tmnam20's picture
Training in progress, step 105000, checkpoint
53de162
raw
history blame
39.4 kB
{
"best_metric": 0.9800674319267273,
"best_model_checkpoint": "/kaggle/output/checkpoint-103000",
"epoch": 4.278031290743155,
"eval_steps": 1000,
"global_step": 105000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 2.7777777777777777e-11,
"loss": 1.2184,
"step": 1
},
{
"epoch": 0.04,
"learning_rate": 2.7750000000000004e-08,
"loss": 1.1394,
"step": 1000
},
{
"epoch": 0.04,
"eval_accuracy": 0.3327345309381238,
"eval_loss": 1.1149410009384155,
"eval_runtime": 20.6803,
"eval_samples_per_second": 242.26,
"eval_steps_per_second": 30.319,
"step": 1000
},
{
"epoch": 0.08,
"learning_rate": 5.5527777777777784e-08,
"loss": 1.1141,
"step": 2000
},
{
"epoch": 0.08,
"eval_accuracy": 0.3401197604790419,
"eval_loss": 1.104099988937378,
"eval_runtime": 20.8477,
"eval_samples_per_second": 240.314,
"eval_steps_per_second": 30.075,
"step": 2000
},
{
"epoch": 0.12,
"learning_rate": 8.330555555555556e-08,
"loss": 1.116,
"step": 3000
},
{
"epoch": 0.12,
"eval_accuracy": 0.3407185628742515,
"eval_loss": 1.1040862798690796,
"eval_runtime": 20.6818,
"eval_samples_per_second": 242.242,
"eval_steps_per_second": 30.317,
"step": 3000
},
{
"epoch": 0.16,
"learning_rate": 1.1108333333333333e-07,
"loss": 1.1158,
"step": 4000
},
{
"epoch": 0.16,
"eval_accuracy": 0.32894211576846305,
"eval_loss": 1.1020556688308716,
"eval_runtime": 20.8541,
"eval_samples_per_second": 240.241,
"eval_steps_per_second": 30.066,
"step": 4000
},
{
"epoch": 0.2,
"learning_rate": 1.3883333333333335e-07,
"loss": 1.1135,
"step": 5000
},
{
"epoch": 0.2,
"eval_accuracy": 0.34271457085828344,
"eval_loss": 1.1008552312850952,
"eval_runtime": 20.8055,
"eval_samples_per_second": 240.802,
"eval_steps_per_second": 30.136,
"step": 5000
},
{
"epoch": 0.24,
"learning_rate": 1.6658333333333335e-07,
"loss": 1.1121,
"step": 6000
},
{
"epoch": 0.24,
"eval_accuracy": 0.3395209580838323,
"eval_loss": 1.1004050970077515,
"eval_runtime": 20.8985,
"eval_samples_per_second": 239.731,
"eval_steps_per_second": 30.002,
"step": 6000
},
{
"epoch": 0.29,
"learning_rate": 1.9436111111111112e-07,
"loss": 1.1089,
"step": 7000
},
{
"epoch": 0.29,
"eval_accuracy": 0.35788423153692617,
"eval_loss": 1.0985721349716187,
"eval_runtime": 20.84,
"eval_samples_per_second": 240.403,
"eval_steps_per_second": 30.086,
"step": 7000
},
{
"epoch": 0.33,
"learning_rate": 2.2213888888888891e-07,
"loss": 1.1079,
"step": 8000
},
{
"epoch": 0.33,
"eval_accuracy": 0.3331337325349301,
"eval_loss": 1.098374843597412,
"eval_runtime": 20.7886,
"eval_samples_per_second": 240.998,
"eval_steps_per_second": 30.161,
"step": 8000
},
{
"epoch": 0.37,
"learning_rate": 2.4988888888888893e-07,
"loss": 1.1087,
"step": 9000
},
{
"epoch": 0.37,
"eval_accuracy": 0.34510978043912177,
"eval_loss": 1.0993521213531494,
"eval_runtime": 20.782,
"eval_samples_per_second": 241.074,
"eval_steps_per_second": 30.17,
"step": 9000
},
{
"epoch": 0.41,
"learning_rate": 2.776666666666667e-07,
"loss": 1.109,
"step": 10000
},
{
"epoch": 0.41,
"eval_accuracy": 0.3475049900199601,
"eval_loss": 1.0967597961425781,
"eval_runtime": 20.6798,
"eval_samples_per_second": 242.265,
"eval_steps_per_second": 30.319,
"step": 10000
},
{
"epoch": 0.45,
"learning_rate": 3.054444444444444e-07,
"loss": 1.1052,
"step": 11000
},
{
"epoch": 0.45,
"eval_accuracy": 0.37544910179640717,
"eval_loss": 1.0941349267959595,
"eval_runtime": 20.8641,
"eval_samples_per_second": 240.126,
"eval_steps_per_second": 30.052,
"step": 11000
},
{
"epoch": 0.49,
"learning_rate": 3.3322222222222225e-07,
"loss": 1.105,
"step": 12000
},
{
"epoch": 0.49,
"eval_accuracy": 0.3834331337325349,
"eval_loss": 1.0927647352218628,
"eval_runtime": 20.6541,
"eval_samples_per_second": 242.567,
"eval_steps_per_second": 30.357,
"step": 12000
},
{
"epoch": 0.53,
"learning_rate": 3.609722222222222e-07,
"loss": 1.1016,
"step": 13000
},
{
"epoch": 0.53,
"eval_accuracy": 0.3457085828343313,
"eval_loss": 1.0942081212997437,
"eval_runtime": 21.0733,
"eval_samples_per_second": 237.742,
"eval_steps_per_second": 29.753,
"step": 13000
},
{
"epoch": 0.57,
"learning_rate": 3.8875e-07,
"loss": 1.1031,
"step": 14000
},
{
"epoch": 0.57,
"eval_accuracy": 0.37005988023952097,
"eval_loss": 1.0918152332305908,
"eval_runtime": 20.9151,
"eval_samples_per_second": 239.54,
"eval_steps_per_second": 29.978,
"step": 14000
},
{
"epoch": 0.61,
"learning_rate": 4.1652777777777786e-07,
"loss": 1.1026,
"step": 15000
},
{
"epoch": 0.61,
"eval_accuracy": 0.3790419161676647,
"eval_loss": 1.0895211696624756,
"eval_runtime": 21.0591,
"eval_samples_per_second": 237.902,
"eval_steps_per_second": 29.773,
"step": 15000
},
{
"epoch": 0.65,
"learning_rate": 4.4427777777777783e-07,
"loss": 1.0988,
"step": 16000
},
{
"epoch": 0.65,
"eval_accuracy": 0.4101796407185629,
"eval_loss": 1.0852997303009033,
"eval_runtime": 20.9509,
"eval_samples_per_second": 239.131,
"eval_steps_per_second": 29.927,
"step": 16000
},
{
"epoch": 0.69,
"learning_rate": 4.720555555555556e-07,
"loss": 1.0974,
"step": 17000
},
{
"epoch": 0.69,
"eval_accuracy": 0.43213572854291415,
"eval_loss": 1.0791982412338257,
"eval_runtime": 20.7526,
"eval_samples_per_second": 241.415,
"eval_steps_per_second": 30.213,
"step": 17000
},
{
"epoch": 0.73,
"learning_rate": 4.998055555555556e-07,
"loss": 1.0932,
"step": 18000
},
{
"epoch": 0.73,
"eval_accuracy": 0.4275449101796407,
"eval_loss": 1.072191596031189,
"eval_runtime": 21.2435,
"eval_samples_per_second": 235.837,
"eval_steps_per_second": 29.515,
"step": 18000
},
{
"epoch": 0.77,
"learning_rate": 5.275833333333334e-07,
"loss": 1.0833,
"step": 19000
},
{
"epoch": 0.77,
"eval_accuracy": 0.43233532934131735,
"eval_loss": 1.06425940990448,
"eval_runtime": 20.7923,
"eval_samples_per_second": 240.955,
"eval_steps_per_second": 30.155,
"step": 19000
},
{
"epoch": 0.81,
"learning_rate": 5.553333333333334e-07,
"loss": 1.0787,
"step": 20000
},
{
"epoch": 0.81,
"eval_accuracy": 0.4295409181636727,
"eval_loss": 1.0638529062271118,
"eval_runtime": 21.0018,
"eval_samples_per_second": 238.551,
"eval_steps_per_second": 29.855,
"step": 20000
},
{
"epoch": 0.86,
"learning_rate": 5.831111111111111e-07,
"loss": 1.0779,
"step": 21000
},
{
"epoch": 0.86,
"eval_accuracy": 0.4243512974051896,
"eval_loss": 1.0603673458099365,
"eval_runtime": 20.9689,
"eval_samples_per_second": 238.926,
"eval_steps_per_second": 29.901,
"step": 21000
},
{
"epoch": 0.9,
"learning_rate": 6.108888888888888e-07,
"loss": 1.0751,
"step": 22000
},
{
"epoch": 0.9,
"eval_accuracy": 0.43233532934131735,
"eval_loss": 1.0603009462356567,
"eval_runtime": 20.8897,
"eval_samples_per_second": 239.831,
"eval_steps_per_second": 30.015,
"step": 22000
},
{
"epoch": 0.94,
"learning_rate": 6.386388888888889e-07,
"loss": 1.0776,
"step": 23000
},
{
"epoch": 0.94,
"eval_accuracy": 0.42734530938123755,
"eval_loss": 1.0591468811035156,
"eval_runtime": 20.964,
"eval_samples_per_second": 238.981,
"eval_steps_per_second": 29.908,
"step": 23000
},
{
"epoch": 0.98,
"learning_rate": 6.664166666666667e-07,
"loss": 1.0754,
"step": 24000
},
{
"epoch": 0.98,
"eval_accuracy": 0.4245508982035928,
"eval_loss": 1.0589721202850342,
"eval_runtime": 20.9053,
"eval_samples_per_second": 239.652,
"eval_steps_per_second": 29.992,
"step": 24000
},
{
"epoch": 1.02,
"learning_rate": 6.941666666666667e-07,
"loss": 1.0736,
"step": 25000
},
{
"epoch": 1.02,
"eval_accuracy": 0.43213572854291415,
"eval_loss": 1.0583962202072144,
"eval_runtime": 21.3265,
"eval_samples_per_second": 234.919,
"eval_steps_per_second": 29.4,
"step": 25000
},
{
"epoch": 1.06,
"learning_rate": 7.219444444444444e-07,
"loss": 1.0717,
"step": 26000
},
{
"epoch": 1.06,
"eval_accuracy": 0.4305389221556886,
"eval_loss": 1.0561293363571167,
"eval_runtime": 21.3034,
"eval_samples_per_second": 235.174,
"eval_steps_per_second": 29.432,
"step": 26000
},
{
"epoch": 1.1,
"learning_rate": 7.496944444444444e-07,
"loss": 1.0709,
"step": 27000
},
{
"epoch": 1.1,
"eval_accuracy": 0.4281437125748503,
"eval_loss": 1.0555357933044434,
"eval_runtime": 21.2178,
"eval_samples_per_second": 236.123,
"eval_steps_per_second": 29.551,
"step": 27000
},
{
"epoch": 1.14,
"learning_rate": 7.774722222222223e-07,
"loss": 1.0701,
"step": 28000
},
{
"epoch": 1.14,
"eval_accuracy": 0.4217564870259481,
"eval_loss": 1.054961085319519,
"eval_runtime": 21.1775,
"eval_samples_per_second": 236.571,
"eval_steps_per_second": 29.607,
"step": 28000
},
{
"epoch": 1.18,
"learning_rate": 8.052222222222223e-07,
"loss": 1.0641,
"step": 29000
},
{
"epoch": 1.18,
"eval_accuracy": 0.4291417165668663,
"eval_loss": 1.0518379211425781,
"eval_runtime": 21.0932,
"eval_samples_per_second": 237.517,
"eval_steps_per_second": 29.725,
"step": 29000
},
{
"epoch": 1.22,
"learning_rate": 8.330000000000001e-07,
"loss": 1.064,
"step": 30000
},
{
"epoch": 1.22,
"eval_accuracy": 0.43173652694610776,
"eval_loss": 1.0493717193603516,
"eval_runtime": 21.2843,
"eval_samples_per_second": 235.385,
"eval_steps_per_second": 29.458,
"step": 30000
},
{
"epoch": 1.26,
"learning_rate": 8.607500000000001e-07,
"loss": 1.0693,
"step": 31000
},
{
"epoch": 1.26,
"eval_accuracy": 0.4291417165668663,
"eval_loss": 1.0521764755249023,
"eval_runtime": 21.2278,
"eval_samples_per_second": 236.011,
"eval_steps_per_second": 29.537,
"step": 31000
},
{
"epoch": 1.3,
"learning_rate": 8.885277777777779e-07,
"loss": 1.0649,
"step": 32000
},
{
"epoch": 1.3,
"eval_accuracy": 0.42375249500998,
"eval_loss": 1.0528494119644165,
"eval_runtime": 21.1249,
"eval_samples_per_second": 237.161,
"eval_steps_per_second": 29.681,
"step": 32000
},
{
"epoch": 1.34,
"learning_rate": 9.163055555555556e-07,
"loss": 1.0619,
"step": 33000
},
{
"epoch": 1.34,
"eval_accuracy": 0.43293413173652695,
"eval_loss": 1.049193263053894,
"eval_runtime": 21.241,
"eval_samples_per_second": 235.865,
"eval_steps_per_second": 29.518,
"step": 33000
},
{
"epoch": 1.39,
"learning_rate": 9.440555555555557e-07,
"loss": 1.0582,
"step": 34000
},
{
"epoch": 1.39,
"eval_accuracy": 0.4357285429141717,
"eval_loss": 1.04512619972229,
"eval_runtime": 21.2531,
"eval_samples_per_second": 235.73,
"eval_steps_per_second": 29.502,
"step": 34000
},
{
"epoch": 1.43,
"learning_rate": 9.718333333333334e-07,
"loss": 1.0629,
"step": 35000
},
{
"epoch": 1.43,
"eval_accuracy": 0.43253493013972055,
"eval_loss": 1.043523907661438,
"eval_runtime": 21.2176,
"eval_samples_per_second": 236.124,
"eval_steps_per_second": 29.551,
"step": 35000
},
{
"epoch": 1.47,
"learning_rate": 9.995833333333334e-07,
"loss": 1.0588,
"step": 36000
},
{
"epoch": 1.47,
"eval_accuracy": 0.4307385229540918,
"eval_loss": 1.0413768291473389,
"eval_runtime": 21.2225,
"eval_samples_per_second": 236.07,
"eval_steps_per_second": 29.544,
"step": 36000
},
{
"epoch": 1.51,
"learning_rate": 1.0273611111111112e-06,
"loss": 1.0552,
"step": 37000
},
{
"epoch": 1.51,
"eval_accuracy": 0.4415169660678643,
"eval_loss": 1.0397700071334839,
"eval_runtime": 21.0651,
"eval_samples_per_second": 237.835,
"eval_steps_per_second": 29.765,
"step": 37000
},
{
"epoch": 1.55,
"learning_rate": 1.055138888888889e-06,
"loss": 1.0567,
"step": 38000
},
{
"epoch": 1.55,
"eval_accuracy": 0.4419161676646707,
"eval_loss": 1.0390877723693848,
"eval_runtime": 21.2858,
"eval_samples_per_second": 235.368,
"eval_steps_per_second": 29.456,
"step": 38000
},
{
"epoch": 1.59,
"learning_rate": 1.0829166666666667e-06,
"loss": 1.054,
"step": 39000
},
{
"epoch": 1.59,
"eval_accuracy": 0.43952095808383235,
"eval_loss": 1.0404103994369507,
"eval_runtime": 21.2544,
"eval_samples_per_second": 235.715,
"eval_steps_per_second": 29.5,
"step": 39000
},
{
"epoch": 1.63,
"learning_rate": 1.1106666666666668e-06,
"loss": 1.049,
"step": 40000
},
{
"epoch": 1.63,
"eval_accuracy": 0.4477045908183633,
"eval_loss": 1.0360453128814697,
"eval_runtime": 21.0993,
"eval_samples_per_second": 237.448,
"eval_steps_per_second": 29.717,
"step": 40000
},
{
"epoch": 1.67,
"learning_rate": 1.1384444444444446e-06,
"loss": 1.0522,
"step": 41000
},
{
"epoch": 1.67,
"eval_accuracy": 0.443313373253493,
"eval_loss": 1.0359567403793335,
"eval_runtime": 21.2828,
"eval_samples_per_second": 235.401,
"eval_steps_per_second": 29.46,
"step": 41000
},
{
"epoch": 1.71,
"learning_rate": 1.1661944444444447e-06,
"loss": 1.0459,
"step": 42000
},
{
"epoch": 1.71,
"eval_accuracy": 0.437125748502994,
"eval_loss": 1.0328505039215088,
"eval_runtime": 21.2626,
"eval_samples_per_second": 235.625,
"eval_steps_per_second": 29.488,
"step": 42000
},
{
"epoch": 1.75,
"learning_rate": 1.1939722222222222e-06,
"loss": 1.0488,
"step": 43000
},
{
"epoch": 1.75,
"eval_accuracy": 0.4379241516966068,
"eval_loss": 1.0460196733474731,
"eval_runtime": 21.2023,
"eval_samples_per_second": 236.295,
"eval_steps_per_second": 29.572,
"step": 43000
},
{
"epoch": 1.79,
"learning_rate": 1.2217222222222223e-06,
"loss": 1.0504,
"step": 44000
},
{
"epoch": 1.79,
"eval_accuracy": 0.4421157684630739,
"eval_loss": 1.0345144271850586,
"eval_runtime": 21.2684,
"eval_samples_per_second": 235.561,
"eval_steps_per_second": 29.48,
"step": 44000
},
{
"epoch": 1.83,
"learning_rate": 1.2494722222222224e-06,
"loss": 1.0482,
"step": 45000
},
{
"epoch": 1.83,
"eval_accuracy": 0.46467065868263474,
"eval_loss": 1.0280784368515015,
"eval_runtime": 21.1069,
"eval_samples_per_second": 237.363,
"eval_steps_per_second": 29.706,
"step": 45000
},
{
"epoch": 1.87,
"learning_rate": 1.2772500000000001e-06,
"loss": 1.046,
"step": 46000
},
{
"epoch": 1.87,
"eval_accuracy": 0.4554890219560878,
"eval_loss": 1.031961441040039,
"eval_runtime": 21.2098,
"eval_samples_per_second": 236.211,
"eval_steps_per_second": 29.562,
"step": 46000
},
{
"epoch": 1.91,
"learning_rate": 1.3050277777777777e-06,
"loss": 1.0391,
"step": 47000
},
{
"epoch": 1.91,
"eval_accuracy": 0.46706586826347307,
"eval_loss": 1.023941159248352,
"eval_runtime": 21.2269,
"eval_samples_per_second": 236.021,
"eval_steps_per_second": 29.538,
"step": 47000
},
{
"epoch": 1.96,
"learning_rate": 1.3328055555555555e-06,
"loss": 1.0444,
"step": 48000
},
{
"epoch": 1.96,
"eval_accuracy": 0.4491017964071856,
"eval_loss": 1.036636471748352,
"eval_runtime": 21.1179,
"eval_samples_per_second": 237.239,
"eval_steps_per_second": 29.69,
"step": 48000
},
{
"epoch": 2.0,
"learning_rate": 1.3605555555555555e-06,
"loss": 1.0386,
"step": 49000
},
{
"epoch": 2.0,
"eval_accuracy": 0.47305389221556887,
"eval_loss": 1.0199605226516724,
"eval_runtime": 21.2669,
"eval_samples_per_second": 235.577,
"eval_steps_per_second": 29.482,
"step": 49000
},
{
"epoch": 2.04,
"learning_rate": 1.3883333333333333e-06,
"loss": 1.0379,
"step": 50000
},
{
"epoch": 2.04,
"eval_accuracy": 0.4738522954091816,
"eval_loss": 1.0184080600738525,
"eval_runtime": 21.2806,
"eval_samples_per_second": 235.426,
"eval_steps_per_second": 29.463,
"step": 50000
},
{
"epoch": 2.08,
"learning_rate": 1.416111111111111e-06,
"loss": 1.0365,
"step": 51000
},
{
"epoch": 2.08,
"eval_accuracy": 0.474251497005988,
"eval_loss": 1.0143232345581055,
"eval_runtime": 21.1003,
"eval_samples_per_second": 237.437,
"eval_steps_per_second": 29.715,
"step": 51000
},
{
"epoch": 2.12,
"learning_rate": 1.4438888888888889e-06,
"loss": 1.0356,
"step": 52000
},
{
"epoch": 2.12,
"eval_accuracy": 0.4782435129740519,
"eval_loss": 1.0179694890975952,
"eval_runtime": 21.2847,
"eval_samples_per_second": 235.38,
"eval_steps_per_second": 29.458,
"step": 52000
},
{
"epoch": 2.16,
"learning_rate": 1.471611111111111e-06,
"loss": 1.0401,
"step": 53000
},
{
"epoch": 2.16,
"eval_accuracy": 0.4864271457085828,
"eval_loss": 1.008955955505371,
"eval_runtime": 21.0808,
"eval_samples_per_second": 237.657,
"eval_steps_per_second": 29.743,
"step": 53000
},
{
"epoch": 2.2,
"learning_rate": 1.4993888888888888e-06,
"loss": 1.0338,
"step": 54000
},
{
"epoch": 2.2,
"eval_accuracy": 0.4870259481037924,
"eval_loss": 1.0100921392440796,
"eval_runtime": 21.2581,
"eval_samples_per_second": 235.674,
"eval_steps_per_second": 29.495,
"step": 54000
},
{
"epoch": 2.24,
"learning_rate": 1.5271388888888889e-06,
"loss": 1.0322,
"step": 55000
},
{
"epoch": 2.24,
"eval_accuracy": 0.48403193612774453,
"eval_loss": 1.0082924365997314,
"eval_runtime": 21.2835,
"eval_samples_per_second": 235.394,
"eval_steps_per_second": 29.459,
"step": 55000
},
{
"epoch": 2.28,
"learning_rate": 1.5549166666666666e-06,
"loss": 1.0287,
"step": 56000
},
{
"epoch": 2.28,
"eval_accuracy": 0.47065868263473054,
"eval_loss": 1.0229328870773315,
"eval_runtime": 21.1356,
"eval_samples_per_second": 237.04,
"eval_steps_per_second": 29.666,
"step": 56000
},
{
"epoch": 2.32,
"learning_rate": 1.5826944444444446e-06,
"loss": 1.0291,
"step": 57000
},
{
"epoch": 2.32,
"eval_accuracy": 0.4844311377245509,
"eval_loss": 1.0074658393859863,
"eval_runtime": 21.2919,
"eval_samples_per_second": 235.301,
"eval_steps_per_second": 29.448,
"step": 57000
},
{
"epoch": 2.36,
"learning_rate": 1.6104722222222222e-06,
"loss": 1.0304,
"step": 58000
},
{
"epoch": 2.36,
"eval_accuracy": 0.4874251497005988,
"eval_loss": 1.007895588874817,
"eval_runtime": 21.1997,
"eval_samples_per_second": 236.324,
"eval_steps_per_second": 29.576,
"step": 58000
},
{
"epoch": 2.4,
"learning_rate": 1.6381944444444445e-06,
"loss": 1.0259,
"step": 59000
},
{
"epoch": 2.4,
"eval_accuracy": 0.492814371257485,
"eval_loss": 0.9993996620178223,
"eval_runtime": 21.0668,
"eval_samples_per_second": 237.815,
"eval_steps_per_second": 29.763,
"step": 59000
},
{
"epoch": 2.44,
"learning_rate": 1.6659722222222223e-06,
"loss": 1.0314,
"step": 60000
},
{
"epoch": 2.44,
"eval_accuracy": 0.48902195608782434,
"eval_loss": 1.0036838054656982,
"eval_runtime": 21.2359,
"eval_samples_per_second": 235.921,
"eval_steps_per_second": 29.525,
"step": 60000
},
{
"epoch": 2.49,
"learning_rate": 1.6937500000000003e-06,
"loss": 1.0276,
"step": 61000
},
{
"epoch": 2.49,
"eval_accuracy": 0.48502994011976047,
"eval_loss": 1.0065844058990479,
"eval_runtime": 21.07,
"eval_samples_per_second": 237.778,
"eval_steps_per_second": 29.758,
"step": 61000
},
{
"epoch": 2.53,
"learning_rate": 1.7215000000000002e-06,
"loss": 1.0298,
"step": 62000
},
{
"epoch": 2.53,
"eval_accuracy": 0.48263473053892214,
"eval_loss": 1.0085350275039673,
"eval_runtime": 21.4949,
"eval_samples_per_second": 233.079,
"eval_steps_per_second": 29.17,
"step": 62000
},
{
"epoch": 2.57,
"learning_rate": 1.7492777777777777e-06,
"loss": 1.0292,
"step": 63000
},
{
"epoch": 2.57,
"eval_accuracy": 0.49560878243512974,
"eval_loss": 0.9984678030014038,
"eval_runtime": 21.4688,
"eval_samples_per_second": 233.362,
"eval_steps_per_second": 29.205,
"step": 63000
},
{
"epoch": 2.61,
"learning_rate": 1.7770277777777778e-06,
"loss": 1.0273,
"step": 64000
},
{
"epoch": 2.61,
"eval_accuracy": 0.4942115768463074,
"eval_loss": 0.9960917830467224,
"eval_runtime": 21.307,
"eval_samples_per_second": 235.134,
"eval_steps_per_second": 29.427,
"step": 64000
},
{
"epoch": 2.65,
"learning_rate": 1.8048055555555558e-06,
"loss": 1.0256,
"step": 65000
},
{
"epoch": 2.65,
"eval_accuracy": 0.49840319361277446,
"eval_loss": 0.9994720816612244,
"eval_runtime": 21.5116,
"eval_samples_per_second": 232.897,
"eval_steps_per_second": 29.147,
"step": 65000
},
{
"epoch": 2.69,
"learning_rate": 1.8325833333333333e-06,
"loss": 1.0306,
"step": 66000
},
{
"epoch": 2.69,
"eval_accuracy": 0.49660678642714573,
"eval_loss": 0.9951530694961548,
"eval_runtime": 21.5301,
"eval_samples_per_second": 232.697,
"eval_steps_per_second": 29.122,
"step": 66000
},
{
"epoch": 2.73,
"learning_rate": 1.8603333333333334e-06,
"loss": 1.0245,
"step": 67000
},
{
"epoch": 2.73,
"eval_accuracy": 0.499001996007984,
"eval_loss": 0.9978940486907959,
"eval_runtime": 21.0729,
"eval_samples_per_second": 237.746,
"eval_steps_per_second": 29.754,
"step": 67000
},
{
"epoch": 2.77,
"learning_rate": 1.8881111111111114e-06,
"loss": 1.0247,
"step": 68000
},
{
"epoch": 2.77,
"eval_accuracy": 0.5,
"eval_loss": 0.9951012134552002,
"eval_runtime": 21.1793,
"eval_samples_per_second": 236.552,
"eval_steps_per_second": 29.604,
"step": 68000
},
{
"epoch": 2.81,
"learning_rate": 1.9158611111111115e-06,
"loss": 1.0241,
"step": 69000
},
{
"epoch": 2.81,
"eval_accuracy": 0.5021956087824351,
"eval_loss": 0.9969767332077026,
"eval_runtime": 21.2229,
"eval_samples_per_second": 236.066,
"eval_steps_per_second": 29.544,
"step": 69000
},
{
"epoch": 2.85,
"learning_rate": 1.943638888888889e-06,
"loss": 1.0242,
"step": 70000
},
{
"epoch": 2.85,
"eval_accuracy": 0.49720558882235527,
"eval_loss": 1.00163733959198,
"eval_runtime": 21.0442,
"eval_samples_per_second": 238.071,
"eval_steps_per_second": 29.794,
"step": 70000
},
{
"epoch": 2.89,
"learning_rate": 1.971416666666667e-06,
"loss": 1.017,
"step": 71000
},
{
"epoch": 2.89,
"eval_accuracy": 0.49600798403193613,
"eval_loss": 1.0031572580337524,
"eval_runtime": 21.1987,
"eval_samples_per_second": 236.336,
"eval_steps_per_second": 29.577,
"step": 71000
},
{
"epoch": 2.93,
"learning_rate": 1.999166666666667e-06,
"loss": 1.0237,
"step": 72000
},
{
"epoch": 2.93,
"eval_accuracy": 0.49001996007984033,
"eval_loss": 1.0067389011383057,
"eval_runtime": 21.2825,
"eval_samples_per_second": 235.405,
"eval_steps_per_second": 29.461,
"step": 72000
},
{
"epoch": 2.97,
"learning_rate": 2.0269444444444444e-06,
"loss": 1.0208,
"step": 73000
},
{
"epoch": 2.97,
"eval_accuracy": 0.49560878243512974,
"eval_loss": 1.0011804103851318,
"eval_runtime": 21.137,
"eval_samples_per_second": 237.025,
"eval_steps_per_second": 29.664,
"step": 73000
},
{
"epoch": 3.01,
"learning_rate": 2.0546944444444447e-06,
"loss": 1.021,
"step": 74000
},
{
"epoch": 3.01,
"eval_accuracy": 0.49580838323353293,
"eval_loss": 0.9941307902336121,
"eval_runtime": 21.161,
"eval_samples_per_second": 236.757,
"eval_steps_per_second": 29.63,
"step": 74000
},
{
"epoch": 3.06,
"learning_rate": 2.0824722222222223e-06,
"loss": 1.0154,
"step": 75000
},
{
"epoch": 3.06,
"eval_accuracy": 0.49580838323353293,
"eval_loss": 0.9971462488174438,
"eval_runtime": 21.0263,
"eval_samples_per_second": 238.273,
"eval_steps_per_second": 29.82,
"step": 75000
},
{
"epoch": 3.1,
"learning_rate": 2.1102222222222226e-06,
"loss": 1.0173,
"step": 76000
},
{
"epoch": 3.1,
"eval_accuracy": 0.5011976047904192,
"eval_loss": 0.9884433150291443,
"eval_runtime": 21.2199,
"eval_samples_per_second": 236.099,
"eval_steps_per_second": 29.548,
"step": 76000
},
{
"epoch": 3.14,
"learning_rate": 2.138e-06,
"loss": 1.0181,
"step": 77000
},
{
"epoch": 3.14,
"eval_accuracy": 0.4808383233532934,
"eval_loss": 1.0162075757980347,
"eval_runtime": 21.2108,
"eval_samples_per_second": 236.201,
"eval_steps_per_second": 29.56,
"step": 77000
},
{
"epoch": 3.18,
"learning_rate": 2.16575e-06,
"loss": 1.0112,
"step": 78000
},
{
"epoch": 3.18,
"eval_accuracy": 0.49880239520958086,
"eval_loss": 0.9902569651603699,
"eval_runtime": 21.032,
"eval_samples_per_second": 238.209,
"eval_steps_per_second": 29.812,
"step": 78000
},
{
"epoch": 3.22,
"learning_rate": 2.193527777777778e-06,
"loss": 1.0166,
"step": 79000
},
{
"epoch": 3.22,
"eval_accuracy": 0.49500998003992014,
"eval_loss": 1.0056451559066772,
"eval_runtime": 21.1866,
"eval_samples_per_second": 236.47,
"eval_steps_per_second": 29.594,
"step": 79000
},
{
"epoch": 3.26,
"learning_rate": 2.221277777777778e-06,
"loss": 1.0135,
"step": 80000
},
{
"epoch": 3.26,
"eval_accuracy": 0.49161676646706587,
"eval_loss": 1.004488229751587,
"eval_runtime": 21.2075,
"eval_samples_per_second": 236.237,
"eval_steps_per_second": 29.565,
"step": 80000
},
{
"epoch": 3.3,
"learning_rate": 2.249055555555556e-06,
"loss": 1.0147,
"step": 81000
},
{
"epoch": 3.3,
"eval_accuracy": 0.49101796407185627,
"eval_loss": 1.0022324323654175,
"eval_runtime": 21.1622,
"eval_samples_per_second": 236.743,
"eval_steps_per_second": 29.628,
"step": 81000
},
{
"epoch": 3.34,
"learning_rate": 2.2768055555555557e-06,
"loss": 1.0249,
"step": 82000
},
{
"epoch": 3.34,
"eval_accuracy": 0.49540918163672654,
"eval_loss": 0.9961836338043213,
"eval_runtime": 21.2661,
"eval_samples_per_second": 235.586,
"eval_steps_per_second": 29.484,
"step": 82000
},
{
"epoch": 3.38,
"learning_rate": 2.3045833333333336e-06,
"loss": 1.012,
"step": 83000
},
{
"epoch": 3.38,
"eval_accuracy": 0.49520958083832334,
"eval_loss": 0.997968316078186,
"eval_runtime": 21.1127,
"eval_samples_per_second": 237.298,
"eval_steps_per_second": 29.698,
"step": 83000
},
{
"epoch": 3.42,
"learning_rate": 2.3323333333333335e-06,
"loss": 1.0153,
"step": 84000
},
{
"epoch": 3.42,
"eval_accuracy": 0.5063872255489021,
"eval_loss": 0.9885319471359253,
"eval_runtime": 21.2015,
"eval_samples_per_second": 236.304,
"eval_steps_per_second": 29.573,
"step": 84000
},
{
"epoch": 3.46,
"learning_rate": 2.360111111111111e-06,
"loss": 1.0113,
"step": 85000
},
{
"epoch": 3.46,
"eval_accuracy": 0.49860279441117766,
"eval_loss": 0.9927662014961243,
"eval_runtime": 21.2447,
"eval_samples_per_second": 235.823,
"eval_steps_per_second": 29.513,
"step": 85000
},
{
"epoch": 3.5,
"learning_rate": 2.3878611111111113e-06,
"loss": 1.0222,
"step": 86000
},
{
"epoch": 3.5,
"eval_accuracy": 0.5073852295409181,
"eval_loss": 0.9855450987815857,
"eval_runtime": 21.03,
"eval_samples_per_second": 238.231,
"eval_steps_per_second": 29.815,
"step": 86000
},
{
"epoch": 3.54,
"learning_rate": 2.4156388888888893e-06,
"loss": 1.0062,
"step": 87000
},
{
"epoch": 3.54,
"eval_accuracy": 0.5075848303393213,
"eval_loss": 0.9856505990028381,
"eval_runtime": 21.1796,
"eval_samples_per_second": 236.548,
"eval_steps_per_second": 29.604,
"step": 87000
},
{
"epoch": 3.59,
"learning_rate": 2.443388888888889e-06,
"loss": 1.0157,
"step": 88000
},
{
"epoch": 3.59,
"eval_accuracy": 0.5115768463073852,
"eval_loss": 0.9844857454299927,
"eval_runtime": 21.2024,
"eval_samples_per_second": 236.294,
"eval_steps_per_second": 29.572,
"step": 88000
},
{
"epoch": 3.63,
"learning_rate": 2.4711666666666668e-06,
"loss": 1.0069,
"step": 89000
},
{
"epoch": 3.63,
"eval_accuracy": 0.499001996007984,
"eval_loss": 0.9894497394561768,
"eval_runtime": 21.2532,
"eval_samples_per_second": 235.729,
"eval_steps_per_second": 29.501,
"step": 89000
},
{
"epoch": 3.67,
"learning_rate": 2.4989166666666666e-06,
"loss": 1.0164,
"step": 90000
},
{
"epoch": 3.67,
"eval_accuracy": 0.5101796407185629,
"eval_loss": 0.9842925667762756,
"eval_runtime": 21.1914,
"eval_samples_per_second": 236.416,
"eval_steps_per_second": 29.587,
"step": 90000
},
{
"epoch": 3.71,
"learning_rate": 2.5266944444444446e-06,
"loss": 1.0175,
"step": 91000
},
{
"epoch": 3.71,
"eval_accuracy": 0.5055888223552895,
"eval_loss": 0.9913986325263977,
"eval_runtime": 21.0589,
"eval_samples_per_second": 237.904,
"eval_steps_per_second": 29.774,
"step": 91000
},
{
"epoch": 3.75,
"learning_rate": 2.554444444444445e-06,
"loss": 1.013,
"step": 92000
},
{
"epoch": 3.75,
"eval_accuracy": 0.5057884231536927,
"eval_loss": 0.9887382388114929,
"eval_runtime": 21.2965,
"eval_samples_per_second": 235.25,
"eval_steps_per_second": 29.441,
"step": 92000
},
{
"epoch": 3.79,
"learning_rate": 2.5822222222222224e-06,
"loss": 1.0157,
"step": 93000
},
{
"epoch": 3.79,
"eval_accuracy": 0.5127744510978044,
"eval_loss": 0.9854277968406677,
"eval_runtime": 21.2839,
"eval_samples_per_second": 235.389,
"eval_steps_per_second": 29.459,
"step": 93000
},
{
"epoch": 3.83,
"learning_rate": 2.6099722222222223e-06,
"loss": 1.0108,
"step": 94000
},
{
"epoch": 3.83,
"eval_accuracy": 0.5039920159680639,
"eval_loss": 0.9907957315444946,
"eval_runtime": 21.0477,
"eval_samples_per_second": 238.03,
"eval_steps_per_second": 29.789,
"step": 94000
},
{
"epoch": 3.87,
"learning_rate": 2.6377500000000003e-06,
"loss": 1.0132,
"step": 95000
},
{
"epoch": 3.87,
"eval_accuracy": 0.5161676646706587,
"eval_loss": 0.9830310940742493,
"eval_runtime": 21.2625,
"eval_samples_per_second": 235.626,
"eval_steps_per_second": 29.489,
"step": 95000
},
{
"epoch": 3.91,
"learning_rate": 2.6655e-06,
"loss": 1.0139,
"step": 96000
},
{
"epoch": 3.91,
"eval_accuracy": 0.5125748502994012,
"eval_loss": 0.9806137681007385,
"eval_runtime": 21.1848,
"eval_samples_per_second": 236.49,
"eval_steps_per_second": 29.597,
"step": 96000
},
{
"epoch": 3.95,
"learning_rate": 2.6932777777777777e-06,
"loss": 1.0104,
"step": 97000
},
{
"epoch": 3.95,
"eval_accuracy": 0.49640718562874253,
"eval_loss": 0.9965940117835999,
"eval_runtime": 21.1145,
"eval_samples_per_second": 237.278,
"eval_steps_per_second": 29.695,
"step": 97000
},
{
"epoch": 3.99,
"learning_rate": 2.721027777777778e-06,
"loss": 1.011,
"step": 98000
},
{
"epoch": 3.99,
"eval_accuracy": 0.49800399201596807,
"eval_loss": 0.9846762418746948,
"eval_runtime": 21.2444,
"eval_samples_per_second": 235.827,
"eval_steps_per_second": 29.514,
"step": 98000
},
{
"epoch": 4.03,
"learning_rate": 2.748805555555556e-06,
"loss": 1.0063,
"step": 99000
},
{
"epoch": 4.03,
"eval_accuracy": 0.5103792415169661,
"eval_loss": 0.9853057265281677,
"eval_runtime": 21.0515,
"eval_samples_per_second": 237.988,
"eval_steps_per_second": 29.784,
"step": 99000
},
{
"epoch": 4.07,
"learning_rate": 2.776555555555556e-06,
"loss": 1.007,
"step": 100000
},
{
"epoch": 4.07,
"eval_accuracy": 0.5115768463073852,
"eval_loss": 0.983065128326416,
"eval_runtime": 21.2494,
"eval_samples_per_second": 235.771,
"eval_steps_per_second": 29.507,
"step": 100000
},
{
"epoch": 4.12,
"learning_rate": 2.8043333333333334e-06,
"loss": 1.0107,
"step": 101000
},
{
"epoch": 4.12,
"eval_accuracy": 0.5129740518962076,
"eval_loss": 0.9863881468772888,
"eval_runtime": 21.2658,
"eval_samples_per_second": 235.59,
"eval_steps_per_second": 29.484,
"step": 101000
},
{
"epoch": 4.16,
"learning_rate": 2.8320833333333333e-06,
"loss": 1.0055,
"step": 102000
},
{
"epoch": 4.16,
"eval_accuracy": 0.49840319361277446,
"eval_loss": 0.9891018867492676,
"eval_runtime": 21.0702,
"eval_samples_per_second": 237.777,
"eval_steps_per_second": 29.758,
"step": 102000
},
{
"epoch": 4.2,
"learning_rate": 2.8598611111111112e-06,
"loss": 1.0115,
"step": 103000
},
{
"epoch": 4.2,
"eval_accuracy": 0.5115768463073852,
"eval_loss": 0.9800674319267273,
"eval_runtime": 21.2539,
"eval_samples_per_second": 235.722,
"eval_steps_per_second": 29.5,
"step": 103000
},
{
"epoch": 4.24,
"learning_rate": 2.8876111111111115e-06,
"loss": 1.0039,
"step": 104000
},
{
"epoch": 4.24,
"eval_accuracy": 0.49500998003992014,
"eval_loss": 0.9979402422904968,
"eval_runtime": 21.1981,
"eval_samples_per_second": 236.342,
"eval_steps_per_second": 29.578,
"step": 104000
},
{
"epoch": 4.28,
"learning_rate": 2.915388888888889e-06,
"loss": 1.0028,
"step": 105000
},
{
"epoch": 4.28,
"eval_accuracy": 0.492814371257485,
"eval_loss": 1.007880687713623,
"eval_runtime": 21.1024,
"eval_samples_per_second": 237.413,
"eval_steps_per_second": 29.712,
"step": 105000
}
],
"logging_steps": 1000,
"max_steps": 10000000,
"num_train_epochs": 408,
"save_steps": 1000,
"total_flos": 7.31660447687639e+16,
"trial_name": null,
"trial_params": null
}