roberta-large-defteval-t6-st2 / trainer_state.json
LL
update
0af2060
{
"best_metric": 43.86838413539317,
"best_model_checkpoint": "results/subtask2-roberta-large-3e-5-32-10-evallabels/checkpoint-950",
"epoch": 10.0,
"global_step": 1890,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.26,
"learning_rate": 7.936507936507936e-06,
"loss": 1.4415,
"step": 50
},
{
"epoch": 0.26,
"eval_accuracy_score": 76.77304173175631,
"eval_f1": 0.29239766081871343,
"eval_loss": 0.6649255752563477,
"eval_precision": 1.2345679012345678,
"eval_recall": 1.355421686746988,
"eval_runtime": 8.6446,
"eval_samples_per_second": 136.386,
"step": 50
},
{
"epoch": 0.53,
"learning_rate": 1.5873015873015872e-05,
"loss": 0.6015,
"step": 100
},
{
"epoch": 0.53,
"eval_accuracy_score": 85.00781075652756,
"eval_f1": 7.7752154879329645,
"eval_loss": 0.44424399733543396,
"eval_precision": 23.591549295774648,
"eval_recall": 20.180722891566266,
"eval_runtime": 8.8048,
"eval_samples_per_second": 133.905,
"step": 100
},
{
"epoch": 0.79,
"learning_rate": 2.380952380952381e-05,
"loss": 0.4787,
"step": 150
},
{
"epoch": 0.79,
"eval_accuracy_score": 84.93639812541844,
"eval_f1": 13.33891292959303,
"eval_loss": 0.42927324771881104,
"eval_precision": 30.36036036036036,
"eval_recall": 50.75301204819277,
"eval_runtime": 8.7806,
"eval_samples_per_second": 134.274,
"step": 150
},
{
"epoch": 1.06,
"learning_rate": 2.980599647266314e-05,
"loss": 0.4004,
"step": 200
},
{
"epoch": 1.06,
"eval_accuracy_score": 86.96273153313993,
"eval_f1": 22.702533225449525,
"eval_loss": 0.3721863925457001,
"eval_precision": 43.04812834224599,
"eval_recall": 48.493975903614455,
"eval_runtime": 8.7281,
"eval_samples_per_second": 135.081,
"step": 200
},
{
"epoch": 1.32,
"learning_rate": 2.8924162257495593e-05,
"loss": 0.3509,
"step": 250
},
{
"epoch": 1.32,
"eval_accuracy_score": 85.72193706761884,
"eval_f1": 30.029422978055926,
"eval_loss": 0.3999338746070862,
"eval_precision": 38.4,
"eval_recall": 57.831325301204814,
"eval_runtime": 8.8031,
"eval_samples_per_second": 133.929,
"step": 250
},
{
"epoch": 1.59,
"learning_rate": 2.8042328042328043e-05,
"loss": 0.3751,
"step": 300
},
{
"epoch": 1.59,
"eval_accuracy_score": 87.88216915866994,
"eval_f1": 36.06508408748763,
"eval_loss": 0.33605942130088806,
"eval_precision": 48.492791612057665,
"eval_recall": 55.72289156626506,
"eval_runtime": 8.8234,
"eval_samples_per_second": 133.622,
"step": 300
},
{
"epoch": 1.85,
"learning_rate": 2.7160493827160493e-05,
"loss": 0.32,
"step": 350
},
{
"epoch": 1.85,
"eval_accuracy_score": 87.23499218924347,
"eval_f1": 38.96036192862586,
"eval_loss": 0.3504337668418884,
"eval_precision": 45.92178770949721,
"eval_recall": 61.897590361445786,
"eval_runtime": 8.8119,
"eval_samples_per_second": 133.797,
"step": 350
},
{
"epoch": 2.12,
"learning_rate": 2.6278659611992947e-05,
"loss": 0.338,
"step": 400
},
{
"epoch": 2.12,
"eval_accuracy_score": 87.51617942423566,
"eval_f1": 37.19003385850202,
"eval_loss": 0.3751635253429413,
"eval_precision": 47.30473047304731,
"eval_recall": 64.7590361445783,
"eval_runtime": 8.8298,
"eval_samples_per_second": 133.526,
"step": 400
},
{
"epoch": 2.38,
"learning_rate": 2.5396825396825397e-05,
"loss": 0.2625,
"step": 450
},
{
"epoch": 2.38,
"eval_accuracy_score": 87.68578442311984,
"eval_f1": 33.11482983162181,
"eval_loss": 0.3540908396244049,
"eval_precision": 51.25944584382871,
"eval_recall": 61.295180722891565,
"eval_runtime": 8.6563,
"eval_samples_per_second": 136.201,
"step": 450
},
{
"epoch": 2.65,
"learning_rate": 2.4514991181657847e-05,
"loss": 0.2718,
"step": 500
},
{
"epoch": 2.65,
"eval_accuracy_score": 88.29279178754742,
"eval_f1": 32.55245599678572,
"eval_loss": 0.3684297502040863,
"eval_precision": 51.84243964421855,
"eval_recall": 61.44578313253012,
"eval_runtime": 8.9791,
"eval_samples_per_second": 131.305,
"step": 500
},
{
"epoch": 2.91,
"learning_rate": 2.36331569664903e-05,
"loss": 0.2708,
"step": 550
},
{
"epoch": 2.91,
"eval_accuracy_score": 86.14594956482928,
"eval_f1": 38.566294416176056,
"eval_loss": 0.3753862679004669,
"eval_precision": 51.88571428571429,
"eval_recall": 68.37349397590361,
"eval_runtime": 8.8041,
"eval_samples_per_second": 133.916,
"step": 550
},
{
"epoch": 3.17,
"learning_rate": 2.275132275132275e-05,
"loss": 0.2305,
"step": 600
},
{
"epoch": 3.17,
"eval_accuracy_score": 88.5918321803169,
"eval_f1": 39.92203233980172,
"eval_loss": 0.3945120573043823,
"eval_precision": 52.08845208845209,
"eval_recall": 63.85542168674698,
"eval_runtime": 8.8109,
"eval_samples_per_second": 133.812,
"step": 600
},
{
"epoch": 3.44,
"learning_rate": 2.1869488536155205e-05,
"loss": 0.2122,
"step": 650
},
{
"epoch": 3.44,
"eval_accuracy_score": 88.7034144164249,
"eval_f1": 37.88744391080894,
"eval_loss": 0.35788846015930176,
"eval_precision": 53.2319391634981,
"eval_recall": 63.25301204819277,
"eval_runtime": 8.7495,
"eval_samples_per_second": 134.751,
"step": 650
},
{
"epoch": 3.7,
"learning_rate": 2.0987654320987655e-05,
"loss": 0.2126,
"step": 700
},
{
"epoch": 3.7,
"eval_accuracy_score": 87.35103771479581,
"eval_f1": 38.094203066236034,
"eval_loss": 0.37450292706489563,
"eval_precision": 52.6071842410197,
"eval_recall": 68.37349397590361,
"eval_runtime": 8.8823,
"eval_samples_per_second": 132.736,
"step": 700
},
{
"epoch": 3.97,
"learning_rate": 2.0105820105820105e-05,
"loss": 0.211,
"step": 750
},
{
"epoch": 3.97,
"eval_accuracy_score": 88.2883284981031,
"eval_f1": 38.51324803050831,
"eval_loss": 0.3612821102142334,
"eval_precision": 61.22159090909091,
"eval_recall": 64.90963855421687,
"eval_runtime": 9.0113,
"eval_samples_per_second": 130.836,
"step": 750
},
{
"epoch": 4.23,
"learning_rate": 1.922398589065256e-05,
"loss": 0.1651,
"step": 800
},
{
"epoch": 4.23,
"eval_accuracy_score": 88.56505244365097,
"eval_f1": 38.932898373925724,
"eval_loss": 0.44898930191993713,
"eval_precision": 57.75749674054759,
"eval_recall": 66.71686746987952,
"eval_runtime": 10.3767,
"eval_samples_per_second": 113.62,
"step": 800
},
{
"epoch": 4.5,
"learning_rate": 1.834215167548501e-05,
"loss": 0.1649,
"step": 850
},
{
"epoch": 4.5,
"eval_accuracy_score": 88.66324481142603,
"eval_f1": 39.08416734249581,
"eval_loss": 0.45583415031433105,
"eval_precision": 58.894878706199464,
"eval_recall": 65.8132530120482,
"eval_runtime": 10.0575,
"eval_samples_per_second": 117.226,
"step": 850
},
{
"epoch": 4.76,
"learning_rate": 1.746031746031746e-05,
"loss": 0.1788,
"step": 900
},
{
"epoch": 4.76,
"eval_accuracy_score": 88.9622852041955,
"eval_f1": 40.02918919780824,
"eval_loss": 0.4240648150444031,
"eval_precision": 60.96866096866097,
"eval_recall": 64.45783132530121,
"eval_runtime": 10.0412,
"eval_samples_per_second": 117.417,
"step": 900
},
{
"epoch": 5.03,
"learning_rate": 1.6578483245149913e-05,
"loss": 0.1676,
"step": 950
},
{
"epoch": 5.03,
"eval_accuracy_score": 87.55634902923455,
"eval_f1": 43.86838413539317,
"eval_loss": 0.4299115240573883,
"eval_precision": 57.91505791505791,
"eval_recall": 67.7710843373494,
"eval_runtime": 10.0147,
"eval_samples_per_second": 117.726,
"step": 950
},
{
"epoch": 5.29,
"learning_rate": 1.5696649029982366e-05,
"loss": 0.1179,
"step": 1000
},
{
"epoch": 5.29,
"eval_accuracy_score": 87.20821245257756,
"eval_f1": 39.553135574552805,
"eval_loss": 0.49208664894104004,
"eval_precision": 56.746532156368225,
"eval_recall": 67.7710843373494,
"eval_runtime": 10.0407,
"eval_samples_per_second": 117.423,
"step": 1000
},
{
"epoch": 5.56,
"learning_rate": 1.4814814814814815e-05,
"loss": 0.1383,
"step": 1050
},
{
"epoch": 5.56,
"eval_accuracy_score": 87.743807185896,
"eval_f1": 38.8202089054134,
"eval_loss": 0.4442407786846161,
"eval_precision": 56.649616368286445,
"eval_recall": 66.71686746987952,
"eval_runtime": 9.9651,
"eval_samples_per_second": 118.313,
"step": 1050
},
{
"epoch": 5.82,
"learning_rate": 1.3932980599647267e-05,
"loss": 0.1381,
"step": 1100
},
{
"epoch": 5.82,
"eval_accuracy_score": 86.94487837536265,
"eval_f1": 38.34834659982452,
"eval_loss": 0.4821961224079132,
"eval_precision": 59.24932975871313,
"eval_recall": 66.56626506024097,
"eval_runtime": 9.9971,
"eval_samples_per_second": 117.934,
"step": 1100
},
{
"epoch": 6.08,
"learning_rate": 1.3051146384479719e-05,
"loss": 0.1308,
"step": 1150
},
{
"epoch": 6.08,
"eval_accuracy_score": 87.83307297478241,
"eval_f1": 43.06179088429465,
"eval_loss": 0.5310231447219849,
"eval_precision": 59.2896174863388,
"eval_recall": 65.36144578313254,
"eval_runtime": 10.0333,
"eval_samples_per_second": 117.508,
"step": 1150
},
{
"epoch": 6.35,
"learning_rate": 1.2169312169312169e-05,
"loss": 0.1008,
"step": 1200
},
{
"epoch": 6.35,
"eval_accuracy_score": 87.8062932381165,
"eval_f1": 38.25726977101606,
"eval_loss": 0.5424754023551941,
"eval_precision": 55.38277511961722,
"eval_recall": 69.7289156626506,
"eval_runtime": 9.9751,
"eval_samples_per_second": 118.194,
"step": 1200
},
{
"epoch": 6.61,
"learning_rate": 1.1287477954144621e-05,
"loss": 0.1133,
"step": 1250
},
{
"epoch": 6.61,
"eval_accuracy_score": 88.02945771033251,
"eval_f1": 40.50705288524587,
"eval_loss": 0.505093514919281,
"eval_precision": 57.824933687002655,
"eval_recall": 65.66265060240963,
"eval_runtime": 10.0854,
"eval_samples_per_second": 116.902,
"step": 1250
},
{
"epoch": 6.88,
"learning_rate": 1.0405643738977073e-05,
"loss": 0.1167,
"step": 1300
},
{
"epoch": 6.88,
"eval_accuracy_score": 87.98928810533363,
"eval_f1": 37.904391097050066,
"eval_loss": 0.4733125865459442,
"eval_precision": 57.469717362045756,
"eval_recall": 64.30722891566265,
"eval_runtime": 8.7518,
"eval_samples_per_second": 134.716,
"step": 1300
},
{
"epoch": 7.14,
"learning_rate": 9.523809523809523e-06,
"loss": 0.0987,
"step": 1350
},
{
"epoch": 7.14,
"eval_accuracy_score": 87.83753626422673,
"eval_f1": 37.322003651376725,
"eval_loss": 0.6008809208869934,
"eval_precision": 55.66502463054187,
"eval_recall": 68.07228915662651,
"eval_runtime": 8.967,
"eval_samples_per_second": 131.482,
"step": 1350
},
{
"epoch": 7.41,
"learning_rate": 8.641975308641977e-06,
"loss": 0.0929,
"step": 1400
},
{
"epoch": 7.41,
"eval_accuracy_score": 88.2883284981031,
"eval_f1": 38.96437550077822,
"eval_loss": 0.5702655911445618,
"eval_precision": 58.97771952817824,
"eval_recall": 67.7710843373494,
"eval_runtime": 10.9203,
"eval_samples_per_second": 107.964,
"step": 1400
},
{
"epoch": 7.67,
"learning_rate": 7.760141093474427e-06,
"loss": 0.0939,
"step": 1450
},
{
"epoch": 7.67,
"eval_accuracy_score": 87.980361526445,
"eval_f1": 36.57320046638653,
"eval_loss": 0.5611082315444946,
"eval_precision": 56.08194622279129,
"eval_recall": 65.96385542168674,
"eval_runtime": 11.1802,
"eval_samples_per_second": 105.454,
"step": 1450
},
{
"epoch": 7.94,
"learning_rate": 6.878306878306878e-06,
"loss": 0.0976,
"step": 1500
},
{
"epoch": 7.94,
"eval_accuracy_score": 88.6364650747601,
"eval_f1": 37.89298254130058,
"eval_loss": 0.6237432360649109,
"eval_precision": 59.53360768175583,
"eval_recall": 65.36144578313254,
"eval_runtime": 10.4805,
"eval_samples_per_second": 112.495,
"step": 1500
},
{
"epoch": 8.2,
"learning_rate": 5.99647266313933e-06,
"loss": 0.0806,
"step": 1550
},
{
"epoch": 8.2,
"eval_accuracy_score": 86.5654987725954,
"eval_f1": 36.826891066079725,
"eval_loss": 0.6972205638885498,
"eval_precision": 57.01643489254109,
"eval_recall": 67.92168674698796,
"eval_runtime": 10.9535,
"eval_samples_per_second": 107.636,
"step": 1550
},
{
"epoch": 8.47,
"learning_rate": 5.114638447971782e-06,
"loss": 0.0784,
"step": 1600
},
{
"epoch": 8.47,
"eval_accuracy_score": 87.85092613255969,
"eval_f1": 37.500099347680326,
"eval_loss": 0.6484220623970032,
"eval_precision": 58.68983957219251,
"eval_recall": 66.1144578313253,
"eval_runtime": 8.8955,
"eval_samples_per_second": 132.539,
"step": 1600
},
{
"epoch": 8.73,
"learning_rate": 4.232804232804233e-06,
"loss": 0.0821,
"step": 1650
},
{
"epoch": 8.73,
"eval_accuracy_score": 87.84646284311538,
"eval_f1": 41.22001286807445,
"eval_loss": 0.6608390808105469,
"eval_precision": 58.84718498659517,
"eval_recall": 66.1144578313253,
"eval_runtime": 8.8611,
"eval_samples_per_second": 133.053,
"step": 1650
},
{
"epoch": 8.99,
"learning_rate": 3.3509700176366843e-06,
"loss": 0.0834,
"step": 1700
},
{
"epoch": 8.99,
"eval_accuracy_score": 87.6768578442312,
"eval_f1": 40.52561281256491,
"eval_loss": 0.6585069298744202,
"eval_precision": 56.80628272251309,
"eval_recall": 65.36144578313254,
"eval_runtime": 8.9947,
"eval_samples_per_second": 131.078,
"step": 1700
},
{
"epoch": 9.26,
"learning_rate": 2.469135802469136e-06,
"loss": 0.0687,
"step": 1750
},
{
"epoch": 9.26,
"eval_accuracy_score": 87.67239455478688,
"eval_f1": 41.42806666569158,
"eval_loss": 0.7016396522521973,
"eval_precision": 57.929226736566186,
"eval_recall": 66.56626506024097,
"eval_runtime": 8.6369,
"eval_samples_per_second": 136.508,
"step": 1750
},
{
"epoch": 9.52,
"learning_rate": 1.5873015873015873e-06,
"loss": 0.0789,
"step": 1800
},
{
"epoch": 9.52,
"eval_accuracy_score": 88.091943762553,
"eval_f1": 40.693681724146366,
"eval_loss": 0.6877384781837463,
"eval_precision": 57.810413885180246,
"eval_recall": 65.21084337349397,
"eval_runtime": 8.9265,
"eval_samples_per_second": 132.079,
"step": 1800
},
{
"epoch": 9.79,
"learning_rate": 7.054673721340388e-07,
"loss": 0.0634,
"step": 1850
},
{
"epoch": 9.79,
"eval_accuracy_score": 87.85985271144834,
"eval_f1": 40.37377917369942,
"eval_loss": 0.7008724212646484,
"eval_precision": 58.56573705179283,
"eval_recall": 66.41566265060241,
"eval_runtime": 8.753,
"eval_samples_per_second": 134.697,
"step": 1850
},
{
"epoch": 10.0,
"step": 1890,
"total_flos": 6.580960338659328e+16,
"train_runtime": 3394.8711,
"train_samples_per_second": 0.557
}
],
"max_steps": 1890,
"num_train_epochs": 10,
"total_flos": 6.580960338659328e+16,
"trial_name": null,
"trial_params": null
}