{ "best_metric": null, "best_model_checkpoint": null, "epoch": 36.0, "eval_steps": 500, "global_step": 2916, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 1.0, "eval_Claim": { "f1-score": 0.44734338904742493, "precision": 0.5477537437603993, "recall": 0.3780431786862655, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.6504965622612681, "precision": 0.5898856944925528, "recall": 0.7249893571732652, "support": 2349.0 }, "eval_O": { "f1-score": 0.8859784283513097, "precision": 0.8689081979599547, "recall": 0.9037328094302554, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8720120805863504, "precision": 0.859258183929738, "recall": 0.8851502916105877, "support": 13374.0 }, "eval_accuracy": 0.8059953068711373, "eval_loss": 0.47828415036201477, "eval_macro avg": { "f1-score": 0.7139576150615883, "precision": 0.7164514550356611, "recall": 0.7229789092250934, "support": 30257.0 }, "eval_runtime": 4.8422, "eval_samples_per_second": 16.728, "eval_steps_per_second": 2.272, "eval_weighted avg": { "f1-score": 0.7984036588902531, "precision": 0.7967665566250599, "recall": 0.8059953068711373, "support": 30257.0 }, "step": 81 }, { "epoch": 2.0, "eval_Claim": { "f1-score": 0.5787561000934482, "precision": 0.5281409891984082, "recall": 0.6401010564997703, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.7702650787552824, "precision": 0.7017850892544627, "recall": 0.8535547041294168, "support": 2349.0 }, "eval_O": { "f1-score": 0.9095337508698677, "precision": 0.9206077681626081, "recall": 0.8987229862475442, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8736648538675224, "precision": 0.9162905211325401, "recall": 0.8348287722446538, "support": 13374.0 }, "eval_accuracy": 0.8297584030141786, "eval_loss": 0.4241589605808258, "eval_macro avg": { "f1-score": 0.7830549458965301, "precision": 0.7667060919370048, "recall": 0.8068018797803463, "support": 30257.0 }, "eval_runtime": 4.8636, "eval_samples_per_second": 16.654, "eval_steps_per_second": 2.262, "eval_weighted avg": { "f1-score": 0.8352680063880269, "precision": 0.8452350051608073, "recall": 0.8297584030141786, "support": 30257.0 }, "step": 162 }, { "epoch": 3.0, "eval_Claim": { "f1-score": 0.5728465083303793, "precision": 0.5899245558530056, "recall": 0.5567294441892513, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.7862876979822088, "precision": 0.8017699115044248, "recall": 0.7713920817369093, "support": 2349.0 }, "eval_O": { "f1-score": 0.914532821534425, "precision": 0.9048168445341794, "recall": 0.9244597249508841, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8880533114924986, "precision": 0.8843330614665975, "recall": 0.8918049947659639, "support": 13374.0 }, "eval_accuracy": 0.8452258981392735, "eval_loss": 0.41067537665367126, "eval_macro avg": { "f1-score": 0.7904300848348779, "precision": 0.7952110933395518, "recall": 0.7860965614107521, "support": 30257.0 }, "eval_runtime": 4.8684, "eval_samples_per_second": 16.638, "eval_steps_per_second": 2.259, "eval_weighted avg": { "f1-score": 0.8437033615676307, "precision": 0.8424495118392472, "recall": 0.8452258981392735, "support": 30257.0 }, "step": 243 }, { "epoch": 4.0, "eval_Claim": { "f1-score": 0.5865384615384616, "precision": 0.5714596949891068, "recall": 0.6024345429490124, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.7783630470016208, "precision": 0.7425589485890993, "recall": 0.8177948063005535, "support": 2349.0 }, "eval_O": { "f1-score": 0.9108591535529275, "precision": 0.9063411787589963, "recall": 0.9154223968565816, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8868256151612411, "precision": 0.9067823097358962, "recall": 0.8677284282937042, "support": 13374.0 }, "eval_accuracy": 0.8417225765938461, "eval_loss": 0.4545813202857971, "eval_macro avg": { "f1-score": 0.7906465693135627, "precision": 0.7817855330182747, "recall": 0.8008450435999629, "support": 30257.0 }, "eval_runtime": 4.9112, "eval_samples_per_second": 16.493, "eval_steps_per_second": 2.24, "eval_weighted avg": { "f1-score": 0.8432797441676475, "precision": 0.8456313015960876, "recall": 0.8417225765938461, "support": 30257.0 }, "step": 324 }, { "epoch": 5.0, "eval_Claim": { "f1-score": 0.597399608790703, "precision": 0.5985704403965875, "recall": 0.5962333486449242, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.7933532168726033, "precision": 0.7940298507462686, "recall": 0.7926777352064709, "support": 2349.0 }, "eval_O": { "f1-score": 0.9101112629318758, "precision": 0.9042862684251358, "recall": 0.9160117878192534, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8923677591320345, "precision": 0.8961019377214808, "recall": 0.8886645730521908, "support": 13374.0 }, "eval_accuracy": 0.848332617245596, "eval_loss": 0.48704993724823, "eval_macro avg": { "f1-score": 0.7983079619318041, "precision": 0.7982471243223681, "recall": 0.7983968611807098, "support": 30257.0 }, "eval_runtime": 4.8769, "eval_samples_per_second": 16.609, "eval_steps_per_second": 2.256, "eval_weighted avg": { "f1-score": 0.8482045037639816, "precision": 0.8481162489521332, "recall": 0.848332617245596, "support": 30257.0 }, "step": 405 }, { "epoch": 6.0, "eval_Claim": { "f1-score": 0.6060809300245921, "precision": 0.5903745644599303, "recall": 0.6226458429030777, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.7858529221479404, "precision": 0.7963286713286714, "recall": 0.7756492124308216, "support": 2349.0 }, "eval_O": { "f1-score": 0.9135615470059578, "precision": 0.9398504051527115, "recall": 0.8887033398821218, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.895557603686636, "precision": 0.88328121591157, "recall": 0.908180050844923, "support": 13374.0 }, "eval_accuracy": 0.8502495290346036, "eval_loss": 0.5423789024353027, "eval_macro avg": { "f1-score": 0.8002632507162816, "precision": 0.8024587142132208, "recall": 0.7987946115152361, "support": 30257.0 }, "eval_runtime": 4.867, "eval_samples_per_second": 16.643, "eval_steps_per_second": 2.26, "eval_weighted avg": { "f1-score": 0.8514422719925407, "precision": 0.8534139871324166, "recall": 0.8502495290346036, "support": 30257.0 }, "step": 486 }, { "epoch": 6.17, "grad_norm": 4.494631767272949, "learning_rate": 1.7530864197530865e-05, "loss": 0.35, "step": 500 }, { "epoch": 7.0, "eval_Claim": { "f1-score": 0.6040515653775324, "precision": 0.6054453161052146, "recall": 0.6026642168121268, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8052777186635455, "precision": 0.8051063829787234, "recall": 0.8054491272882077, "support": 2349.0 }, "eval_O": { "f1-score": 0.9184720354599332, "precision": 0.9315082331548641, "recall": 0.9057956777996071, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8908606921029282, "precision": 0.8810881965774463, "recall": 0.9008524001794527, "support": 13374.0 }, "eval_accuracy": 0.8521994910268698, "eval_loss": 0.5764245986938477, "eval_macro avg": { "f1-score": 0.8046655029009848, "precision": 0.8057870322040621, "recall": 0.8036903555198486, "support": 30257.0 }, "eval_runtime": 4.8398, "eval_samples_per_second": 16.736, "eval_steps_per_second": 2.273, "eval_weighted avg": { "f1-score": 0.8522343290465387, "precision": 0.8524880574571969, "recall": 0.8521994910268698, "support": 30257.0 }, "step": 567 }, { "epoch": 8.0, "eval_Claim": { "f1-score": 0.597346719254213, "precision": 0.6227261400448543, "recall": 0.5739549839228296, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8047001620745543, "precision": 0.7676845767298028, "recall": 0.8454661558109834, "support": 2349.0 }, "eval_O": { "f1-score": 0.9099175618286286, "precision": 0.9258769700050838, "recall": 0.8944990176817289, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8942491542873952, "precision": 0.8797569092750688, "recall": 0.9092268580828473, "support": 13374.0 }, "eval_accuracy": 0.8510757841160723, "eval_loss": 0.5954988598823547, "eval_macro avg": { "f1-score": 0.8015533993611977, "precision": 0.7990111490137024, "recall": 0.8057867538745973, "support": 30257.0 }, "eval_runtime": 4.8609, "eval_samples_per_second": 16.663, "eval_steps_per_second": 2.263, "eval_weighted avg": { "f1-score": 0.8498442431569896, "precision": 0.8495864475589161, "recall": 0.8510757841160723, "support": 30257.0 }, "step": 648 }, { "epoch": 9.0, "eval_Claim": { "f1-score": 0.6049033474776048, "precision": 0.6213075060532688, "recall": 0.5893431327514929, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.823503227242377, "precision": 0.8628731343283582, "recall": 0.787569178373776, "support": 2349.0 }, "eval_O": { "f1-score": 0.9189658498938837, "precision": 0.9027672479150872, "recall": 0.9357563850687622, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8989367655288193, "precision": 0.8970292606656243, "recall": 0.9008524001794527, "support": 13374.0 }, "eval_accuracy": 0.8589747826949136, "eval_loss": 0.5689579844474792, "eval_macro avg": { "f1-score": 0.8115772975356712, "precision": 0.8209942872405847, "recall": 0.803380274093371, "support": 30257.0 }, "eval_runtime": 4.8593, "eval_samples_per_second": 16.669, "eval_steps_per_second": 2.264, "eval_weighted avg": { "f1-score": 0.8575077142417292, "precision": 0.8566315824374819, "recall": 0.8589747826949136, "support": 30257.0 }, "step": 729 }, { "epoch": 10.0, "eval_Claim": { "f1-score": 0.60162224797219, "precision": 0.6071094480823199, "recall": 0.5962333486449242, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8131676477000218, "precision": 0.8333333333333334, "recall": 0.7939548744146445, "support": 2349.0 }, "eval_O": { "f1-score": 0.9197247706422018, "precision": 0.9338801134062373, "recall": 0.905992141453831, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.9008479864909512, "precision": 0.884834499170693, "recall": 0.9174517720951099, "support": 13374.0 }, "eval_accuracy": 0.8577849753775986, "eval_loss": 0.7088457942008972, "eval_macro avg": { "f1-score": 0.8088406632013413, "precision": 0.8147893484981459, "recall": 0.8034080341521274, "support": 30257.0 }, "eval_runtime": 4.9043, "eval_samples_per_second": 16.516, "eval_steps_per_second": 2.243, "eval_weighted avg": { "f1-score": 0.8573332851434664, "precision": 0.8573728619273148, "recall": 0.8577849753775986, "support": 30257.0 }, "step": 810 }, { "epoch": 11.0, "eval_Claim": { "f1-score": 0.6253176253176252, "precision": 0.6289498141263941, "recall": 0.6217271474506201, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.7947071414672116, "precision": 0.7318996415770609, "recall": 0.8693060876968923, "support": 2349.0 }, "eval_O": { "f1-score": 0.9110249527786062, "precision": 0.9221171261823304, "recall": 0.900196463654224, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8996578818752584, "precision": 0.9047258979206049, "recall": 0.8946463286974727, "support": 13374.0 }, "eval_accuracy": 0.8552731599299336, "eval_loss": 0.7312190532684326, "eval_macro avg": { "f1-score": 0.8076769003596753, "precision": 0.7969231199515976, "recall": 0.8214690068748023, "support": 30257.0 }, "eval_runtime": 4.9029, "eval_samples_per_second": 16.521, "eval_steps_per_second": 2.244, "eval_weighted avg": { "f1-score": 0.8558567785115953, "precision": 0.857475501606145, "recall": 0.8552731599299336, "support": 30257.0 }, "step": 891 }, { "epoch": 12.0, "eval_Claim": { "f1-score": 0.633677135241627, "precision": 0.6395321637426901, "recall": 0.6279283417547084, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8290997764681975, "precision": 0.7931570762052877, "recall": 0.8684546615581098, "support": 2349.0 }, "eval_O": { "f1-score": 0.916678984971668, "precision": 0.9196243203163619, "recall": 0.9137524557956778, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.9029209944129889, "precision": 0.905603610379842, "recall": 0.9002542246149244, "support": 13374.0 }, "eval_accuracy": 0.8631391083055161, "eval_loss": 0.7092207074165344, "eval_macro avg": { "f1-score": 0.8205942227736204, "precision": 0.8144792926610454, "recall": 0.8275974209308551, "support": 30257.0 }, "eval_runtime": 4.8734, "eval_samples_per_second": 16.621, "eval_steps_per_second": 2.257, "eval_weighted avg": { "f1-score": 0.8630744313070275, "precision": 0.8633032778855295, "recall": 0.8631391083055161, "support": 30257.0 }, "step": 972 }, { "epoch": 12.35, "grad_norm": 4.426708698272705, "learning_rate": 1.506172839506173e-05, "loss": 0.063, "step": 1000 }, { "epoch": 13.0, "eval_Claim": { "f1-score": 0.5936837143530521, "precision": 0.6096321393998064, "recall": 0.5785484611851172, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8065005417118094, "precision": 0.821270962047661, "recall": 0.7922520221370796, "support": 2349.0 }, "eval_O": { "f1-score": 0.9133826964170352, "precision": 0.9242683294780247, "recall": 0.9027504911591355, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.901282521069989, "precision": 0.883730957171601, "recall": 0.9195453865709585, "support": 13374.0 }, "eval_accuracy": 0.854942657897346, "eval_loss": 0.7807008624076843, "eval_macro avg": { "f1-score": 0.8037123683879714, "precision": 0.8097255970242733, "recall": 0.7982740902630727, "support": 30257.0 }, "eval_runtime": 4.8517, "eval_samples_per_second": 16.695, "eval_steps_per_second": 2.267, "eval_weighted avg": { "f1-score": 0.8537315976828396, "precision": 0.8530777420132861, "recall": 0.854942657897346, "support": 30257.0 }, "step": 1053 }, { "epoch": 14.0, "eval_Claim": { "f1-score": 0.6335702820454024, "precision": 0.632554945054945, "recall": 0.6345888837850253, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8183556405353729, "precision": 0.816793893129771, "recall": 0.8199233716475096, "support": 2349.0 }, "eval_O": { "f1-score": 0.9172396976347231, "precision": 0.9107990314769976, "recall": 0.9237721021611002, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8981941309255079, "precision": 0.9039073148568832, "recall": 0.8925527142216241, "support": 13374.0 }, "eval_accuracy": 0.8602967908252636, "eval_loss": 0.7879041433334351, "eval_macro avg": { "f1-score": 0.8168399377852517, "precision": 0.8160137961296492, "recall": 0.8177092679538147, "support": 30257.0 }, "eval_runtime": 4.864, "eval_samples_per_second": 16.653, "eval_steps_per_second": 2.262, "eval_weighted avg": { "f1-score": 0.8603242501425289, "precision": 0.8604152313535001, "recall": 0.8602967908252636, "support": 30257.0 }, "step": 1134 }, { "epoch": 15.0, "eval_Claim": { "f1-score": 0.6212592513139547, "precision": 0.5828134433487623, "recall": 0.6651355075792375, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.7988422575976845, "precision": 0.7765273311897106, "recall": 0.822477650063857, "support": 2349.0 }, "eval_O": { "f1-score": 0.9143625765954267, "precision": 0.9276255938542404, "recall": 0.9014734774066797, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8928883984627678, "precision": 0.9090416053304409, "recall": 0.8772992373261552, "support": 13374.0 }, "eval_accuracy": 0.8506461314737086, "eval_loss": 0.8185116052627563, "eval_macro avg": { "f1-score": 0.8068381209924584, "precision": 0.7990019934307886, "recall": 0.8165964680939823, "support": 30257.0 }, "eval_runtime": 4.8575, "eval_samples_per_second": 16.675, "eval_steps_per_second": 2.265, "eval_weighted avg": { "f1-score": 0.8537245501570023, "precision": 0.8580620487302317, "recall": 0.8506461314737086, "support": 30257.0 }, "step": 1215 }, { "epoch": 16.0, "eval_Claim": { "f1-score": 0.620457750298297, "precision": 0.5878725590955807, "recall": 0.6568672485071199, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8217720482913852, "precision": 0.7911741528762806, "recall": 0.8548318433375904, "support": 2349.0 }, "eval_O": { "f1-score": 0.9171081176529942, "precision": 0.9427445285758739, "recall": 0.8928290766208251, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8952495580546884, "precision": 0.900703852266707, "recall": 0.8898609241812472, "support": 13374.0 }, "eval_accuracy": 0.8546121558647586, "eval_loss": 0.8933361768722534, "eval_macro avg": { "f1-score": 0.8136468685743412, "precision": 0.8056237732036106, "recall": 0.8235972731616956, "support": 30257.0 }, "eval_runtime": 4.839, "eval_samples_per_second": 16.739, "eval_steps_per_second": 2.273, "eval_weighted avg": { "f1-score": 0.8573567707758909, "precision": 0.8613285464033407, "recall": 0.8546121558647586, "support": 30257.0 }, "step": 1296 }, { "epoch": 17.0, "eval_Claim": { "f1-score": 0.577664702731655, "precision": 0.6926782273603083, "recall": 0.49540652273771246, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8222534645511148, "precision": 0.7783269961977186, "recall": 0.8714346530438485, "support": 2349.0 }, "eval_O": { "f1-score": 0.9173265135494024, "precision": 0.9185462424899045, "recall": 0.9161100196463654, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.901492752578063, "precision": 0.8705431754874652, "recall": 0.9347240915208613, "support": 13374.0 }, "eval_accuracy": 0.8603298410285223, "eval_loss": 0.8753499984741211, "eval_macro avg": { "f1-score": 0.8046843583525589, "precision": 0.8150236603838491, "recall": 0.8044188217371968, "support": 30257.0 }, "eval_runtime": 4.8417, "eval_samples_per_second": 16.73, "eval_steps_per_second": 2.272, "eval_weighted avg": { "f1-score": 0.8540692561997597, "precision": 0.8539397922302876, "recall": 0.8603298410285223, "support": 30257.0 }, "step": 1377 }, { "epoch": 18.0, "eval_Claim": { "f1-score": 0.6204532891100056, "precision": 0.5981667021956939, "recall": 0.6444648598989435, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8362068965517241, "precision": 0.807372175980975, "recall": 0.8671775223499362, "support": 2349.0 }, "eval_O": { "f1-score": 0.9045747829039926, "precision": 0.8936068244991853, "recall": 0.9158153241650294, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8850831280788178, "precision": 0.9118953211736717, "recall": 0.8598026020637057, "support": 13374.0 }, "eval_accuracy": 0.8482334666358198, "eval_loss": 0.9685499668121338, "eval_macro avg": { "f1-score": 0.811579524161135, "precision": 0.8027602559623814, "recall": 0.8218150771194037, "support": 30257.0 }, "eval_runtime": 4.8521, "eval_samples_per_second": 16.694, "eval_steps_per_second": 2.267, "eval_weighted avg": { "f1-score": 0.8497662248627994, "precision": 0.8524817583209754, "recall": 0.8482334666358198, "support": 30257.0 }, "step": 1458 }, { "epoch": 18.52, "grad_norm": 0.44517388939857483, "learning_rate": 1.2592592592592593e-05, "loss": 0.0216, "step": 1500 }, { "epoch": 19.0, "eval_Claim": { "f1-score": 0.6178126034654013, "precision": 0.5946462715105163, "recall": 0.6428571428571429, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8151062155782848, "precision": 0.7575868372943327, "recall": 0.8820774797786292, "support": 2349.0 }, "eval_O": { "f1-score": 0.9146588828971454, "precision": 0.9213595136051032, "recall": 0.9080550098231827, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8898914207065302, "precision": 0.9104991394148021, "recall": 0.870195902497383, "support": 13374.0 }, "eval_accuracy": 0.8511418845225898, "eval_loss": 0.9391252398490906, "eval_macro avg": { "f1-score": 0.8093672806618404, "precision": 0.7960229404561886, "recall": 0.8257963837390844, "support": 30257.0 }, "eval_runtime": 4.8661, "eval_samples_per_second": 16.646, "eval_steps_per_second": 2.261, "eval_weighted avg": { "f1-score": 0.853266214902463, "precision": 0.8568303759789367, "recall": 0.8511418845225898, "support": 30257.0 }, "step": 1539 }, { "epoch": 20.0, "eval_Claim": { "f1-score": 0.6097087378640776, "precision": 0.6064530788457169, "recall": 0.6129995406522738, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8268961376826096, "precision": 0.7802114803625377, "recall": 0.8795232013622818, "support": 2349.0 }, "eval_O": { "f1-score": 0.9215637325877477, "precision": 0.9370494466443293, "recall": 0.906581532416503, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.896644596566042, "precision": 0.8971479901190209, "recall": 0.8961417676087932, "support": 13374.0 }, "eval_accuracy": 0.8576197243613048, "eval_loss": 0.8915661573410034, "eval_macro avg": { "f1-score": 0.8137033011751192, "precision": 0.8052154989929012, "recall": 0.8238115105099629, "support": 30257.0 }, "eval_runtime": 4.864, "eval_samples_per_second": 16.653, "eval_steps_per_second": 2.261, "eval_weighted avg": { "f1-score": 0.8583235120565212, "precision": 0.8596633525913644, "recall": 0.8576197243613048, "support": 30257.0 }, "step": 1620 }, { "epoch": 21.0, "eval_Claim": { "f1-score": 0.6211416490486258, "precision": 0.5754014884449667, "recall": 0.6747818098300413, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8162756902636495, "precision": 0.7965964343598055, "recall": 0.8369518944231588, "support": 2349.0 }, "eval_O": { "f1-score": 0.911611088790679, "precision": 0.9325935059597205, "recall": 0.8915520628683693, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8885090218423553, "precision": 0.9030190718863408, "recall": 0.8744579033946464, "support": 13374.0 }, "eval_accuracy": 0.8485639686684073, "eval_loss": 0.9740728735923767, "eval_macro avg": { "f1-score": 0.8093843624863275, "precision": 0.8019026251627084, "recall": 0.8194359176290539, "support": 30257.0 }, "eval_runtime": 4.884, "eval_samples_per_second": 16.585, "eval_steps_per_second": 2.252, "eval_weighted avg": { "f1-score": 0.852199586158436, "precision": 0.8575629461968616, "recall": 0.8485639686684073, "support": 30257.0 }, "step": 1701 }, { "epoch": 22.0, "eval_Claim": { "f1-score": 0.6145704312577412, "precision": 0.6028274795670422, "recall": 0.6267799724391364, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8148905706688483, "precision": 0.784251968503937, "recall": 0.8480204342273308, "support": 2349.0 }, "eval_O": { "f1-score": 0.9175837613108339, "precision": 0.9137847053093034, "recall": 0.9214145383104125, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8915167877105593, "precision": 0.9070019342359767, "recall": 0.876551517870495, "support": 13374.0 }, "eval_accuracy": 0.853488448953961, "eval_loss": 0.9675940871238708, "eval_macro avg": { "f1-score": 0.8096403877369956, "precision": 0.8019665219040648, "recall": 0.8181916157118437, "support": 30257.0 }, "eval_runtime": 4.8521, "eval_samples_per_second": 16.694, "eval_steps_per_second": 2.267, "eval_weighted avg": { "f1-score": 0.8544854353433136, "precision": 0.8559834381654265, "recall": 0.853488448953961, "support": 30257.0 }, "step": 1782 }, { "epoch": 23.0, "eval_Claim": { "f1-score": 0.5983347015362965, "precision": 0.6113108075724898, "recall": 0.5858980248047773, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8269188475918188, "precision": 0.8553230209281165, "recall": 0.800340570455513, "support": 2349.0 }, "eval_O": { "f1-score": 0.9219349129879069, "precision": 0.9227514268844715, "recall": 0.9211198428290767, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8949737988043397, "precision": 0.8835616438356164, "recall": 0.9066846119336025, "support": 13374.0 }, "eval_accuracy": 0.8571239713124236, "eval_loss": 0.9897539019584656, "eval_macro avg": { "f1-score": 0.8105405652300904, "precision": 0.8182367248051736, "recall": 0.8035107625057423, "support": 30257.0 }, "eval_runtime": 4.8416, "eval_samples_per_second": 16.73, "eval_steps_per_second": 2.272, "eval_weighted avg": { "f1-score": 0.8560749136698399, "precision": 0.8553777302003576, "recall": 0.8571239713124236, "support": 30257.0 }, "step": 1863 }, { "epoch": 24.0, "eval_Claim": { "f1-score": 0.6090506975161619, "precision": 0.6016132646202106, "recall": 0.6166743224621039, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.7960317460317461, "precision": 0.7454477889260498, "recall": 0.853980417198808, "support": 2349.0 }, "eval_O": { "f1-score": 0.9115434760806339, "precision": 0.9310592091784471, "recall": 0.8928290766208251, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8995695302264645, "precision": 0.9006821077880219, "recall": 0.89845969792134, "support": 13374.0 }, "eval_accuracy": 0.852563043262716, "eval_loss": 0.9720127582550049, "eval_macro avg": { "f1-score": 0.8040488624637516, "precision": 0.7947005926281823, "recall": 0.8154858785507693, "support": 30257.0 }, "eval_runtime": 4.8503, "eval_samples_per_second": 16.7, "eval_steps_per_second": 2.268, "eval_weighted avg": { "f1-score": 0.8537541987693271, "precision": 0.8558147294621835, "recall": 0.852563043262716, "support": 30257.0 }, "step": 1944 }, { "epoch": 24.69, "grad_norm": 3.446370840072632, "learning_rate": 1.0123456790123458e-05, "loss": 0.0116, "step": 2000 }, { "epoch": 25.0, "eval_Claim": { "f1-score": 0.6102259215219976, "precision": 0.6326429980276134, "recall": 0.5893431327514929, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8064243448858834, "precision": 0.8006714225765842, "recall": 0.8122605363984674, "support": 2349.0 }, "eval_O": { "f1-score": 0.9169322709163347, "precision": 0.9298989898989899, "recall": 0.9043222003929273, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8993844349992672, "precision": 0.8818077309958328, "recall": 0.917676087931808, "support": 13374.0 }, "eval_accuracy": 0.8577519251743398, "eval_loss": 1.0132465362548828, "eval_macro avg": { "f1-score": 0.8082417430808707, "precision": 0.811255285374755, "recall": 0.8059004893686739, "support": 30257.0 }, "eval_runtime": 4.8662, "eval_samples_per_second": 16.645, "eval_steps_per_second": 2.26, "eval_weighted avg": { "f1-score": 0.8564613940592988, "precision": 0.8558341242209939, "recall": 0.8577519251743398, "support": 30257.0 }, "step": 2025 }, { "epoch": 26.0, "eval_Claim": { "f1-score": 0.6123184326089406, "precision": 0.6006185111552905, "recall": 0.6244832338079926, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.7939849624060151, "precision": 0.8013876843018214, "recall": 0.7867177522349936, "support": 2349.0 }, "eval_O": { "f1-score": 0.9205986178093769, "precision": 0.93204469948656, "recall": 0.9094302554027505, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8996091568955891, "precision": 0.8957082499444073, "recall": 0.9035441902198296, "support": 13374.0 }, "eval_accuracy": 0.8562977162309549, "eval_loss": 1.0289610624313354, "eval_macro avg": { "f1-score": 0.8066277924299805, "precision": 0.8074397862220198, "recall": 0.8060438579163915, "support": 30257.0 }, "eval_runtime": 4.8488, "eval_samples_per_second": 16.705, "eval_steps_per_second": 2.269, "eval_weighted avg": { "f1-score": 0.8571296534981037, "precision": 0.8581475309358099, "recall": 0.8562977162309549, "support": 30257.0 }, "step": 2106 }, { "epoch": 27.0, "eval_Claim": { "f1-score": 0.6028743397617001, "precision": 0.6480063374702931, "recall": 0.5636196600826826, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8248976954555244, "precision": 0.8347863993025283, "recall": 0.8152405278842061, "support": 2349.0 }, "eval_O": { "f1-score": 0.9177434297137843, "precision": 0.9111240197502178, "recall": 0.9244597249508841, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8988648469931302, "precision": 0.883512674225464, "recall": 0.9147599820547331, "support": 13374.0 }, "eval_accuracy": 0.8597679875731236, "eval_loss": 1.0470740795135498, "eval_macro avg": { "f1-score": 0.8110950779810349, "precision": 0.8193573576871258, "recall": 0.8045199737431264, "support": 30257.0 }, "eval_runtime": 4.8822, "eval_samples_per_second": 16.591, "eval_steps_per_second": 2.253, "eval_weighted avg": { "f1-score": 0.8568809247486504, "precision": 0.8551301805022264, "recall": 0.8597679875731236, "support": 30257.0 }, "step": 2187 }, { "epoch": 28.0, "eval_Claim": { "f1-score": 0.6148011951275568, "precision": 0.6152253909843606, "recall": 0.6143775838309601, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.80912003014886, "precision": 0.7258282623394186, "recall": 0.9140059599829715, "support": 2349.0 }, "eval_O": { "f1-score": 0.9129290270324997, "precision": 0.9418216001671192, "recall": 0.8857563850687623, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.897985122051512, "precision": 0.8978844284966734, "recall": 0.8980858381935098, "support": 13374.0 }, "eval_accuracy": 0.8543477542386886, "eval_loss": 1.0588206052780151, "eval_macro avg": { "f1-score": 0.8087088435901071, "precision": 0.7951899204968929, "recall": 0.8280564417690509, "support": 30257.0 }, "eval_runtime": 4.8745, "eval_samples_per_second": 16.617, "eval_steps_per_second": 2.257, "eval_weighted avg": { "f1-score": 0.8553636471531488, "precision": 0.8586347680535737, "recall": 0.8543477542386886, "support": 30257.0 }, "step": 2268 }, { "epoch": 29.0, "eval_Claim": { "f1-score": 0.6258845012554212, "precision": 0.6220508166969148, "recall": 0.6297657326596233, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8182007064201122, "precision": 0.7991071428571429, "recall": 0.8382290336313325, "support": 2349.0 }, "eval_O": { "f1-score": 0.9212816455696202, "precision": 0.9275189167662286, "recall": 0.9151277013752456, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8977727868238817, "precision": 0.8988831421932388, "recall": 0.8966651712277554, "support": 13374.0 }, "eval_accuracy": 0.8599332385894173, "eval_loss": 1.013275384902954, "eval_macro avg": { "f1-score": 0.8157849100172588, "precision": 0.8118900046283812, "recall": 0.8199469097234893, "support": 30257.0 }, "eval_runtime": 4.8916, "eval_samples_per_second": 16.559, "eval_steps_per_second": 2.249, "eval_weighted avg": { "f1-score": 0.860379911449525, "precision": 0.8609352431120858, "recall": 0.8599332385894173, "support": 30257.0 }, "step": 2349 }, { "epoch": 30.0, "eval_Claim": { "f1-score": 0.6191074795725958, "precision": 0.5691448382126348, "recall": 0.6786862655029857, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.802213568354171, "precision": 0.7735177865612648, "recall": 0.8331204767986378, "support": 2349.0 }, "eval_O": { "f1-score": 0.9179413083553385, "precision": 0.9313517338995047, "recall": 0.9049115913555992, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8899992313014068, "precision": 0.9156912369503322, "recall": 0.8657095857634216, "support": 13374.0 }, "eval_accuracy": 0.8494563241563936, "eval_loss": 1.0923724174499512, "eval_macro avg": { "f1-score": 0.8073153968958781, "precision": 0.797426398905934, "recall": 0.8206069798551611, "support": 30257.0 }, "eval_runtime": 4.9081, "eval_samples_per_second": 16.503, "eval_steps_per_second": 2.241, "eval_weighted avg": { "f1-score": 0.853603657884304, "precision": 0.8600543728816776, "recall": 0.8494563241563936, "support": 30257.0 }, "step": 2430 }, { "epoch": 30.86, "grad_norm": 1.9078470468521118, "learning_rate": 7.654320987654322e-06, "loss": 0.0059, "step": 2500 }, { "epoch": 31.0, "eval_Claim": { "f1-score": 0.6110219368646335, "precision": 0.5720296533760769, "recall": 0.655718879191548, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8156672723601858, "precision": 0.7757296466973886, "recall": 0.8599404001702853, "support": 2349.0 }, "eval_O": { "f1-score": 0.9190938511326862, "precision": 0.9318525996971226, "recall": 0.906679764243615, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.890513183575064, "precision": 0.9120482872148624, "recall": 0.8699715866606849, "support": 13374.0 }, "eval_accuracy": 0.850712231880226, "eval_loss": 1.0502196550369263, "eval_macro avg": { "f1-score": 0.8090740609831424, "precision": 0.7979150467463627, "recall": 0.8230776575665333, "support": 30257.0 }, "eval_runtime": 4.8419, "eval_samples_per_second": 16.729, "eval_steps_per_second": 2.272, "eval_weighted avg": { "f1-score": 0.8540995689442556, "precision": 0.8591995012400396, "recall": 0.850712231880226, "support": 30257.0 }, "step": 2511 }, { "epoch": 32.0, "eval_Claim": { "f1-score": 0.5963855421686747, "precision": 0.6272174353775976, "recall": 0.5684428112080845, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8180682764363031, "precision": 0.8004073319755601, "recall": 0.8365261813537676, "support": 2349.0 }, "eval_O": { "f1-score": 0.9205133306804616, "precision": 0.9323861346231358, "recall": 0.9089390962671906, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.9026587563172929, "precision": 0.8845822566752799, "recall": 0.9214894571556752, "support": 13374.0 }, "eval_accuracy": 0.8598671381828998, "eval_loss": 1.0191655158996582, "eval_macro avg": { "f1-score": 0.8094064764006831, "precision": 0.8111482896628934, "recall": 0.8088493864961794, "support": 30257.0 }, "eval_runtime": 4.8429, "eval_samples_per_second": 16.725, "eval_steps_per_second": 2.271, "eval_weighted avg": { "f1-score": 0.8580258765001771, "precision": 0.8570960600087044, "recall": 0.8598671381828998, "support": 30257.0 }, "step": 2592 }, { "epoch": 33.0, "eval_Claim": { "f1-score": 0.6162361623616236, "precision": 0.618805002315887, "recall": 0.6136885622416169, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8282912580847068, "precision": 0.8121931260229133, "recall": 0.8450404427415922, "support": 2349.0 }, "eval_O": { "f1-score": 0.9207272010641968, "precision": 0.9235939507759218, "recall": 0.9178781925343812, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8987739234449761, "precision": 0.8986395574824338, "recall": 0.8989083295947361, "support": 13374.0 }, "eval_accuracy": 0.8600654394024523, "eval_loss": 1.0603935718536377, "eval_macro avg": { "f1-score": 0.8160071362388759, "precision": 0.813307909149289, "recall": 0.8188788817780815, "support": 30257.0 }, "eval_runtime": 4.8566, "eval_samples_per_second": 16.678, "eval_steps_per_second": 2.265, "eval_weighted avg": { "f1-score": 0.8600308614585094, "precision": 0.8600558711630416, "recall": 0.8600654394024523, "support": 30257.0 }, "step": 2673 }, { "epoch": 34.0, "eval_Claim": { "f1-score": 0.6198869304955105, "precision": 0.5991000642811227, "recall": 0.6421681212677998, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8214212991244145, "precision": 0.7872755659640905, "recall": 0.8586632609621115, "support": 2349.0 }, "eval_O": { "f1-score": 0.9196205691462807, "precision": 0.9350253807106599, "recall": 0.9047151277013753, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.897408858089786, "precision": 0.9040825618455001, "recall": 0.8908329594736055, "support": 13374.0 }, "eval_accuracy": 0.8572231219221998, "eval_loss": 1.002353310585022, "eval_macro avg": { "f1-score": 0.8145844142139979, "precision": 0.8063708932003434, "recall": 0.8240948673512231, "support": 30257.0 }, "eval_runtime": 4.8525, "eval_samples_per_second": 16.692, "eval_steps_per_second": 2.267, "eval_weighted avg": { "f1-score": 0.8590471556672056, "precision": 0.8615378438736785, "recall": 0.8572231219221998, "support": 30257.0 }, "step": 2754 }, { "epoch": 35.0, "eval_Claim": { "f1-score": 0.6195019781242729, "precision": 0.6278301886792453, "recall": 0.6113918236104732, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8297086581862947, "precision": 0.8007920792079208, "recall": 0.8607918263090677, "support": 2349.0 }, "eval_O": { "f1-score": 0.9212497535974769, "precision": 0.9243473101265823, "recall": 0.9181728880157171, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.9000523286237572, "precision": 0.8998505231689088, "recall": 0.9002542246149244, "support": 13374.0 }, "eval_accuracy": 0.8616518491588723, "eval_loss": 1.005292296409607, "eval_macro avg": { "f1-score": 0.8176281796329504, "precision": 0.8132050252956643, "recall": 0.8226526906375455, "support": 30257.0 }, "eval_runtime": 4.883, "eval_samples_per_second": 16.588, "eval_steps_per_second": 2.253, "eval_weighted avg": { "f1-score": 0.8613517396129535, "precision": 0.8612582129595939, "recall": 0.8616518491588723, "support": 30257.0 }, "step": 2835 }, { "epoch": 36.0, "eval_Claim": { "f1-score": 0.6118366866690199, "precision": 0.6272617611580217, "recall": 0.5971520440973818, "support": 4354.0 }, "eval_MajorClaim": { "f1-score": 0.8270394325057375, "precision": 0.8109656301145662, "recall": 0.8437633035334184, "support": 2349.0 }, "eval_O": { "f1-score": 0.9219198416625434, "precision": 0.9288135593220339, "recall": 0.9151277013752456, "support": 10180.0 }, "eval_Premise": { "f1-score": 0.8994520953650229, "precision": 0.8907464437600822, "recall": 0.908329594736055, "support": 13374.0 }, "eval_accuracy": 0.8608255940774036, "eval_loss": 1.0221480131149292, "eval_macro avg": { "f1-score": 0.8150620140505809, "precision": 0.814446848588676, "recall": 0.8160931609355251, "support": 30257.0 }, "eval_runtime": 4.8502, "eval_samples_per_second": 16.7, "eval_steps_per_second": 2.268, "eval_weighted avg": { "f1-score": 0.8600016152377763, "precision": 0.8594447878496476, "recall": 0.8608255940774036, "support": 30257.0 }, "step": 2916 } ], "logging_steps": 500, "max_steps": 4050, "num_input_tokens_seen": 0, "num_train_epochs": 50, "save_steps": 500, "total_flos": 5160318619226400.0, "train_batch_size": 4, "trial_name": null, "trial_params": null }