longformer-sep_tok / checkpoint-2754 /trainer_state.json
Theoreticallyhugo's picture
Training in progress, epoch 34, checkpoint
50f83e4 verified
raw
history blame
45.8 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 34.0,
"eval_steps": 500,
"global_step": 2754,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 1.0,
"eval_Claim": {
"f1-score": 0.5385059461955894,
"precision": 0.6267132491265789,
"recall": 0.4720647773279352,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.7814784727863526,
"precision": 0.9588039867109635,
"recall": 0.6595063985374772,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9938608715302625,
"precision": 0.9996969237763298,
"recall": 0.9880925634688834,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.8968837690149166,
"precision": 0.8455112497215416,
"recall": 0.954902824076986,
"support": 15899.0
},
"eval_accuracy": 0.8837548103353491,
"eval_loss": 0.2860471308231354,
"eval_macro avg": {
"f1-score": 0.8026822648817803,
"precision": 0.8576813523338533,
"recall": 0.7686416408528204,
"support": 36380.0
},
"eval_runtime": 4.9572,
"eval_samples_per_second": 16.34,
"eval_steps_per_second": 2.219,
"eval_weighted avg": {
"f1-score": 0.8768739289107891,
"precision": 0.8792073379911217,
"recall": 0.8837548103353491,
"support": 36380.0
},
"step": 81
},
{
"epoch": 2.0,
"eval_Claim": {
"f1-score": 0.6469459868753155,
"precision": 0.6453172205438067,
"recall": 0.648582995951417,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.8435970110796187,
"precision": 0.9669226225634967,
"recall": 0.7481718464351006,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.998877245508982,
"precision": 0.9983541557567143,
"recall": 0.9994008836965476,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9099026477336144,
"precision": 0.8972179761540813,
"recall": 0.922951129001824,
"support": 15899.0
},
"eval_accuracy": 0.903243540406817,
"eval_loss": 0.24885737895965576,
"eval_macro avg": {
"f1-score": 0.8498307227993828,
"precision": 0.8769529937545246,
"recall": 0.8297767137712222,
"support": 36380.0
},
"eval_runtime": 5.0097,
"eval_samples_per_second": 16.169,
"eval_steps_per_second": 2.196,
"eval_weighted avg": {
"f1-score": 0.9028656814459712,
"precision": 0.9043261520711511,
"recall": 0.903243540406817,
"support": 36380.0
},
"step": 162
},
{
"epoch": 3.0,
"eval_Claim": {
"f1-score": 0.7040630102775911,
"precision": 0.6698958142935478,
"recall": 0.7419028340080972,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.8888411094388303,
"precision": 0.8392204628501827,
"recall": 0.9446983546617916,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.998989104796136,
"precision": 0.9988769092542678,
"recall": 0.9991013255448213,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.914582593823615,
"precision": 0.9390987408880053,
"recall": 0.8913139191144097,
"support": 15899.0
},
"eval_accuracy": 0.9137987905442551,
"eval_loss": 0.24298855662345886,
"eval_macro avg": {
"f1-score": 0.8766189545840432,
"precision": 0.8617729818215009,
"recall": 0.8942541083322799,
"support": 36380.0
},
"eval_runtime": 4.9876,
"eval_samples_per_second": 16.24,
"eval_steps_per_second": 2.205,
"eval_weighted avg": {
"f1-score": 0.9154289662937856,
"precision": 0.9184781183611038,
"recall": 0.9137987905442551,
"support": 36380.0
},
"step": 243
},
{
"epoch": 4.0,
"eval_Claim": {
"f1-score": 0.6981611893583725,
"precision": 0.6754352763058289,
"recall": 0.7224696356275304,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.901231802911534,
"precision": 0.8836187966622749,
"recall": 0.9195612431444241,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9989130842172332,
"precision": 0.9998499399759904,
"recall": 0.9979779824758481,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9146224912392481,
"precision": 0.9266670970240785,
"recall": 0.9028869740235235,
"support": 15899.0
},
"eval_accuracy": 0.9142935678944475,
"eval_loss": 0.23980534076690674,
"eval_macro avg": {
"f1-score": 0.8782321419315969,
"precision": 0.8713927774920431,
"recall": 0.8857239588178316,
"support": 36380.0
},
"eval_runtime": 4.9726,
"eval_samples_per_second": 16.289,
"eval_steps_per_second": 2.212,
"eval_weighted avg": {
"f1-score": 0.9153623106642749,
"precision": 0.9168247558035478,
"recall": 0.9142935678944475,
"support": 36380.0
},
"step": 324
},
{
"epoch": 5.0,
"eval_Claim": {
"f1-score": 0.6650318783717509,
"precision": 0.6450999048525214,
"recall": 0.6862348178137652,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9172031076581575,
"precision": 0.8916702632714717,
"recall": 0.9442413162705667,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9982383147794145,
"precision": 0.9992495872730002,
"recall": 0.9972290870965326,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.8998438545616774,
"precision": 0.9119622787753521,
"recall": 0.8880432731618341,
"support": 15899.0
},
"eval_accuracy": 0.9040956569543706,
"eval_loss": 0.2932502031326294,
"eval_macro avg": {
"f1-score": 0.87007928884275,
"precision": 0.8619955085430864,
"recall": 0.8789371235856747,
"support": 36380.0
},
"eval_runtime": 4.9595,
"eval_samples_per_second": 16.332,
"eval_steps_per_second": 2.218,
"eval_weighted avg": {
"f1-score": 0.9051179642561333,
"precision": 0.9065430476942037,
"recall": 0.9040956569543706,
"support": 36380.0
},
"step": 405
},
{
"epoch": 6.0,
"eval_Claim": {
"f1-score": 0.6622698072805139,
"precision": 0.5740163325909429,
"recall": 0.7825910931174089,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9223704351629127,
"precision": 0.8842767295597485,
"recall": 0.9638939670932358,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9982004948639124,
"precision": 0.9993994444861497,
"recall": 0.9970044184827379,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.8780749379985253,
"precision": 0.9398091685199799,
"recall": 0.8239511918988616,
"support": 15899.0
},
"eval_accuracy": 0.8902693787795491,
"eval_loss": 0.3604726195335388,
"eval_macro avg": {
"f1-score": 0.865228918826466,
"precision": 0.8493754187892053,
"recall": 0.8918601676480611,
"support": 36380.0
},
"eval_runtime": 4.991,
"eval_samples_per_second": 16.229,
"eval_steps_per_second": 2.204,
"eval_weighted avg": {
"f1-score": 0.8955262233990811,
"precision": 0.9086708334194257,
"recall": 0.8902693787795491,
"support": 36380.0
},
"step": 486
},
{
"epoch": 6.17,
"grad_norm": 17.04454231262207,
"learning_rate": 1.7530864197530865e-05,
"loss": 0.2104,
"step": 500
},
{
"epoch": 7.0,
"eval_Claim": {
"f1-score": 0.7037529399734124,
"precision": 0.7110973341599504,
"recall": 0.6965587044534413,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9196885428253615,
"precision": 0.8959687906371911,
"recall": 0.9446983546617916,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9987257327036954,
"precision": 0.9996248780853777,
"recall": 0.997828203399985,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9188781285372909,
"precision": 0.9187048098082364,
"recall": 0.9190515126737531,
"support": 15899.0
},
"eval_accuracy": 0.9192963166575041,
"eval_loss": 0.36414769291877747,
"eval_macro avg": {
"f1-score": 0.8852613360099401,
"precision": 0.8813489531726889,
"recall": 0.8895341937972427,
"support": 36380.0
},
"eval_runtime": 4.9835,
"eval_samples_per_second": 16.254,
"eval_steps_per_second": 2.207,
"eval_weighted avg": {
"f1-score": 0.9190227083446231,
"precision": 0.9188476996393494,
"recall": 0.9192963166575041,
"support": 36380.0
},
"step": 567
},
{
"epoch": 8.0,
"eval_Claim": {
"f1-score": 0.7025399811853246,
"precision": 0.6562390158172232,
"recall": 0.7558704453441295,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9105835367206567,
"precision": 0.8848641655886158,
"recall": 0.9378427787934186,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9986513823331086,
"precision": 0.9991005172026085,
"recall": 0.9982026510896428,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9083384525849526,
"precision": 0.9345974717232203,
"recall": 0.8835146864582678,
"support": 15899.0
},
"eval_accuracy": 0.911544804837823,
"eval_loss": 0.35222452878952026,
"eval_macro avg": {
"f1-score": 0.8800283382060107,
"precision": 0.8687002925829169,
"recall": 0.8938576404213647,
"support": 36380.0
},
"eval_runtime": 4.9741,
"eval_samples_per_second": 16.285,
"eval_steps_per_second": 2.211,
"eval_weighted avg": {
"f1-score": 0.9136769997620249,
"precision": 0.9174837257168467,
"recall": 0.911544804837823,
"support": 36380.0
},
"step": 648
},
{
"epoch": 9.0,
"eval_Claim": {
"f1-score": 0.6566933693658913,
"precision": 0.6275594908688434,
"recall": 0.6886639676113361,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.8702745722244328,
"precision": 0.7706131078224101,
"recall": 0.9995429616087751,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9987648313807688,
"precision": 0.9983537862915295,
"recall": 0.9991762150827529,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9001826722338205,
"precision": 0.9350138917124077,
"recall": 0.8678533241084345,
"support": 15899.0
},
"eval_accuracy": 0.8996426608026388,
"eval_loss": 0.4436376690864563,
"eval_macro avg": {
"f1-score": 0.8564788613012284,
"precision": 0.8328850691737977,
"recall": 0.8888091171028246,
"support": 36380.0
},
"eval_runtime": 5.0402,
"eval_samples_per_second": 16.071,
"eval_steps_per_second": 2.182,
"eval_weighted avg": {
"f1-score": 0.9015045934020747,
"precision": 0.90662587513727,
"recall": 0.8996426608026388,
"support": 36380.0
},
"step": 729
},
{
"epoch": 10.0,
"eval_Claim": {
"f1-score": 0.7371103705155853,
"precision": 0.7145572025845686,
"recall": 0.7611336032388664,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9365328549002018,
"precision": 0.9194187582562747,
"recall": 0.9542961608775137,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9991386090408599,
"precision": 0.9993257416841475,
"recall": 0.9989515464689583,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9225428371233836,
"precision": 0.9344473837021743,
"recall": 0.9109377948298635,
"support": 15899.0
},
"eval_accuracy": 0.9255085211654756,
"eval_loss": 0.44094258546829224,
"eval_macro avg": {
"f1-score": 0.8988311678950076,
"precision": 0.8919372715567913,
"recall": 0.9063297763538004,
"support": 36380.0
},
"eval_runtime": 4.9458,
"eval_samples_per_second": 16.378,
"eval_steps_per_second": 2.224,
"eval_weighted avg": {
"f1-score": 0.9263184587909816,
"precision": 0.927497977075915,
"recall": 0.9255085211654756,
"support": 36380.0
},
"step": 810
},
{
"epoch": 11.0,
"eval_Claim": {
"f1-score": 0.6750768578394996,
"precision": 0.7086579123080348,
"recall": 0.6445344129554655,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9008264462809917,
"precision": 0.8220211161387632,
"recall": 0.9963436928702011,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9990631440884392,
"precision": 0.9998499849984999,
"recall": 0.9982775406275743,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9168605748066159,
"precision": 0.9167452681883921,
"recall": 0.9169759104346186,
"support": 15899.0
},
"eval_accuracy": 0.9145959318306762,
"eval_loss": 0.4954843521118164,
"eval_macro avg": {
"f1-score": 0.8729567557538865,
"precision": 0.8618185704084225,
"recall": 0.8890328892219649,
"support": 36380.0
},
"eval_runtime": 5.0503,
"eval_samples_per_second": 16.039,
"eval_steps_per_second": 2.178,
"eval_weighted avg": {
"f1-score": 0.9132365141301059,
"precision": 0.9132952764575458,
"recall": 0.9145959318306762,
"support": 36380.0
},
"step": 891
},
{
"epoch": 12.0,
"eval_Claim": {
"f1-score": 0.6857200770244248,
"precision": 0.6866247209255125,
"recall": 0.6848178137651821,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9214571488734471,
"precision": 0.8543537680593518,
"recall": 1.0,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9987652935234033,
"precision": 0.9979811574697174,
"recall": 0.9995506627724107,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9127542413343095,
"precision": 0.9239592731022039,
"recall": 0.9018177243851815,
"support": 15899.0
},
"eval_accuracy": 0.91412864211105,
"eval_loss": 0.5210011005401611,
"eval_macro avg": {
"f1-score": 0.8796741901888963,
"precision": 0.8657297298891964,
"recall": 0.8965465502306935,
"support": 36380.0
},
"eval_runtime": 4.9497,
"eval_samples_per_second": 16.365,
"eval_steps_per_second": 2.222,
"eval_weighted avg": {
"f1-score": 0.9140186385274314,
"precision": 0.9147147620844138,
"recall": 0.91412864211105,
"support": 36380.0
},
"step": 972
},
{
"epoch": 12.35,
"grad_norm": 3.2667109966278076,
"learning_rate": 1.506172839506173e-05,
"loss": 0.0286,
"step": 1000
},
{
"epoch": 13.0,
"eval_Claim": {
"f1-score": 0.7024380291747424,
"precision": 0.7079991774624718,
"recall": 0.6969635627530364,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9075119409465915,
"precision": 0.8643507030603805,
"recall": 0.9552102376599635,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9986895802912875,
"precision": 0.9985774183887391,
"recall": 0.9988017673930952,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9200429808482398,
"precision": 0.9246014101505431,
"recall": 0.9155292785709793,
"support": 15899.0
},
"eval_accuracy": 0.9188015393073117,
"eval_loss": 0.5909315347671509,
"eval_macro avg": {
"f1-score": 0.8821706328152152,
"precision": 0.8738821772655336,
"recall": 0.8916262115942686,
"support": 36380.0
},
"eval_runtime": 4.9699,
"eval_samples_per_second": 16.298,
"eval_steps_per_second": 2.213,
"eval_weighted avg": {
"f1-score": 0.9186076225687217,
"precision": 0.918717904416961,
"recall": 0.9188015393073117,
"support": 36380.0
},
"step": 1053
},
{
"epoch": 14.0,
"eval_Claim": {
"f1-score": 0.7095000509113125,
"precision": 0.7137881581643106,
"recall": 0.7052631578947368,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.921028466483012,
"precision": 0.9252767527675276,
"recall": 0.916819012797075,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9994754982766372,
"precision": 1.0,
"recall": 0.9989515464689583,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9206986297074409,
"precision": 0.9180215107553776,
"recall": 0.9233914082646707,
"support": 15899.0
},
"eval_accuracy": 0.9211105002748763,
"eval_loss": 0.559357225894928,
"eval_macro avg": {
"f1-score": 0.8876756613446006,
"precision": 0.889271605421804,
"recall": 0.8861062813563602,
"support": 36380.0
},
"eval_runtime": 4.9811,
"eval_samples_per_second": 16.262,
"eval_steps_per_second": 2.208,
"eval_weighted avg": {
"f1-score": 0.9209544908843665,
"precision": 0.9208148168193181,
"recall": 0.9211105002748763,
"support": 36380.0
},
"step": 1134
},
{
"epoch": 15.0,
"eval_Claim": {
"f1-score": 0.6915445526531356,
"precision": 0.6282454109475774,
"recall": 0.7690283400809717,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9327135900841081,
"precision": 0.9042918454935622,
"recall": 0.9629798903107861,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9995130539011873,
"precision": 0.9998501199040767,
"recall": 0.9991762150827529,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.8984226716408141,
"precision": 0.936421311139914,
"recall": 0.8633876344424177,
"support": 15899.0
},
"eval_accuracy": 0.9064046179219352,
"eval_loss": 0.6213898658752441,
"eval_macro avg": {
"f1-score": 0.8805484670698113,
"precision": 0.8672021718712826,
"recall": 0.898643019979232,
"support": 36380.0
},
"eval_runtime": 4.9756,
"eval_samples_per_second": 16.28,
"eval_steps_per_second": 2.211,
"eval_weighted avg": {
"f1-score": 0.9094977265082566,
"precision": 0.9159231436204942,
"recall": 0.9064046179219352,
"support": 36380.0
},
"step": 1215
},
{
"epoch": 16.0,
"eval_Claim": {
"f1-score": 0.6819089456869009,
"precision": 0.6727738376674547,
"recall": 0.6912955465587044,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9290586630286495,
"precision": 0.9244343891402715,
"recall": 0.9337294332723949,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9998127410958391,
"precision": 1.0,
"recall": 0.9996255523103422,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9087691578448412,
"precision": 0.9131843007748,
"recall": 0.9043965029247123,
"support": 15899.0
},
"eval_accuracy": 0.9121770203408466,
"eval_loss": 0.6214332580566406,
"eval_macro avg": {
"f1-score": 0.8798873769140577,
"precision": 0.8775981318956315,
"recall": 0.8822617587665385,
"support": 36380.0
},
"eval_runtime": 4.962,
"eval_samples_per_second": 16.324,
"eval_steps_per_second": 2.217,
"eval_weighted avg": {
"f1-score": 0.9126011797368843,
"precision": 0.9130808795913876,
"recall": 0.9121770203408466,
"support": 36380.0
},
"step": 1296
},
{
"epoch": 17.0,
"eval_Claim": {
"f1-score": 0.6977198697068404,
"precision": 0.7524590163934426,
"recall": 0.6504048582995952,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.935002195871761,
"precision": 0.8998309382924767,
"recall": 0.9730347349177331,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9994757732344791,
"precision": 0.9994757732344791,
"recall": 0.9994757732344791,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9219572623103128,
"precision": 0.908120309926179,
"recall": 0.9362224039247752,
"support": 15899.0
},
"eval_accuracy": 0.9228422210005498,
"eval_loss": 0.6075455546379089,
"eval_macro avg": {
"f1-score": 0.8885387752808483,
"precision": 0.8899715094616443,
"recall": 0.8897844425941457,
"support": 36380.0
},
"eval_runtime": 5.0046,
"eval_samples_per_second": 16.185,
"eval_steps_per_second": 2.198,
"eval_weighted avg": {
"f1-score": 0.9207454500932067,
"precision": 0.9200160099363349,
"recall": 0.9228422210005498,
"support": 36380.0
},
"step": 1377
},
{
"epoch": 18.0,
"eval_Claim": {
"f1-score": 0.6748943740271293,
"precision": 0.7486433152442032,
"recall": 0.6143724696356275,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9180620674257648,
"precision": 0.8974247053688346,
"recall": 0.9396709323583181,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9997753145596164,
"precision": 0.9998501984870047,
"recall": 0.9997004418482738,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9206027683147655,
"precision": 0.898945097099017,
"recall": 0.9433297691678721,
"support": 15899.0
},
"eval_accuracy": 0.9191313908741067,
"eval_loss": 0.6335331201553345,
"eval_macro avg": {
"f1-score": 0.878333631081819,
"precision": 0.8862158290497649,
"recall": 0.874268403252523,
"support": 36380.0
},
"eval_runtime": 4.9489,
"eval_samples_per_second": 16.367,
"eval_steps_per_second": 2.223,
"eval_weighted avg": {
"f1-score": 0.9161451676737934,
"precision": 0.9154807870211,
"recall": 0.9191313908741067,
"support": 36380.0
},
"step": 1458
},
{
"epoch": 18.52,
"grad_norm": 0.10011027753353119,
"learning_rate": 1.2592592592592593e-05,
"loss": 0.0091,
"step": 1500
},
{
"epoch": 19.0,
"eval_Claim": {
"f1-score": 0.7046035805626598,
"precision": 0.7439243924392439,
"recall": 0.6692307692307692,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9264422045893201,
"precision": 0.8727272727272727,
"recall": 0.9872029250457038,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.999213041034289,
"precision": 1.0,
"recall": 0.9984273197034375,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.921880854252529,
"precision": 0.9153078306156612,
"recall": 0.9285489653437323,
"support": 15899.0
},
"eval_accuracy": 0.9225123694337548,
"eval_loss": 0.6328368782997131,
"eval_macro avg": {
"f1-score": 0.8880349201096995,
"precision": 0.8829898739455444,
"recall": 0.8958524948309107,
"support": 36380.0
},
"eval_runtime": 4.9991,
"eval_samples_per_second": 16.203,
"eval_steps_per_second": 2.2,
"eval_weighted avg": {
"f1-score": 0.9210355324440019,
"precision": 0.9205605544347316,
"recall": 0.9225123694337548,
"support": 36380.0
},
"step": 1539
},
{
"epoch": 20.0,
"eval_Claim": {
"f1-score": 0.6881697835679765,
"precision": 0.7153778942769768,
"recall": 0.6629554655870445,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9309440559440559,
"precision": 0.8919597989949749,
"recall": 0.973491773308958,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9997753313862053,
"precision": 0.9997753313862053,
"recall": 0.9997753313862053,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9161451814768461,
"precision": 0.9115248116555632,
"recall": 0.92081262972514,
"support": 15899.0
},
"eval_accuracy": 0.9179494227597581,
"eval_loss": 0.6887677907943726,
"eval_macro avg": {
"f1-score": 0.883758588093771,
"precision": 0.87965945907843,
"recall": 0.8892588000018369,
"support": 36380.0
},
"eval_runtime": 5.0229,
"eval_samples_per_second": 16.126,
"eval_steps_per_second": 2.19,
"eval_weighted avg": {
"f1-score": 0.916774507023963,
"precision": 0.9161052176591827,
"recall": 0.9179494227597581,
"support": 36380.0
},
"step": 1620
},
{
"epoch": 21.0,
"eval_Claim": {
"f1-score": 0.6423871114136601,
"precision": 0.7307692307692307,
"recall": 0.573076923076923,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.8879392212725545,
"precision": 0.9239130434782609,
"recall": 0.8546617915904936,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9984999624990625,
"precision": 1.0,
"recall": 0.9970044184827379,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9172614007499698,
"precision": 0.883336245558856,
"recall": 0.9538964714761935,
"support": 15899.0
},
"eval_accuracy": 0.9120395821880154,
"eval_loss": 0.7241741418838501,
"eval_macro avg": {
"f1-score": 0.8615219239838117,
"precision": 0.8845046299515869,
"recall": 0.844659901156587,
"support": 36380.0
},
"eval_runtime": 4.9949,
"eval_samples_per_second": 16.217,
"eval_steps_per_second": 2.202,
"eval_weighted avg": {
"f1-score": 0.9079909938510604,
"precision": 0.9078802833224489,
"recall": 0.9120395821880154,
"support": 36380.0
},
"step": 1701
},
{
"epoch": 22.0,
"eval_Claim": {
"f1-score": 0.6720596061734965,
"precision": 0.7086419753086419,
"recall": 0.6390688259109312,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9051685393258427,
"precision": 0.8903625110521662,
"recall": 0.9204753199268738,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9997004867090977,
"precision": 0.9995507973347308,
"recall": 0.9998502209241369,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9166278528178853,
"precision": 0.9051882742548755,
"recall": 0.9283602742310837,
"support": 15899.0
},
"eval_accuracy": 0.9148433205057724,
"eval_loss": 0.7163732051849365,
"eval_macro avg": {
"f1-score": 0.8733891212565805,
"precision": 0.8759358894876037,
"recall": 0.8719386602482564,
"support": 36380.0
},
"eval_runtime": 4.9476,
"eval_samples_per_second": 16.371,
"eval_steps_per_second": 2.223,
"eval_weighted avg": {
"f1-score": 0.9132201772820274,
"precision": 0.9122428450081295,
"recall": 0.9148433205057724,
"support": 36380.0
},
"step": 1782
},
{
"epoch": 23.0,
"eval_Claim": {
"f1-score": 0.7022826614861584,
"precision": 0.6750700280112045,
"recall": 0.7317813765182186,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9080485115766262,
"precision": 0.8772901576480614,
"recall": 0.9410420475319927,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9988753936122357,
"precision": 1.0,
"recall": 0.9977533138620535,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9134190823574582,
"precision": 0.9295994789970694,
"recall": 0.8977923139820114,
"support": 15899.0
},
"eval_accuracy": 0.9145409565695437,
"eval_loss": 0.695353627204895,
"eval_macro avg": {
"f1-score": 0.8806564122581195,
"precision": 0.8704899161640838,
"recall": 0.892092262973569,
"support": 36380.0
},
"eval_runtime": 4.9626,
"eval_samples_per_second": 16.322,
"eval_steps_per_second": 2.217,
"eval_weighted avg": {
"f1-score": 0.9157921828580727,
"precision": 0.9177311412832249,
"recall": 0.9145409565695437,
"support": 36380.0
},
"step": 1863
},
{
"epoch": 24.0,
"eval_Claim": {
"f1-score": 0.6945771284068559,
"precision": 0.6463308349311487,
"recall": 0.7506072874493928,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9287416071041802,
"precision": 0.8826677645121449,
"recall": 0.979890310786106,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9992880428673137,
"precision": 1.0,
"recall": 0.9985770987793006,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9030182916923878,
"precision": 0.9339381720430108,
"recall": 0.8740801308258381,
"support": 15899.0
},
"eval_accuracy": 0.9093732820230896,
"eval_loss": 0.7689566016197205,
"eval_macro avg": {
"f1-score": 0.8814062675176844,
"precision": 0.8657341928715762,
"recall": 0.9007887069601594,
"support": 36380.0
},
"eval_runtime": 4.9669,
"eval_samples_per_second": 16.308,
"eval_steps_per_second": 2.215,
"eval_weighted avg": {
"f1-score": 0.911596446033489,
"precision": 0.9160482240413489,
"recall": 0.9093732820230896,
"support": 36380.0
},
"step": 1944
},
{
"epoch": 24.69,
"grad_norm": 0.003183299908414483,
"learning_rate": 1.0123456790123458e-05,
"loss": 0.0045,
"step": 2000
},
{
"epoch": 25.0,
"eval_Claim": {
"f1-score": 0.6596491228070176,
"precision": 0.7481386392811297,
"recall": 0.5898785425101215,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9256909832351607,
"precision": 0.9177897574123989,
"recall": 0.9337294332723949,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9998876614866129,
"precision": 0.9999251048532055,
"recall": 0.9998502209241369,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9174541242455649,
"precision": 0.8901046903649376,
"recall": 0.9465375180828983,
"support": 15899.0
},
"eval_accuracy": 0.9169048927982408,
"eval_loss": 0.7267045378684998,
"eval_macro avg": {
"f1-score": 0.8756704729435891,
"precision": 0.888989547977918,
"recall": 0.867498928697388,
"support": 36380.0
},
"eval_runtime": 4.9577,
"eval_samples_per_second": 16.338,
"eval_steps_per_second": 2.219,
"eval_weighted avg": {
"f1-score": 0.9131990545133639,
"precision": 0.9128010792876335,
"recall": 0.9169048927982408,
"support": 36380.0
},
"step": 2025
},
{
"epoch": 26.0,
"eval_Claim": {
"f1-score": 0.7098703888334996,
"precision": 0.6994106090373281,
"recall": 0.7206477732793523,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9220324134910205,
"precision": 0.8851976450798991,
"recall": 0.9620658135283364,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9995504607777028,
"precision": 1.0,
"recall": 0.9991013255448213,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9171909755322529,
"precision": 0.9268511977393873,
"recall": 0.9077300459148374,
"support": 15899.0
},
"eval_accuracy": 0.9191313908741067,
"eval_loss": 0.6802815794944763,
"eval_macro avg": {
"f1-score": 0.887161059658619,
"precision": 0.8778648629641536,
"recall": 0.8973862395668368,
"support": 36380.0
},
"eval_runtime": 4.9696,
"eval_samples_per_second": 16.299,
"eval_steps_per_second": 2.213,
"eval_weighted avg": {
"f1-score": 0.9195597378864155,
"precision": 0.9203108314716255,
"recall": 0.9191313908741067,
"support": 36380.0
},
"step": 2106
},
{
"epoch": 27.0,
"eval_Claim": {
"f1-score": 0.698643105080467,
"precision": 0.7271731990365666,
"recall": 0.6722672064777327,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9165368350767862,
"precision": 0.8932754880694144,
"recall": 0.9410420475319927,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9991755359016639,
"precision": 1.0,
"recall": 0.9983524301655059,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9206260132186058,
"precision": 0.9127155838536193,
"recall": 0.9286747594188314,
"support": 15899.0
},
"eval_accuracy": 0.9201759208356239,
"eval_loss": 0.6692858934402466,
"eval_macro avg": {
"f1-score": 0.8837453723193808,
"precision": 0.8832910677399001,
"recall": 0.8850841108985157,
"support": 36380.0
},
"eval_runtime": 4.9586,
"eval_samples_per_second": 16.335,
"eval_steps_per_second": 2.218,
"eval_weighted avg": {
"f1-score": 0.9190682641369721,
"precision": 0.918388879571886,
"recall": 0.9201759208356239,
"support": 36380.0
},
"step": 2187
},
{
"epoch": 28.0,
"eval_Claim": {
"f1-score": 0.6998441558441558,
"precision": 0.7188900747065101,
"recall": 0.6817813765182186,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9112884834663626,
"precision": 0.9094219390077378,
"recall": 0.9131627056672761,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9997003745318352,
"precision": 0.9999250767962838,
"recall": 0.9994757732344791,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9211232449297972,
"precision": 0.9139372175097518,
"recall": 0.9284231712686333,
"support": 15899.0
},
"eval_accuracy": 0.9200934579439253,
"eval_loss": 0.6845241189002991,
"eval_macro avg": {
"f1-score": 0.8829890646930376,
"precision": 0.8855435770050709,
"recall": 0.8807107566721517,
"support": 36380.0
},
"eval_runtime": 4.9593,
"eval_samples_per_second": 16.333,
"eval_steps_per_second": 2.218,
"eval_weighted avg": {
"f1-score": 0.9193256433193232,
"precision": 0.9187416037176308,
"recall": 0.9200934579439253,
"support": 36380.0
},
"step": 2268
},
{
"epoch": 29.0,
"eval_Claim": {
"f1-score": 0.6987782896054476,
"precision": 0.6914387633769322,
"recall": 0.7062753036437247,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9134720700985761,
"precision": 0.8771560790912916,
"recall": 0.9529250457038391,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9993630811884155,
"precision": 0.9999250262408157,
"recall": 0.9988017673930952,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9151595913446284,
"precision": 0.9233625712273513,
"recall": 0.907101075539342,
"support": 15899.0
},
"eval_accuracy": 0.916245189664651,
"eval_loss": 0.7010424137115479,
"eval_macro avg": {
"f1-score": 0.8816932580592669,
"precision": 0.8729706099840976,
"recall": 0.8912757980700003,
"support": 36380.0
},
"eval_runtime": 4.9512,
"eval_samples_per_second": 16.36,
"eval_steps_per_second": 2.222,
"eval_weighted avg": {
"f1-score": 0.9165821661881187,
"precision": 0.9171925615027781,
"recall": 0.916245189664651,
"support": 36380.0
},
"step": 2349
},
{
"epoch": 30.0,
"eval_Claim": {
"f1-score": 0.6948434432158727,
"precision": 0.7097318978256281,
"recall": 0.6805668016194332,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9181940098345999,
"precision": 0.8985126859142607,
"recall": 0.9387568555758684,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9995504944560982,
"precision": 0.9999250543356066,
"recall": 0.9991762150827529,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9180584714693073,
"precision": 0.91476208317722,
"recall": 0.9213787030630858,
"support": 15899.0
},
"eval_accuracy": 0.9182792743265531,
"eval_loss": 0.7106152176856995,
"eval_macro avg": {
"f1-score": 0.8826616047439695,
"precision": 0.8807329303131789,
"recall": 0.8849696438352851,
"support": 36380.0
},
"eval_runtime": 4.9623,
"eval_samples_per_second": 16.323,
"eval_steps_per_second": 2.217,
"eval_weighted avg": {
"f1-score": 0.917667523182169,
"precision": 0.9172023898575309,
"recall": 0.9182792743265531,
"support": 36380.0
},
"step": 2430
},
{
"epoch": 30.86,
"grad_norm": 0.0027945609763264656,
"learning_rate": 7.654320987654322e-06,
"loss": 0.0031,
"step": 2500
},
{
"epoch": 31.0,
"eval_Claim": {
"f1-score": 0.7131089189464049,
"precision": 0.7165338238299611,
"recall": 0.7097165991902834,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9162036008001779,
"precision": 0.8918217221981826,
"recall": 0.9419561243144424,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9994380549207659,
"precision": 0.9999250374812594,
"recall": 0.9989515464689583,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9217583110130771,
"precision": 0.9235918161151806,
"recall": 0.9199320711994465,
"support": 15899.0
},
"eval_accuracy": 0.9217152281473338,
"eval_loss": 0.6842674016952515,
"eval_macro avg": {
"f1-score": 0.8876272214201064,
"precision": 0.882968099906146,
"recall": 0.8926390852932827,
"support": 36380.0
},
"eval_runtime": 4.9684,
"eval_samples_per_second": 16.303,
"eval_steps_per_second": 2.214,
"eval_weighted avg": {
"f1-score": 0.9216037183150064,
"precision": 0.9215824169266671,
"recall": 0.9217152281473338,
"support": 36380.0
},
"step": 2511
},
{
"epoch": 32.0,
"eval_Claim": {
"f1-score": 0.6948140387637506,
"precision": 0.720086862106406,
"recall": 0.671255060728745,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9145506734378449,
"precision": 0.8846646732165742,
"recall": 0.946526508226691,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9991005172026085,
"precision": 1.0,
"recall": 0.9982026510896428,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9190726159230096,
"precision": 0.9131946600434647,
"recall": 0.9250267312409586,
"support": 15899.0
},
"eval_accuracy": 0.918719076415613,
"eval_loss": 0.7216777205467224,
"eval_macro avg": {
"f1-score": 0.8818844613318033,
"precision": 0.8794865488416111,
"recall": 0.8852527378215094,
"support": 36380.0
},
"eval_runtime": 4.9621,
"eval_samples_per_second": 16.324,
"eval_steps_per_second": 2.217,
"eval_weighted avg": {
"f1-score": 0.9177224560676552,
"precision": 0.9171181227002353,
"recall": 0.918719076415613,
"support": 36380.0
},
"step": 2592
},
{
"epoch": 33.0,
"eval_Claim": {
"f1-score": 0.7047761512982157,
"precision": 0.7463226974428604,
"recall": 0.6676113360323886,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.921081556386019,
"precision": 0.8873358746293943,
"recall": 0.9574954296160878,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9991380279578758,
"precision": 1.0,
"recall": 0.9982775406275743,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9233112623954739,
"precision": 0.9127842655193608,
"recall": 0.9340839046480911,
"support": 15899.0
},
"eval_accuracy": 0.922869708631116,
"eval_loss": 0.7047613263130188,
"eval_macro avg": {
"f1-score": 0.8870767495093961,
"precision": 0.8866107093979039,
"recall": 0.8893670527310353,
"support": 36380.0
},
"eval_runtime": 4.9399,
"eval_samples_per_second": 16.397,
"eval_steps_per_second": 2.227,
"eval_weighted avg": {
"f1-score": 0.9213341528568705,
"precision": 0.9206619586736987,
"recall": 0.922869708631116,
"support": 36380.0
},
"step": 2673
},
{
"epoch": 34.0,
"eval_Claim": {
"f1-score": 0.704132582598534,
"precision": 0.7408897831433042,
"recall": 0.6708502024291498,
"support": 4940.0
},
"eval_MajorClaim": {
"f1-score": 0.9225707391971923,
"precision": 0.8869675242513707,
"recall": 0.9611517367458866,
"support": 2188.0
},
"eval_O": {
"f1-score": 0.9989880439263896,
"precision": 0.9999249699879952,
"recall": 0.9980528720137797,
"support": 13353.0
},
"eval_Premise": {
"f1-score": 0.9222287974584982,
"precision": 0.9134378084896347,
"recall": 0.9311906409208126,
"support": 15899.0
},
"eval_accuracy": 0.9221825178669598,
"eval_loss": 0.7385492324829102,
"eval_macro avg": {
"f1-score": 0.8869800407951536,
"precision": 0.8853050214680762,
"recall": 0.8903113630274071,
"support": 36380.0
},
"eval_runtime": 4.9565,
"eval_samples_per_second": 16.342,
"eval_steps_per_second": 2.219,
"eval_weighted avg": {
"f1-score": 0.9208082115651994,
"precision": 0.9201601515452535,
"recall": 0.9221825178669598,
"support": 36380.0
},
"step": 2754
}
],
"logging_steps": 500,
"max_steps": 4050,
"num_input_tokens_seen": 0,
"num_train_epochs": 50,
"save_steps": 500,
"total_flos": 4873634251491600.0,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}