albert-base-v2_mnli_bc / trainer_state.json
mjtwins's picture
commit from mujeen
f7c73a6
{
"best_metric": 0.9400777047220562,
"best_model_checkpoint": "/hdd1/mujeen/retrieval_prf/output/albert-base-v2_mnli_bc/checkpoint-49089",
"epoch": 3.0,
"global_step": 49089,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.03,
"learning_rate": 1.979832549043574e-05,
"loss": 0.4104,
"step": 500
},
{
"epoch": 0.06,
"learning_rate": 1.9595021287864898e-05,
"loss": 0.3325,
"step": 1000
},
{
"epoch": 0.09,
"learning_rate": 1.939171708529406e-05,
"loss": 0.3067,
"step": 1500
},
{
"epoch": 0.12,
"learning_rate": 1.9188005459471572e-05,
"loss": 0.3078,
"step": 2000
},
{
"epoch": 0.15,
"learning_rate": 1.898429383364909e-05,
"loss": 0.2841,
"step": 2500
},
{
"epoch": 0.18,
"learning_rate": 1.8780582207826603e-05,
"loss": 0.2747,
"step": 3000
},
{
"epoch": 0.21,
"learning_rate": 1.8576870582004117e-05,
"loss": 0.2749,
"step": 3500
},
{
"epoch": 0.24,
"learning_rate": 1.837315895618163e-05,
"loss": 0.2612,
"step": 4000
},
{
"epoch": 0.28,
"learning_rate": 1.8169447330359144e-05,
"loss": 0.2629,
"step": 4500
},
{
"epoch": 0.31,
"learning_rate": 1.796573570453666e-05,
"loss": 0.2586,
"step": 5000
},
{
"epoch": 0.34,
"learning_rate": 1.776202407871417e-05,
"loss": 0.2599,
"step": 5500
},
{
"epoch": 0.37,
"learning_rate": 1.755831245289169e-05,
"loss": 0.2579,
"step": 6000
},
{
"epoch": 0.4,
"learning_rate": 1.7354600827069202e-05,
"loss": 0.2479,
"step": 6500
},
{
"epoch": 0.43,
"learning_rate": 1.7150889201246716e-05,
"loss": 0.2444,
"step": 7000
},
{
"epoch": 0.46,
"learning_rate": 1.6947177575424233e-05,
"loss": 0.2419,
"step": 7500
},
{
"epoch": 0.49,
"learning_rate": 1.6743465949601747e-05,
"loss": 0.2452,
"step": 8000
},
{
"epoch": 0.52,
"learning_rate": 1.653975432377926e-05,
"loss": 0.2493,
"step": 8500
},
{
"epoch": 0.55,
"learning_rate": 1.6336042697956774e-05,
"loss": 0.2343,
"step": 9000
},
{
"epoch": 0.58,
"learning_rate": 1.6132738495385935e-05,
"loss": 0.2464,
"step": 9500
},
{
"epoch": 0.61,
"learning_rate": 1.5929026869563445e-05,
"loss": 0.2305,
"step": 10000
},
{
"epoch": 0.64,
"learning_rate": 1.5725315243740962e-05,
"loss": 0.231,
"step": 10500
},
{
"epoch": 0.67,
"learning_rate": 1.5521603617918476e-05,
"loss": 0.2339,
"step": 11000
},
{
"epoch": 0.7,
"learning_rate": 1.5318299415347636e-05,
"loss": 0.2267,
"step": 11500
},
{
"epoch": 0.73,
"learning_rate": 1.511458778952515e-05,
"loss": 0.2249,
"step": 12000
},
{
"epoch": 0.76,
"learning_rate": 1.4910876163702664e-05,
"loss": 0.233,
"step": 12500
},
{
"epoch": 0.79,
"learning_rate": 1.4707164537880179e-05,
"loss": 0.2257,
"step": 13000
},
{
"epoch": 0.83,
"learning_rate": 1.4503452912057691e-05,
"loss": 0.2277,
"step": 13500
},
{
"epoch": 0.86,
"learning_rate": 1.4299741286235206e-05,
"loss": 0.2225,
"step": 14000
},
{
"epoch": 0.89,
"learning_rate": 1.409602966041272e-05,
"loss": 0.226,
"step": 14500
},
{
"epoch": 0.92,
"learning_rate": 1.3892318034590235e-05,
"loss": 0.2173,
"step": 15000
},
{
"epoch": 0.95,
"learning_rate": 1.3689013832019396e-05,
"loss": 0.2191,
"step": 15500
},
{
"epoch": 0.98,
"learning_rate": 1.3485302206196908e-05,
"loss": 0.2159,
"step": 16000
},
{
"epoch": 1.0,
"eval_accuracy": 0.9248356246264196,
"eval_loss": 0.22675447165966034,
"eval_runtime": 19.8714,
"eval_samples_per_second": 336.766,
"eval_steps_per_second": 42.121,
"step": 16363
},
{
"epoch": 1.01,
"learning_rate": 1.3281590580374423e-05,
"loss": 0.2072,
"step": 16500
},
{
"epoch": 1.04,
"learning_rate": 1.3077878954551937e-05,
"loss": 0.1826,
"step": 17000
},
{
"epoch": 1.07,
"learning_rate": 1.2874574751981098e-05,
"loss": 0.1817,
"step": 17500
},
{
"epoch": 1.1,
"learning_rate": 1.267086312615861e-05,
"loss": 0.1745,
"step": 18000
},
{
"epoch": 1.13,
"learning_rate": 1.2467151500336125e-05,
"loss": 0.1841,
"step": 18500
},
{
"epoch": 1.16,
"learning_rate": 1.2263439874513639e-05,
"loss": 0.1856,
"step": 19000
},
{
"epoch": 1.19,
"learning_rate": 1.2059728248691154e-05,
"loss": 0.19,
"step": 19500
},
{
"epoch": 1.22,
"learning_rate": 1.185601662286867e-05,
"loss": 0.2017,
"step": 20000
},
{
"epoch": 1.25,
"learning_rate": 1.1652304997046182e-05,
"loss": 0.1885,
"step": 20500
},
{
"epoch": 1.28,
"learning_rate": 1.1449000794475342e-05,
"loss": 0.1732,
"step": 21000
},
{
"epoch": 1.31,
"learning_rate": 1.12456965919045e-05,
"loss": 0.1852,
"step": 21500
},
{
"epoch": 1.34,
"learning_rate": 1.104239238933366e-05,
"loss": 0.1899,
"step": 22000
},
{
"epoch": 1.38,
"learning_rate": 1.0838680763511174e-05,
"loss": 0.172,
"step": 22500
},
{
"epoch": 1.41,
"learning_rate": 1.063496913768869e-05,
"loss": 0.1869,
"step": 23000
},
{
"epoch": 1.44,
"learning_rate": 1.0431257511866205e-05,
"loss": 0.1956,
"step": 23500
},
{
"epoch": 1.47,
"learning_rate": 1.0227545886043717e-05,
"loss": 0.1925,
"step": 24000
},
{
"epoch": 1.5,
"learning_rate": 1.0023834260221232e-05,
"loss": 0.1867,
"step": 24500
},
{
"epoch": 1.53,
"learning_rate": 9.820122634398746e-06,
"loss": 0.1829,
"step": 25000
},
{
"epoch": 1.56,
"learning_rate": 9.616411008576261e-06,
"loss": 0.1764,
"step": 25500
},
{
"epoch": 1.59,
"learning_rate": 9.412699382753775e-06,
"loss": 0.1717,
"step": 26000
},
{
"epoch": 1.62,
"learning_rate": 9.208987756931288e-06,
"loss": 0.1767,
"step": 26500
},
{
"epoch": 1.65,
"learning_rate": 9.005276131108804e-06,
"loss": 0.1723,
"step": 27000
},
{
"epoch": 1.68,
"learning_rate": 8.801971928537963e-06,
"loss": 0.1753,
"step": 27500
},
{
"epoch": 1.71,
"learning_rate": 8.598667725967122e-06,
"loss": 0.1865,
"step": 28000
},
{
"epoch": 1.74,
"learning_rate": 8.394956100144636e-06,
"loss": 0.1708,
"step": 28500
},
{
"epoch": 1.77,
"learning_rate": 8.19124447432215e-06,
"loss": 0.1731,
"step": 29000
},
{
"epoch": 1.8,
"learning_rate": 7.987532848499665e-06,
"loss": 0.1667,
"step": 29500
},
{
"epoch": 1.83,
"learning_rate": 7.78382122267718e-06,
"loss": 0.1852,
"step": 30000
},
{
"epoch": 1.86,
"learning_rate": 7.580109596854694e-06,
"loss": 0.1737,
"step": 30500
},
{
"epoch": 1.89,
"learning_rate": 7.376397971032207e-06,
"loss": 0.1806,
"step": 31000
},
{
"epoch": 1.93,
"learning_rate": 7.172686345209722e-06,
"loss": 0.1698,
"step": 31500
},
{
"epoch": 1.96,
"learning_rate": 6.9689747193872355e-06,
"loss": 0.175,
"step": 32000
},
{
"epoch": 1.99,
"learning_rate": 6.76526309356475e-06,
"loss": 0.1817,
"step": 32500
},
{
"epoch": 2.0,
"eval_accuracy": 0.9346981470412433,
"eval_loss": 0.2335003763437271,
"eval_runtime": 19.6848,
"eval_samples_per_second": 339.958,
"eval_steps_per_second": 42.52,
"step": 32726
},
{
"epoch": 2.02,
"learning_rate": 6.5615514677422645e-06,
"loss": 0.1377,
"step": 33000
},
{
"epoch": 2.05,
"learning_rate": 6.358247265171424e-06,
"loss": 0.0941,
"step": 33500
},
{
"epoch": 2.08,
"learning_rate": 6.154535639348939e-06,
"loss": 0.1108,
"step": 34000
},
{
"epoch": 2.11,
"learning_rate": 5.950824013526453e-06,
"loss": 0.1082,
"step": 34500
},
{
"epoch": 2.14,
"learning_rate": 5.747112387703967e-06,
"loss": 0.1117,
"step": 35000
},
{
"epoch": 2.17,
"learning_rate": 5.543400761881481e-06,
"loss": 0.1155,
"step": 35500
},
{
"epoch": 2.2,
"learning_rate": 5.339689136058995e-06,
"loss": 0.1013,
"step": 36000
},
{
"epoch": 2.23,
"learning_rate": 5.13597751023651e-06,
"loss": 0.1156,
"step": 36500
},
{
"epoch": 2.26,
"learning_rate": 4.9322658844140235e-06,
"loss": 0.1014,
"step": 37000
},
{
"epoch": 2.29,
"learning_rate": 4.729369105094828e-06,
"loss": 0.1124,
"step": 37500
},
{
"epoch": 2.32,
"learning_rate": 4.5256574792723424e-06,
"loss": 0.1178,
"step": 38000
},
{
"epoch": 2.35,
"learning_rate": 4.321945853449856e-06,
"loss": 0.1088,
"step": 38500
},
{
"epoch": 2.38,
"learning_rate": 4.118641650879016e-06,
"loss": 0.1216,
"step": 39000
},
{
"epoch": 2.41,
"learning_rate": 3.9149300250565305e-06,
"loss": 0.1087,
"step": 39500
},
{
"epoch": 2.44,
"learning_rate": 3.7112183992340446e-06,
"loss": 0.1094,
"step": 40000
},
{
"epoch": 2.48,
"learning_rate": 3.5075067734115587e-06,
"loss": 0.1137,
"step": 40500
},
{
"epoch": 2.51,
"learning_rate": 3.303795147589073e-06,
"loss": 0.1161,
"step": 41000
},
{
"epoch": 2.54,
"learning_rate": 3.1000835217665877e-06,
"loss": 0.1055,
"step": 41500
},
{
"epoch": 2.57,
"learning_rate": 2.896371895944102e-06,
"loss": 0.1058,
"step": 42000
},
{
"epoch": 2.6,
"learning_rate": 2.692660270121616e-06,
"loss": 0.1037,
"step": 42500
},
{
"epoch": 2.63,
"learning_rate": 2.4889486442991304e-06,
"loss": 0.1023,
"step": 43000
},
{
"epoch": 2.66,
"learning_rate": 2.2852370184766445e-06,
"loss": 0.0998,
"step": 43500
},
{
"epoch": 2.69,
"learning_rate": 2.081932815905804e-06,
"loss": 0.1097,
"step": 44000
},
{
"epoch": 2.72,
"learning_rate": 1.8782211900833183e-06,
"loss": 0.0914,
"step": 44500
},
{
"epoch": 2.75,
"learning_rate": 1.6745095642608326e-06,
"loss": 0.1139,
"step": 45000
},
{
"epoch": 2.78,
"learning_rate": 1.4707979384383466e-06,
"loss": 0.0998,
"step": 45500
},
{
"epoch": 2.81,
"learning_rate": 1.2670863126158612e-06,
"loss": 0.0959,
"step": 46000
},
{
"epoch": 2.84,
"learning_rate": 1.0633746867933755e-06,
"loss": 0.0994,
"step": 46500
},
{
"epoch": 2.87,
"learning_rate": 8.596630609708896e-07,
"loss": 0.1093,
"step": 47000
},
{
"epoch": 2.9,
"learning_rate": 6.559514351484039e-07,
"loss": 0.0967,
"step": 47500
},
{
"epoch": 2.93,
"learning_rate": 4.5264723257756327e-07,
"loss": 0.102,
"step": 48000
},
{
"epoch": 2.96,
"learning_rate": 2.4893560675507757e-07,
"loss": 0.0966,
"step": 48500
},
{
"epoch": 2.99,
"learning_rate": 4.522398093259183e-08,
"loss": 0.0863,
"step": 49000
},
{
"epoch": 3.0,
"eval_accuracy": 0.9400777047220562,
"eval_loss": 0.3014400601387024,
"eval_runtime": 20.103,
"eval_samples_per_second": 332.885,
"eval_steps_per_second": 41.636,
"step": 49089
},
{
"epoch": 3.0,
"step": 49089,
"total_flos": 3063521562716160.0,
"train_loss": 0.18007777646369438,
"train_runtime": 3964.6017,
"train_samples_per_second": 198.105,
"train_steps_per_second": 12.382
}
],
"max_steps": 49089,
"num_train_epochs": 3,
"total_flos": 3063521562716160.0,
"trial_name": null,
"trial_params": null
}