longformer-sep_tok / checkpoint-1134 /trainer_state.json
Theoreticallyhugo's picture
Training in progress, epoch 14, checkpoint
25cd488 verified
{
"best_metric": 0.2527652680873871,
"best_model_checkpoint": "longformer-sep_tok/checkpoint-324",
"epoch": 14.0,
"eval_steps": 500,
"global_step": 1134,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 1.0,
"eval_Claim": {
"f1-score": 0.580553004795429,
"precision": 0.5136306192453511,
"recall": 0.6675269826372595,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.6929252383341696,
"precision": 0.7583745194947831,
"recall": 0.6378752886836028,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.9968272281511393,
"precision": 0.9967861557478368,
"recall": 0.996868303939344,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.8769487750556793,
"precision": 0.9107806691449815,
"recall": 0.8455403021704119,
"support": 13039.0
},
"eval_accuracy": 0.8654113924050633,
"eval_loss": 0.32089120149612427,
"eval_macro avg": {
"f1-score": 0.7868135615841043,
"precision": 0.7948929909082382,
"recall": 0.7869527193576545,
"support": 31600.0
},
"eval_runtime": 5.9486,
"eval_samples_per_second": 13.449,
"eval_steps_per_second": 1.681,
"eval_weighted avg": {
"f1-score": 0.8703967313850799,
"precision": 0.8797989523023909,
"recall": 0.8654113924050633,
"support": 31600.0
},
"step": 81
},
{
"epoch": 2.0,
"eval_Claim": {
"f1-score": 0.6029902384776968,
"precision": 0.6369094231271208,
"recall": 0.5725011731581418,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.8179940325912326,
"precision": 0.8129562043795621,
"recall": 0.8230946882217091,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.9984359565360553,
"precision": 0.9972866304884065,
"recall": 0.9995879347288611,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.9017917895214335,
"precision": 0.8891539321654864,
"recall": 0.9147940793005599,
"support": 13039.0
},
"eval_accuracy": 0.8949050632911393,
"eval_loss": 0.25671303272247314,
"eval_macro avg": {
"f1-score": 0.8303030042816045,
"precision": 0.8340765475401439,
"recall": 0.8274944688523179,
"support": 31600.0
},
"eval_runtime": 5.97,
"eval_samples_per_second": 13.4,
"eval_steps_per_second": 1.675,
"eval_weighted avg": {
"f1-score": 0.8928603328205832,
"precision": 0.8914339316361279,
"recall": 0.8949050632911393,
"support": 31600.0
},
"step": 162
},
{
"epoch": 3.0,
"eval_Claim": {
"f1-score": 0.6010706092276319,
"precision": 0.6579241071428571,
"recall": 0.5532613796339747,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.8101840181497353,
"precision": 0.8917869034406215,
"recall": 0.7422632794457275,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.997983456109305,
"precision": 0.9967118783394986,
"recall": 0.9992582825119499,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.9085203780271707,
"precision": 0.8758630507509432,
"recall": 0.9437073395199018,
"support": 13039.0
},
"eval_accuracy": 0.8985759493670886,
"eval_loss": 0.2674676775932312,
"eval_macro avg": {
"f1-score": 0.8294396153784607,
"precision": 0.8555714849184801,
"recall": 0.8096225702778885,
"support": 31600.0
},
"eval_runtime": 6.0212,
"eval_samples_per_second": 13.286,
"eval_steps_per_second": 1.661,
"eval_weighted avg": {
"f1-score": 0.894668981055346,
"precision": 0.8939642861109123,
"recall": 0.8985759493670886,
"support": 31600.0
},
"step": 243
},
{
"epoch": 4.0,
"eval_Claim": {
"f1-score": 0.66,
"precision": 0.6399294843543412,
"recall": 0.681370248709526,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.8682903759754079,
"precision": 0.8895348837209303,
"recall": 0.8480369515011548,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.9983125488743466,
"precision": 0.9971224204554797,
"recall": 0.9995055216746332,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.9011362757980985,
"precision": 0.9082976236852357,
"recall": 0.8940869698596519,
"support": 13039.0
},
"eval_accuracy": 0.9027215189873418,
"eval_loss": 0.2527652680873871,
"eval_macro avg": {
"f1-score": 0.8569348001619632,
"precision": 0.8587211030539967,
"recall": 0.8557499229362415,
"support": 31600.0
},
"eval_runtime": 5.9641,
"eval_samples_per_second": 13.414,
"eval_steps_per_second": 1.677,
"eval_weighted avg": {
"f1-score": 0.9036775010177053,
"precision": 0.9049240079307782,
"recall": 0.9027215189873418,
"support": 31600.0
},
"step": 324
},
{
"epoch": 5.0,
"eval_Claim": {
"f1-score": 0.6614122219822932,
"precision": 0.6126,
"recall": 0.7186766776161426,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.805946376426865,
"precision": 0.947565543071161,
"recall": 0.7011547344110854,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.998518030627367,
"precision": 0.9975324888962,
"recall": 0.9995055216746332,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.9051354379999227,
"precision": 0.9121495327102803,
"recall": 0.8982283917478334,
"support": 13039.0
},
"eval_accuracy": 0.8993987341772152,
"eval_loss": 0.29903507232666016,
"eval_macro avg": {
"f1-score": 0.842753016759112,
"precision": 0.8674618911694103,
"recall": 0.8293913313624237,
"support": 31600.0
},
"eval_runtime": 5.9489,
"eval_samples_per_second": 13.448,
"eval_steps_per_second": 1.681,
"eval_weighted avg": {
"f1-score": 0.9013256821128531,
"precision": 0.9069606828488892,
"recall": 0.8993987341772152,
"support": 31600.0
},
"step": 405
},
{
"epoch": 6.0,
"eval_Claim": {
"f1-score": 0.6276952563488261,
"precision": 0.6412139011257953,
"recall": 0.6147348662599719,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.8774834437086093,
"precision": 0.8991759573436743,
"recall": 0.8568129330254042,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.9990939044481055,
"precision": 0.99860036225918,
"recall": 0.9995879347288611,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.8991800789553598,
"precision": 0.8901916572717024,
"recall": 0.9083518674744996,
"support": 13039.0
},
"eval_accuracy": 0.900253164556962,
"eval_loss": 0.3113674819469452,
"eval_macro avg": {
"f1-score": 0.8508631708652252,
"precision": 0.8572954695000881,
"recall": 0.8448719003721842,
"support": 31600.0
},
"eval_runtime": 5.9773,
"eval_samples_per_second": 13.384,
"eval_steps_per_second": 1.673,
"eval_weighted avg": {
"f1-score": 0.8994431431727875,
"precision": 0.8988542850970194,
"recall": 0.900253164556962,
"support": 31600.0
},
"step": 486
},
{
"epoch": 6.17,
"grad_norm": 5.353315830230713,
"learning_rate": 1.3827160493827162e-05,
"loss": 0.2362,
"step": 500
},
{
"epoch": 7.0,
"eval_Claim": {
"f1-score": 0.6218705409261902,
"precision": 0.6953872932985204,
"recall": 0.5624120131393712,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.8918461166223083,
"precision": 0.9364837398373984,
"recall": 0.851270207852194,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.9988881110241733,
"precision": 0.9982714626718249,
"recall": 0.9995055216746332,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.9056325023084026,
"precision": 0.8734682245654033,
"recall": 0.9402561546130839,
"support": 13039.0
},
"eval_accuracy": 0.9059493670886076,
"eval_loss": 0.38884854316711426,
"eval_macro avg": {
"f1-score": 0.8545593177202686,
"precision": 0.8759026800932868,
"recall": 0.8383609743198206,
"support": 31600.0
},
"eval_runtime": 5.9526,
"eval_samples_per_second": 13.44,
"eval_steps_per_second": 1.68,
"eval_weighted avg": {
"f1-score": 0.9022249881228259,
"precision": 0.9016900648403317,
"recall": 0.9059493670886076,
"support": 31600.0
},
"step": 567
},
{
"epoch": 8.0,
"eval_Claim": {
"f1-score": 0.6671109630123293,
"precision": 0.6334106728538283,
"recall": 0.7045987799155327,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.8920657612580415,
"precision": 0.9212598425196851,
"recall": 0.8646651270207852,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.998846882464377,
"precision": 0.9982713203819559,
"recall": 0.9994231086204055,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.8988257251730306,
"precision": 0.9115860872308542,
"recall": 0.8864176700667229,
"support": 13039.0
},
"eval_accuracy": 0.9037974683544304,
"eval_loss": 0.40879106521606445,
"eval_macro avg": {
"f1-score": 0.8642123329769446,
"precision": 0.8661319807465809,
"recall": 0.8637761714058616,
"support": 31600.0
},
"eval_runtime": 5.9421,
"eval_samples_per_second": 13.463,
"eval_steps_per_second": 1.683,
"eval_weighted avg": {
"f1-score": 0.9055172784758262,
"precision": 0.9080164253061992,
"recall": 0.9037974683544304,
"support": 31600.0
},
"step": 648
},
{
"epoch": 9.0,
"eval_Claim": {
"f1-score": 0.6533434157934537,
"precision": 0.6054054054054054,
"recall": 0.7095260441107462,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.8944609072258947,
"precision": 0.882988298829883,
"recall": 0.9062355658198614,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.9986414721501792,
"precision": 0.9976968001974171,
"recall": 0.9995879347288611,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.8881852364931724,
"precision": 0.9177163422214952,
"recall": 0.8604954367666232,
"support": 13039.0
},
"eval_accuracy": 0.8966772151898734,
"eval_loss": 0.4774320125579834,
"eval_macro avg": {
"f1-score": 0.8586577579156749,
"precision": 0.8509517116635501,
"recall": 0.868961245356523,
"support": 31600.0
},
"eval_runtime": 5.9344,
"eval_samples_per_second": 13.481,
"eval_steps_per_second": 1.685,
"eval_weighted avg": {
"f1-score": 0.899355076707611,
"precision": 0.903926071665382,
"recall": 0.8966772151898734,
"support": 31600.0
},
"step": 729
},
{
"epoch": 10.0,
"eval_Claim": {
"f1-score": 0.6297460701330109,
"precision": 0.6497005988023952,
"recall": 0.6109807602064758,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.8433544303797468,
"precision": 0.9827904118008605,
"recall": 0.7385681293302541,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.9991762089134196,
"precision": 0.9987648221343873,
"recall": 0.9995879347288611,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.9050632911392404,
"precision": 0.8794587945879458,
"recall": 0.9322033898305084,
"support": 13039.0
},
"eval_accuracy": 0.901487341772152,
"eval_loss": 0.514373242855072,
"eval_macro avg": {
"f1-score": 0.8443350001413544,
"precision": 0.8776786568313971,
"recall": 0.8203350535240248,
"support": 31600.0
},
"eval_runtime": 5.9234,
"eval_samples_per_second": 13.506,
"eval_steps_per_second": 1.688,
"eval_weighted avg": {
"f1-score": 0.8998406476202226,
"precision": 0.901362049622011,
"recall": 0.901487341772152,
"support": 31600.0
},
"step": 810
},
{
"epoch": 11.0,
"eval_Claim": {
"f1-score": 0.6562464084587979,
"precision": 0.6431628745212886,
"recall": 0.6698732989206945,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.8588056680161943,
"precision": 0.9496362618914381,
"recall": 0.7838337182448037,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.9985593743568636,
"precision": 0.9974508675273415,
"recall": 0.9996703477830888,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.9037025750419017,
"precision": 0.8977522137289033,
"recall": 0.9097323414372268,
"support": 13039.0
},
"eval_accuracy": 0.9032911392405063,
"eval_loss": 0.5588687658309937,
"eval_macro avg": {
"f1-score": 0.8543285064684394,
"precision": 0.8720005544172429,
"recall": 0.8407774265964535,
"support": 31600.0
},
"eval_runtime": 5.9421,
"eval_samples_per_second": 13.463,
"eval_steps_per_second": 1.683,
"eval_weighted avg": {
"f1-score": 0.9036751198899997,
"precision": 0.9052526145440706,
"recall": 0.9032911392405063,
"support": 31600.0
},
"step": 891
},
{
"epoch": 12.0,
"eval_Claim": {
"f1-score": 0.6462172647914647,
"precision": 0.668590065228299,
"recall": 0.6252932895354294,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.8801742919389979,
"precision": 0.9247202441505595,
"recall": 0.8397228637413395,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.9987650255228059,
"precision": 0.99777924000658,
"recall": 0.9997527608373167,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.9052734743111314,
"precision": 0.8901408450704226,
"recall": 0.920929519134903,
"support": 13039.0
},
"eval_accuracy": 0.905759493670886,
"eval_loss": 0.6348227262496948,
"eval_macro avg": {
"f1-score": 0.8576075141410999,
"precision": 0.8703075986139653,
"recall": 0.8464246083122471,
"support": 31600.0
},
"eval_runtime": 5.9311,
"eval_samples_per_second": 13.488,
"eval_steps_per_second": 1.686,
"eval_weighted avg": {
"f1-score": 0.9045136384754975,
"precision": 0.9039604418893055,
"recall": 0.905759493670886,
"support": 31600.0
},
"step": 972
},
{
"epoch": 12.35,
"grad_norm": 1.0823522806167603,
"learning_rate": 7.654320987654322e-06,
"loss": 0.0312,
"step": 1000
},
{
"epoch": 13.0,
"eval_Claim": {
"f1-score": 0.6595566207309765,
"precision": 0.6740142052412442,
"recall": 0.6457062412013139,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.8993576017130621,
"precision": 0.9273797841020608,
"recall": 0.8729792147806005,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.9989296006587073,
"precision": 0.998025666337611,
"recall": 0.9998351738915444,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.9065321295804567,
"precision": 0.8968700743075884,
"recall": 0.9164046322570749,
"support": 13039.0
},
"eval_accuracy": 0.9089556962025317,
"eval_loss": 0.5934906005859375,
"eval_macro avg": {
"f1-score": 0.8660939881708007,
"precision": 0.8740724324971261,
"recall": 0.8587313155326335,
"support": 31600.0
},
"eval_runtime": 5.948,
"eval_samples_per_second": 13.45,
"eval_steps_per_second": 1.681,
"eval_weighted avg": {
"f1-score": 0.9082096119384979,
"precision": 0.9077455097960874,
"recall": 0.9089556962025317,
"support": 31600.0
},
"step": 1053
},
{
"epoch": 14.0,
"eval_Claim": {
"f1-score": 0.6832563099893353,
"precision": 0.6902082834570266,
"recall": 0.6764429845143125,
"support": 4262.0
},
"eval_MajorClaim": {
"f1-score": 0.892668902731193,
"precision": 0.9273270283723245,
"recall": 0.8605080831408776,
"support": 2165.0
},
"eval_O": {
"f1-score": 0.9988875612871329,
"precision": 0.9987641097470544,
"recall": 0.9990110433492665,
"support": 12134.0
},
"eval_Premise": {
"f1-score": 0.9105487156102752,
"precision": 0.9023875875574301,
"recall": 0.9188588081908122,
"support": 13039.0
},
"eval_accuracy": 0.9129430379746836,
"eval_loss": 0.6126174330711365,
"eval_macro avg": {
"f1-score": 0.8713403724044841,
"precision": 0.879671752283459,
"recall": 0.8637052297988173,
"support": 31600.0
},
"eval_runtime": 5.9438,
"eval_samples_per_second": 13.459,
"eval_steps_per_second": 1.682,
"eval_weighted avg": {
"f1-score": 0.9125890170597477,
"precision": 0.9124862715934182,
"recall": 0.9129430379746836,
"support": 31600.0
},
"step": 1134
}
],
"logging_steps": 500,
"max_steps": 1620,
"num_input_tokens_seen": 0,
"num_train_epochs": 20,
"save_steps": 500,
"total_flos": 2013042258175200.0,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}