,
"learning_rate": 5.3166986564299424e-05,
"loss": 0.0182,
"step": 4590
},
{
"epoch": 2.94,
"learning_rate": 5.284708893154191e-05,
"loss": 0.0238,
"step": 4600
},
{
"epoch": 2.94,
"eval_accuracy": 0.9742,
"eval_loss": 0.09226856380701065,
"eval_runtime": 62.7299,
"eval_samples_per_second": 159.414,
"eval_steps_per_second": 9.963,
"step": 4600
},
{
"epoch": 2.95,
"learning_rate": 5.25271912987844e-05,
"loss": 0.0036,
"step": 4610
},
{
"epoch": 2.96,
"learning_rate": 5.220729366602687e-05,
"loss": 0.0672,
"step": 4620
},
{
"epoch": 2.96,
"learning_rate": 5.188739603326935e-05,
"loss": 0.0383,
"step": 4630
},
{
"epoch": 2.97,
"learning_rate": 5.1567498400511836e-05,
"loss": 0.0044,
"step": 4640
},
{
"epoch": 2.98,
"learning_rate": 5.124760076775432e-05,
"loss": 0.0257,
"step": 4650
},
{
"epoch": 2.98,
"learning_rate": 5.092770313499681e-05,
"loss": 0.0011,
"step": 4660
},
{
"epoch": 2.99,
"learning_rate": 5.060780550223928e-05,
"loss": 0.0034,
"step": 4670
},
{
"epoch": 2.99,
"learning_rate": 5.028790786948176e-05,
"loss": 0.0256,
"step": 4680
},
{
"epoch": 3.0,
"learning_rate": 4.996801023672425e-05,
"loss": 0.0129,
"step": 4690
},
{
"epoch": 3.01,
"learning_rate": 4.9648112603966736e-05,
"loss": 0.0015,
"step": 4700
},
{
"epoch": 3.01,
"eval_accuracy": 0.9748,
"eval_loss": 0.09875385463237762,
"eval_runtime": 62.4215,
"eval_samples_per_second": 160.201,
"eval_steps_per_second": 10.013,
"step": 4700
},
{
"epoch": 3.01,
"learning_rate": 4.9328214971209215e-05,
"loss": 0.0154,
"step": 4710
},
{
"epoch": 3.02,
"learning_rate": 4.9008317338451695e-05,
"loss": 0.0012,
"step": 4720
},
{
"epoch": 3.03,
"learning_rate": 4.868841970569418e-05,
"loss": 0.001,
"step": 4730
},
{
"epoch": 3.03,
"learning_rate": 4.836852207293666e-05,
"loss": 0.0069,
"step": 4740
},
{
"epoch": 3.04,
"learning_rate": 4.804862444017915e-05,
"loss": 0.0035,
"step": 4750
},
{
"epoch": 3.05,
"learning_rate": 4.772872680742163e-05,
"loss": 0.0024,
"step": 4760
},
{
"epoch": 3.05,
"learning_rate": 4.740882917466411e-05,
"loss": 0.0026,
"step": 4770
},
{
"epoch": 3.06,
"learning_rate": 4.7088931541906594e-05,
"loss": 0.0014,
"step": 4780
},
{
"epoch": 3.06,
"learning_rate": 4.6769033909149074e-05,
"loss": 0.0008,
"step": 4790
},
{
"epoch": 3.07,
"learning_rate": 4.644913627639156e-05,
"loss": 0.0016,
"step": 4800
},
{
"epoch": 3.07,
"eval_accuracy": 0.9791,
"eval_loss": 0.08298031985759735,
"eval_runtime": 61.1981,
"eval_samples_per_second": 163.404,
"eval_steps_per_second": 10.213,
"step": 4800
},
{
"epoch": 3.08,
"learning_rate": 4.612923864363404e-05,
"loss": 0.0024,
"step": 4810
},
{
"epoch": 3.08,
"learning_rate": 4.580934101087652e-05,
"loss": 0.0015,
"step": 4820
},
{
"epoch": 3.09,
"learning_rate": 4.548944337811901e-05,
"loss": 0.0007,
"step": 4830
},
{
"epoch": 3.1,
"learning_rate": 4.516954574536149e-05,
"loss": 0.0008,
"step": 4840
},
{
"epoch": 3.1,
"learning_rate": 4.4849648112603973e-05,
"loss": 0.0139,
"step": 4850
},
{
"epoch": 3.11,
"learning_rate": 4.4529750479846447e-05,
"loss": 0.0019,
"step": 4860
},
{
"epoch": 3.12,
"learning_rate": 4.420985284708893e-05,
"loss": 0.0011,
"step": 4870
},
{
"epoch": 3.12,
"learning_rate": 4.388995521433142e-05,
"loss": 0.001,
"step": 4880
},
{
"epoch": 3.13,
"learning_rate": 4.35700575815739e-05,
"loss": 0.0009,
"step": 4890
},
{
"epoch": 3.13,
"learning_rate": 4.325015994881638e-05,
"loss": 0.001,
"step": 4900
},
{
"epoch": 3.13,
"eval_accuracy": 0.9795,
"eval_loss": 0.08240580558776855,
"eval_runtime": 61.0908,
"eval_samples_per_second": 163.691,
"eval_steps_per_second": 10.231,
"step": 4900
},
{
"epoch": 3.14,
"learning_rate": 4.293026231605886e-05,
"loss": 0.001,
"step": 4910
},
{
"epoch": 3.15,
"learning_rate": 4.2610364683301346e-05,
"loss": 0.0007,
"step": 4920
},
{
"epoch": 3.15,
"learning_rate": 4.229046705054383e-05,
"loss": 0.0011,
"step": 4930
},
{
"epoch": 3.16,
"learning_rate": 4.197056941778631e-05,
"loss": 0.0008,
"step": 4940
},
{
"epoch": 3.17,
"learning_rate": 4.165067178502879e-05,
"loss": 0.0013,
"step": 4950
},
{
"epoch": 3.17,
"learning_rate": 4.133077415227127e-05,
"loss": 0.0035,
"step": 4960
},
{
"epoch": 3.18,
"learning_rate": 4.101087651951376e-05,
"loss": 0.0008,
"step": 4970
},
{
"epoch": 3.19,
"learning_rate": 4.0690978886756245e-05,
"loss": 0.0035,
"step": 4980
},
{
"epoch": 3.19,
"learning_rate": 4.0371081253998725e-05,
"loss": 0.001,
"step": 4990
},
{
"epoch": 3.2,
"learning_rate": 4.0051183621241205e-05,
"loss": 0.0008,
"step": 5000
},
{
"epoch": 3.2,
"eval_accuracy": 0.976,
"eval_loss": 0.0969667136669159,
"eval_runtime": 61.996,
"eval_samples_per_second": 161.301,
"eval_steps_per_second": 10.081,
"step": 5000
},
{
"epoch": 3.21,
"learning_rate": 3.9731285988483684e-05,
"loss": 0.0133,
"step": 5010
},
{
"epoch": 3.21,
"learning_rate": 3.941138835572617e-05,
"loss": 0.0008,
"step": 5020
},
{
"epoch": 3.22,
"learning_rate": 3.909149072296866e-05,
"loss": 0.0037,
"step": 5030
},
{
"epoch": 3.22,
"learning_rate": 3.877159309021113e-05,
"loss": 0.0007,
"step": 5040
},
{
"epoch": 3.23,
"learning_rate": 3.845169545745362e-05,
"loss": 0.0015,
"step": 5050
},
{
"epoch": 3.24,
"learning_rate": 3.81317978246961e-05,
"loss": 0.0008,
"step": 5060
},
{
"epoch": 3.24,
"learning_rate": 3.7811900191938584e-05,
"loss": 0.0014,
"step": 5070
},
{
"epoch": 3.25,
"learning_rate": 3.7492002559181063e-05,
"loss": 0.0007,
"step": 5080
},
{
"epoch": 3.26,
"learning_rate": 3.717210492642354e-05,
"loss": 0.0012,
"step": 5090
},
{
"epoch": 3.26,
"learning_rate": 3.685220729366603e-05,
"loss": 0.0149,
"step": 5100
},
{
"epoch": 3.26,
"eval_accuracy": 0.9796,
"eval_loss": 0.0864199697971344,
"eval_runtime": 61.7022,
"eval_samples_per_second": 162.069,
"eval_steps_per_second": 10.129,
"step": 5100
},
{
"epoch": 3.27,
"learning_rate": 3.653230966090851e-05,
"loss": 0.001,
"step": 5110
},
{
"epoch": 3.28,
"learning_rate": 3.6212412028150996e-05,
"loss": 0.001,
"step": 5120
},
{
"epoch": 3.28,
"learning_rate": 3.5892514395393476e-05,
"loss": 0.0015,
"step": 5130
},
{
"epoch": 3.29,
"learning_rate": 3.5572616762635956e-05,
"loss": 0.0018,
"step": 5140
},
{
"epoch": 3.29,
"learning_rate": 3.525271912987844e-05,
"loss": 0.002,
"step": 5150
},
{
"epoch": 3.3,
"learning_rate": 3.493282149712092e-05,
"loss": 0.0019,
"step": 5160
},
{
"epoch": 3.31,
"learning_rate": 3.461292386436341e-05,
"loss": 0.0063,
"step": 5170
},
{
"epoch": 3.31,
"learning_rate": 3.429302623160589e-05,
"loss": 0.0008,
"step": 5180
},
{
"epoch": 3.32,
"learning_rate": 3.397312859884837e-05,
"loss": 0.0273,
"step": 5190
},
{
"epoch": 3.33,
"learning_rate": 3.3653230966090855e-05,
"loss": 0.008,
"step": 5200
},
{
"epoch": 3.33,
"eval_accuracy": 0.9778,
"eval_loss": 0.09051262587308884,
"eval_runtime": 62.6123,
"eval_samples_per_second": 159.713,
"eval_steps_per_second": 9.982,
"step": 5200
},
{
"epoch": 3.33,
"learning_rate": 3.3333333333333335e-05,
"loss": 0.0007,
"step": 5210
},
{
"epoch": 3.34,
"learning_rate": 3.3013435700575815e-05,
"loss": 0.001,
"step": 5220
},
{
"epoch": 3.35,
"learning_rate": 3.26935380678183e-05,
"loss": 0.0065,
"step": 5230
},
{
"epoch": 3.35,
"learning_rate": 3.237364043506078e-05,
"loss": 0.0006,
"step": 5240
},
{
"epoch": 3.36,
"learning_rate": 3.205374280230327e-05,
"loss": 0.0007,
"step": 5250
},
{
"epoch": 3.37,
"learning_rate": 3.173384516954575e-05,
"loss": 0.0165,
"step": 5260
},
{
"epoch": 3.37,
"learning_rate": 3.141394753678823e-05,
"loss": 0.0119,
"step": 5270
},
{
"epoch": 3.38,
"learning_rate": 3.1094049904030714e-05,
"loss": 0.0006,
"step": 5280
},
{
"epoch": 3.38,
"learning_rate": 3.0774152271273194e-05,
"loss": 0.0074,
"step": 5290
},
{
"epoch": 3.39,
"learning_rate": 3.0454254638515677e-05,
"loss": 0.0064,
"step": 5300
},
{
"epoch": 3.39,
"eval_accuracy": 0.9799,
"eval_loss": 0.08600812405347824,
"eval_runtime": 63.7405,
"eval_samples_per_second": 156.886,
"eval_steps_per_second": 9.805,
"step": 5300
},
{
"epoch": 3.4,
"learning_rate": 3.0134357005758157e-05,
"loss": 0.0027,
"step": 5310
},
{
"epoch": 3.4,
"learning_rate": 2.981445937300064e-05,
"loss": 0.0008,
"step": 5320
},
{
"epoch": 3.41,
"learning_rate": 2.9494561740243127e-05,
"loss": 0.0008,
"step": 5330
},
{
"epoch": 3.42,
"learning_rate": 2.9174664107485606e-05,
"loss": 0.0016,
"step": 5340
},
{
"epoch": 3.42,
"learning_rate": 2.885476647472809e-05,
"loss": 0.0174,
"step": 5350
},
{
"epoch": 3.43,
"learning_rate": 2.853486884197057e-05,
"loss": 0.0014,
"step": 5360
},
{
"epoch": 3.44,
"learning_rate": 2.8214971209213053e-05,
"loss": 0.0007,
"step": 5370
},
{
"epoch": 3.44,
"learning_rate": 2.789507357645554e-05,
"loss": 0.0007,
"step": 5380
},
{
"epoch": 3.45,
"learning_rate": 2.7575175943698016e-05,
"loss": 0.0035,
"step": 5390
},
{
"epoch": 3.45,
"learning_rate": 2.7255278310940502e-05,
"loss": 0.0009,
"step": 5400
},
{
"epoch": 3.45,
"eval_accuracy": 0.9781,
"eval_loss": 0.08675622195005417,
"eval_runtime": 61.5807,
"eval_samples_per_second": 162.388,
"eval_steps_per_second": 10.149,
"step": 5400
},
{
"epoch": 3.46,
"learning_rate": 2.6935380678182982e-05,
"loss": 0.0007,
"step": 5410
},
{
"epoch": 3.47,
"learning_rate": 2.6615483045425465e-05,
"loss": 0.0009,
"step": 5420
},
{
"epoch": 3.47,
"learning_rate": 2.6295585412667952e-05,
"loss": 0.0011,
"step": 5430
},
{
"epoch": 3.48,
"learning_rate": 2.5975687779910428e-05,
"loss": 0.0006,
"step": 5440
},
{
"epoch": 3.49,
"learning_rate": 2.5655790147152915e-05,
"loss": 0.0007,
"step": 5450
},
{
"epoch": 3.49,
"learning_rate": 2.533589251439539e-05,
"loss": 0.0009,
"step": 5460
},
{
"epoch": 3.5,
"learning_rate": 2.5015994881637878e-05,
"loss": 0.0112,
"step": 5470
},
{
"epoch": 3.51,
"learning_rate": 2.4696097248880358e-05,
"loss": 0.0006,
"step": 5480
},
{
"epoch": 3.51,
"learning_rate": 2.437619961612284e-05,
"loss": 0.0007,
"step": 5490
},
{
"epoch": 3.52,
"learning_rate": 2.4056301983365324e-05,
"loss": 0.0007,
"step": 5500
},
{
"epoch": 3.52,
"eval_accuracy": 0.9786,
"eval_loss": 0.08598293364048004,
"eval_runtime": 60.1457,
"eval_samples_per_second": 166.263,
"eval_steps_per_second": 10.391,
"step": 5500
},
{
"epoch": 3.53,
"learning_rate": 2.3736404350607807e-05,
"loss": 0.0008,
"step": 5510
},
{
"epoch": 3.53,
"learning_rate": 2.341650671785029e-05,
"loss": 0.0033,
"step": 5520
},
{
"epoch": 3.54,
"learning_rate": 2.309660908509277e-05,
"loss": 0.0009,
"step": 5530
},
{
"epoch": 3.54,
"learning_rate": 2.2776711452335254e-05,
"loss": 0.0006,
"step": 5540
},
{
"epoch": 3.55,
"learning_rate": 2.2456813819577733e-05,
"loss": 0.0007,
"step": 5550
},
{
"epoch": 3.56,
"learning_rate": 2.213691618682022e-05,
"loss": 0.0006,
"step": 5560
},
{
"epoch": 3.56,
"learning_rate": 2.1817018554062703e-05,
"loss": 0.0013,
"step": 5570
},
{
"epoch": 3.57,
"learning_rate": 2.1497120921305183e-05,
"loss": 0.0037,
"step": 5580
},
{
"epoch": 3.58,
"learning_rate": 2.1177223288547666e-05,
"loss": 0.0009,
"step": 5590
},
{
"epoch": 3.58,
"learning_rate": 2.0857325655790146e-05,
"loss": 0.0006,
"step": 5600
},
{
"epoch": 3.58,
"eval_accuracy": 0.9796,
"eval_loss": 0.0850202739238739,
"eval_runtime": 62.0241,
"eval_samples_per_second": 161.228,
"eval_steps_per_second": 10.077,
"step": 5600
},
{
"epoch": 3.59,
"learning_rate": 2.0537428023032633e-05,
"loss": 0.0047,
"step": 5610
},
{
"epoch": 3.6,
"learning_rate": 2.0217530390275112e-05,
"loss": 0.0013,
"step": 5620
},
{
"epoch": 3.6,
"learning_rate": 1.9897632757517596e-05,
"loss": 0.0018,
"step": 5630
},
{
"epoch": 3.61,
"learning_rate": 1.957773512476008e-05,
"loss": 0.0006,
"step": 5640
},
{
"epoch": 3.61,
"learning_rate": 1.925783749200256e-05,
"loss": 0.0134,
"step": 5650
},
{
"epoch": 3.62,
"learning_rate": 1.8937939859245045e-05,
"loss": 0.0014,
"step": 5660
},
{
"epoch": 3.63,
"learning_rate": 1.8618042226487525e-05,
"loss": 0.0018,
"step": 5670
},
{
"epoch": 3.63,
"learning_rate": 1.8298144593730008e-05,
"loss": 0.0007,
"step": 5680
},
{
"epoch": 3.64,
"learning_rate": 1.7978246960972488e-05,
"loss": 0.0006,
"step": 5690
},
{
"epoch": 3.65,
"learning_rate": 1.765834932821497e-05,
"loss": 0.0005,
"step": 5700
},
{
"epoch": 3.65,
"eval_accuracy": 0.9804,
"eval_loss": 0.08065660297870636,
"eval_runtime": 63.2588,
"eval_samples_per_second": 158.081,
"eval_steps_per_second": 9.88,
"step": 5700
},
{
"epoch": 3.65,
"learning_rate": 1.7338451695457454e-05,
"loss": 0.01,
"step": 5710
},
{
"epoch": 3.66,
"learning_rate": 1.7018554062699938e-05,
"loss": 0.0005,
"step": 5720
},
{
"epoch": 3.67,
"learning_rate": 1.669865642994242e-05,
"loss": 0.0009,
"step": 5730
},
{
"epoch": 3.67,
"learning_rate": 1.63787587971849e-05,
"loss": 0.0005,
"step": 5740
},
{
"epoch": 3.68,
"learning_rate": 1.6058861164427384e-05,
"loss": 0.0007,
"step": 5750
},
{
"epoch": 3.69,
"learning_rate": 1.5738963531669867e-05,
"loss": 0.002,
"step": 5760
},
{
"epoch": 3.69,
"learning_rate": 1.541906589891235e-05,
"loss": 0.0005,
"step": 5770
},
{
"epoch": 3.7,
"learning_rate": 1.5099168266154832e-05,
"loss": 0.0008,
"step": 5780
},
{
"epoch": 3.7,
"learning_rate": 1.4779270633397313e-05,
"loss": 0.0026,
"step": 5790
},
{
"epoch": 3.71,
"learning_rate": 1.4459373000639795e-05,
"loss": 0.0005,
"step": 5800
},
{
"epoch": 3.71,
"eval_accuracy": 0.981,
"eval_loss": 0.07981295883655548,
"eval_runtime": 62.5683,
"eval_samples_per_second": 159.825,
"eval_steps_per_second": 9.989,
"step": 5800
},
{
"epoch": 3.72,
"learning_rate": 1.413947536788228e-05,
"loss": 0.0015,
"step": 5810
},
{
"epoch": 3.72,
"learning_rate": 1.3819577735124761e-05,
"loss": 0.0007,
"step": 5820
},
{
"epoch": 3.73,
"learning_rate": 1.3499680102367243e-05,
"loss": 0.0058,
"step": 5830
},
{
"epoch": 3.74,
"learning_rate": 1.3179782469609724e-05,
"loss": 0.0006,
"step": 5840
},
{
"epoch": 3.74,
"learning_rate": 1.2859884836852207e-05,
"loss": 0.0058,
"step": 5850
},
{
"epoch": 3.75,
"learning_rate": 1.253998720409469e-05,
"loss": 0.0007,
"step": 5860
},
{
"epoch": 3.76,
"learning_rate": 1.2220089571337174e-05,
"loss": 0.0005,
"step": 5870
},
{
"epoch": 3.76,
"learning_rate": 1.1900191938579655e-05,
"loss": 0.0008,
"step": 5880
},
{
"epoch": 3.77,
"learning_rate": 1.1580294305822137e-05,
"loss": 0.0004,
"step": 5890
},
{
"epoch": 3.77,
"learning_rate": 1.126039667306462e-05,
"loss": 0.0016,
"step": 5900
},
{
"epoch": 3.77,
"eval_accuracy": 0.9808,
"eval_loss": 0.07946650683879852,
"eval_runtime": 60.9852,
"eval_samples_per_second": 163.974,
"eval_steps_per_second": 10.248,
"step": 5900
},
{
"epoch": 3.78,
"learning_rate": 1.0940499040307102e-05,
"loss": 0.0006,
"step": 5910
},
{
"epoch": 3.79,
"learning_rate": 1.0620601407549585e-05,
"loss": 0.0058,
"step": 5920
},
{
"epoch": 3.79,
"learning_rate": 1.0300703774792066e-05,
"loss": 0.0035,
"step": 5930
},
{
"epoch": 3.8,
"learning_rate": 9.98080614203455e-06,
"loss": 0.0024,
"step": 5940
},
{
"epoch": 3.81,
"learning_rate": 9.660908509277033e-06,
"loss": 0.0005,
"step": 5950
},
{
"epoch": 3.81,
"learning_rate": 9.341010876519514e-06,
"loss": 0.0048,
"step": 5960
},
{
"epoch": 3.82,
"learning_rate": 9.021113243761997e-06,
"loss": 0.001,
"step": 5970
},
{
"epoch": 3.83,
"learning_rate": 8.701215611004479e-06,
"loss": 0.0006,
"step": 5980
},
{
"epoch": 3.83,
"learning_rate": 8.38131797824696e-06,
"loss": 0.001,
"step": 5990
},
{
"epoch": 3.84,
"learning_rate": 8.061420345489444e-06,
"loss": 0.0068,
"step": 6000
},
{
"epoch": 3.84,
"eval_accuracy": 0.9806,
"eval_loss": 0.08210848271846771,
"eval_runtime": 64.268,
"eval_samples_per_second": 155.598,
"eval_steps_per_second": 9.725,
"step": 6000
},
{
"epoch": 3.85,
"learning_rate": 7.741522712731925e-06,
"loss": 0.0004,
"step": 6010
},
{
"epoch": 3.85,
"learning_rate": 7.421625079974409e-06,
"loss": 0.0005,
"step": 6020
},
{
"epoch": 3.86,
"learning_rate": 7.101727447216891e-06,
"loss": 0.0007,
"step": 6030
},
{
"epoch": 3.86,
"learning_rate": 6.781829814459373e-06,
"loss": 0.0005,
"step": 6040
},
{
"epoch": 3.87,
"learning_rate": 6.461932181701856e-06,
"loss": 0.0164,
"step": 6050
},
{
"epoch": 3.88,
"learning_rate": 6.142034548944338e-06,
"loss": 0.0005,
"step": 6060
},
{
"epoch": 3.88,
"learning_rate": 5.82213691618682e-06,
"loss": 0.0007,
"step": 6070
},
{
"epoch": 3.89,
"learning_rate": 5.502239283429303e-06,
"loss": 0.0006,
"step": 6080
},
{
"epoch": 3.9,
"learning_rate": 5.182341650671786e-06,
"loss": 0.0007,
"step": 6090
},
{
"epoch": 3.9,
"learning_rate": 4.862444017914268e-06,
"loss": 0.001,
"step": 6100
},
{
"epoch": 3.9,
"eval_accuracy": 0.981,
"eval_loss": 0.08095718175172806,
"eval_runtime": 62.2344,
"eval_samples_per_second": 160.683,
"eval_steps_per_second": 10.043,
"step": 6100
},
{
"epoch": 3.91,
"learning_rate": 4.5425463851567495e-06,
"loss": 0.0005,
"step": 6110
},
{
"epoch": 3.92,
"learning_rate": 4.222648752399233e-06,
"loss": 0.0006,
"step": 6120
},
{
"epoch": 3.92,
"learning_rate": 3.902751119641715e-06,
"loss": 0.0067,
"step": 6130
},
{
"epoch": 3.93,
"learning_rate": 3.5828534868841974e-06,
"loss": 0.001,
"step": 6140
},
{
"epoch": 3.93,
"learning_rate": 3.2629558541266794e-06,
"loss": 0.0009,
"step": 6150
},
{
"epoch": 3.94,
"learning_rate": 2.943058221369162e-06,
"loss": 0.0005,
"step": 6160
},
{
"epoch": 3.95,
"learning_rate": 2.6231605886116445e-06,
"loss": 0.0005,
"step": 6170
},
{
"epoch": 3.95,
"learning_rate": 2.303262955854127e-06,
"loss": 0.0006,
"step": 6180
},
{
"epoch": 3.96,
"learning_rate": 1.9833653230966092e-06,
"loss": 0.005,
"step": 6190
},
{
"epoch": 3.97,
"learning_rate": 1.6634676903390916e-06,
"loss": 0.0009,
"step": 6200
},
{
"epoch": 3.97,
"eval_accuracy": 0.9812,
"eval_loss": 0.07883724570274353,
"eval_runtime": 62.3579,
"eval_samples_per_second": 160.364,
"eval_steps_per_second": 10.023,
"step": 6200
},
{
"epoch": 3.97,
"learning_rate": 1.343570057581574e-06,
"loss": 0.0006,
"step": 6210
},
{
"epoch": 3.98,
"learning_rate": 1.0236724248240563e-06,
"loss": 0.0005,
"step": 6220
},
{
"epoch": 3.99,
"learning_rate": 7.037747920665388e-07,
"loss": 0.0005,
"step": 6230
},
{
"epoch": 3.99,
"learning_rate": 3.838771593090211e-07,
"loss": 0.0005,
"step": 6240
},
{
"epoch": 4.0,
"learning_rate": 6.397952655150352e-08,
"loss": 0.0005,
"step": 6250
},
{
"epoch": 4.0,
"step": 6252,
"total_flos": 1.54995091808256e+19,
"train_loss": 0.08309991182349008,
"train_runtime": 5212.8854,
"train_samples_per_second": 38.366,
"train_steps_per_second": 1.199
}
],
"max_steps": 6252,
"num_train_epochs": 4,
"total_flos": 1.54995091808256e+19,
"trial_name": null,
"trial_params": null
}