{ "best_metric": 0.4725272059440613, "best_model_checkpoint": "longformer-simple/checkpoint-324", "epoch": 18.0, "eval_steps": 500, "global_step": 1458, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 1.0, "eval_Claim": { "f1-score": 0.46242900830056793, "precision": 0.43257049448304047, "recall": 0.496715157203191, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.5626477541371159, "precision": 0.6522533495736906, "recall": 0.49468822170900695, "support": 2165.0 }, "eval_O": { "f1-score": 0.8691687871077185, "precision": 0.911293908403735, "recall": 0.8307661126874747, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8615178272046495, "precision": 0.8377046804810897, "recall": 0.88672444205844, "support": 13039.0 }, "eval_accuracy": 0.7823004022635849, "eval_loss": 0.5488611459732056, "eval_macro avg": { "f1-score": 0.6889408441875129, "precision": 0.708455608235389, "recall": 0.6772234834145282, "support": 29334.0 }, "eval_runtime": 5.8698, "eval_samples_per_second": 13.629, "eval_steps_per_second": 1.704, "eval_weighted avg": { "f1-score": 0.7840489998358311, "precision": 0.7899101236188295, "recall": 0.7823004022635849, "support": 29334.0 }, "step": 81 }, { "epoch": 2.0, "eval_Claim": { "f1-score": 0.5457327068079991, "precision": 0.5107383923092657, "recall": 0.5858751759737213, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.6722085773663329, "precision": 0.5687340153452686, "recall": 0.8217090069284064, "support": 2165.0 }, "eval_O": { "f1-score": 0.8773589977955805, "precision": 0.9344863131370977, "recall": 0.8268139440616133, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8678243902439025, "precision": 0.8834419195931988, "recall": 0.852749443975765, "support": 13039.0 }, "eval_accuracy": 0.8029590236585532, "eval_loss": 0.5075464248657227, "eval_macro avg": { "f1-score": 0.7407811680534537, "precision": 0.7243501600962077, "recall": 0.7717868927348766, "support": 29334.0 }, "eval_runtime": 5.9473, "eval_samples_per_second": 13.451, "eval_steps_per_second": 1.681, "eval_weighted avg": { "f1-score": 0.8097969994222006, "precision": 0.8232353684753937, "recall": 0.8029590236585532, "support": 29334.0 }, "step": 162 }, { "epoch": 3.0, "eval_Claim": { "f1-score": 0.4978109708987896, "precision": 0.5516552511415526, "recall": 0.45354293758798686, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7357876283496118, "precision": 0.8036105032822757, "recall": 0.6785219399538106, "support": 2165.0 }, "eval_O": { "f1-score": 0.8891011355194736, "precision": 0.9331699710403207, "recall": 0.8490068909606809, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8770979581655561, "precision": 0.8191560170394037, "recall": 0.9438607255157604, "support": 13039.0 }, "eval_accuracy": 0.821129065248517, "eval_loss": 0.5219377279281616, "eval_macro avg": { "f1-score": 0.7499494232333578, "precision": 0.7768979356258882, "recall": 0.7312331235045597, "support": 29334.0 }, "eval_runtime": 5.8364, "eval_samples_per_second": 13.707, "eval_steps_per_second": 1.713, "eval_weighted avg": { "f1-score": 0.815598992812927, "precision": 0.8174973750724106, "recall": 0.821129065248517, "support": 29334.0 }, "step": 243 }, { "epoch": 4.0, "eval_Claim": { "f1-score": 0.5811481218993623, "precision": 0.5851569933396765, "recall": 0.5771938057250118, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.761714412613813, "precision": 0.7335329341317365, "recall": 0.792147806004619, "support": 2165.0 }, "eval_O": { "f1-score": 0.8977645611156685, "precision": 0.9085720215857203, "recall": 0.8872111876773409, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8891009734682191, "precision": 0.8851474612344178, "recall": 0.8930899608865711, "support": 13039.0 }, "eval_accuracy": 0.8377650507943001, "eval_loss": 0.4725272059440613, "eval_macro avg": { "f1-score": 0.7824320172742658, "precision": 0.7781023525728877, "recall": 0.7874106900733857, "support": 29334.0 }, "eval_runtime": 5.8763, "eval_samples_per_second": 13.614, "eval_steps_per_second": 1.702, "eval_weighted avg": { "f1-score": 0.8378705011585708, "precision": 0.8382513248807655, "recall": 0.8377650507943001, "support": 29334.0 }, "step": 324 }, { "epoch": 5.0, "eval_Claim": { "f1-score": 0.5790953831731895, "precision": 0.5784176029962547, "recall": 0.5797747536367902, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7450592885375493, "precision": 0.8008497079129049, "recall": 0.6965357967667436, "support": 2165.0 }, "eval_O": { "f1-score": 0.8897467572575665, "precision": 0.9040794979079498, "recall": 0.8758613700851237, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8861129867206842, "precision": 0.8672442910639547, "recall": 0.905820998542833, "support": 13039.0 }, "eval_accuracy": 0.8329242517215518, "eval_loss": 0.5538631081581116, "eval_macro avg": { "f1-score": 0.7750036039222473, "precision": 0.787647774970266, "recall": 0.7644982297578727, "support": 29334.0 }, "eval_runtime": 5.8534, "eval_samples_per_second": 13.667, "eval_steps_per_second": 1.708, "eval_weighted avg": { "f1-score": 0.8323176558681596, "precision": 0.8327711951367025, "recall": 0.8329242517215518, "support": 29334.0 }, "step": 405 }, { "epoch": 6.0, "eval_Claim": { "f1-score": 0.5991994252283692, "precision": 0.5325670498084292, "recall": 0.6848897231346786, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7586672669968483, "precision": 0.7400087834870444, "recall": 0.7782909930715936, "support": 2165.0 }, "eval_O": { "f1-score": 0.8972759409440632, "precision": 0.9212211784799317, "recall": 0.8745439805431698, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8764605695726225, "precision": 0.9062909567496723, "recall": 0.8485313290896541, "support": 13039.0 }, "eval_accuracy": 0.8283220835890094, "eval_loss": 0.5790187120437622, "eval_macro avg": { "f1-score": 0.7829008006854759, "precision": 0.7750219921312693, "recall": 0.796564006459774, "support": 29334.0 }, "eval_runtime": 5.8468, "eval_samples_per_second": 13.683, "eval_steps_per_second": 1.71, "eval_weighted avg": { "f1-score": 0.8344852708551485, "precision": 0.8447418748493872, "recall": 0.8283220835890094, "support": 29334.0 }, "step": 486 }, { "epoch": 6.17, "grad_norm": 6.636661529541016, "learning_rate": 1.3827160493827162e-05, "loss": 0.3561, "step": 500 }, { "epoch": 7.0, "eval_Claim": { "f1-score": 0.5628712871287128, "precision": 0.5955997904662127, "recall": 0.5335523228531206, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7734199350487135, "precision": 0.8422198041349293, "recall": 0.7150115473441109, "support": 2165.0 }, "eval_O": { "f1-score": 0.9010954940057875, "precision": 0.9193378321383383, "recall": 0.8835630320226996, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8842947894099071, "precision": 0.848316189939411, "recall": 0.9234603880665695, "support": 13039.0 }, "eval_accuracy": 0.838003681734506, "eval_loss": 0.7058164477348328, "eval_macro avg": { "f1-score": 0.7804203763982802, "precision": 0.8013684041697228, "recall": 0.7638968225716252, "support": 29334.0 }, "eval_runtime": 5.8534, "eval_samples_per_second": 13.667, "eval_steps_per_second": 1.708, "eval_weighted avg": { "f1-score": 0.835063123988816, "precision": 0.8350403187795808, "recall": 0.838003681734506, "support": 29334.0 }, "step": 567 }, { "epoch": 8.0, "eval_Claim": { "f1-score": 0.5638466220328667, "precision": 0.5858841386288894, "recall": 0.5434068512435476, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7606318347509112, "precision": 0.8025641025641026, "recall": 0.7228637413394919, "support": 2165.0 }, "eval_O": { "f1-score": 0.896161170692593, "precision": 0.8864763037874281, "recall": 0.9060599918929875, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8909187386294724, "precision": 0.8807043836642937, "recall": 0.9013728046629342, "support": 13039.0 }, "eval_accuracy": 0.8377650507943001, "eval_loss": 0.6875935196876526, "eval_macro avg": { "f1-score": 0.7778895915264609, "precision": 0.7889072321611785, "recall": 0.7684258472847404, "support": 29334.0 }, "eval_runtime": 5.8627, "eval_samples_per_second": 13.646, "eval_steps_per_second": 1.706, "eval_weighted avg": { "f1-score": 0.8355454452418354, "precision": 0.8340438435010799, "recall": 0.8377650507943001, "support": 29334.0 }, "step": 648 }, { "epoch": 9.0, "eval_Claim": { "f1-score": 0.5931978798586572, "precision": 0.5602836879432624, "recall": 0.6302205537306429, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7739087788131438, "precision": 0.8248823836905385, "recall": 0.7288683602771363, "support": 2165.0 }, "eval_O": { "f1-score": 0.8940459939094823, "precision": 0.927653083460449, "recall": 0.8627888123226591, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8843249773482331, "precision": 0.8708454160160607, "recall": 0.8982283917478334, "support": 13039.0 }, "eval_accuracy": 0.8348673893775141, "eval_loss": 0.6963120698928833, "eval_macro avg": { "f1-score": 0.786369407482379, "precision": 0.7959161427775777, "recall": 0.7800265295195679, "support": 29334.0 }, "eval_runtime": 5.8671, "eval_samples_per_second": 13.635, "eval_steps_per_second": 1.704, "eval_weighted avg": { "f1-score": 0.8371473756606818, "precision": 0.8414411074427396, "recall": 0.8348673893775141, "support": 29334.0 }, "step": 729 }, { "epoch": 10.0, "eval_Claim": { "f1-score": 0.5676917638741604, "precision": 0.5701775147928994, "recall": 0.5652275926794932, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7616811458088754, "precision": 0.7745940783190067, "recall": 0.7491916859122402, "support": 2165.0 }, "eval_O": { "f1-score": 0.9002143118498772, "precision": 0.9296124365756234, "recall": 0.8726185650587759, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8900003732596766, "precision": 0.8669284467713787, "recall": 0.9143339213129841, "support": 13039.0 }, "eval_accuracy": 0.8373900593168337, "eval_loss": 0.7715491056442261, "eval_macro avg": { "f1-score": 0.7798968986981474, "precision": 0.785328119114727, "recall": 0.7753429412408733, "support": 29334.0 }, "eval_runtime": 5.8552, "eval_samples_per_second": 13.663, "eval_steps_per_second": 1.708, "eval_weighted avg": { "f1-score": 0.8371368267053725, "precision": 0.8380850988337167, "recall": 0.8373900593168337, "support": 29334.0 }, "step": 810 }, { "epoch": 11.0, "eval_Claim": { "f1-score": 0.5986891586977544, "precision": 0.5522299306243805, "recall": 0.6536837165649929, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.760990850256639, "precision": 0.7361830742659758, "recall": 0.7875288683602771, "support": 2165.0 }, "eval_O": { "f1-score": 0.8930523028883683, "precision": 0.9179415855354659, "recall": 0.8694770976895014, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8843171634521723, "precision": 0.8987802946301283, "recall": 0.8703121405015722, "support": 13039.0 }, "eval_accuracy": 0.83244698984114, "eval_loss": 0.7798097133636475, "eval_macro avg": { "f1-score": 0.7842623688237336, "precision": 0.7762837212639876, "recall": 0.795250455779086, "support": 29334.0 }, "eval_runtime": 5.8447, "eval_samples_per_second": 13.688, "eval_steps_per_second": 1.711, "eval_weighted avg": { "f1-score": 0.8366540534646059, "precision": 0.8428746215263233, "recall": 0.83244698984114, "support": 29334.0 }, "step": 891 }, { "epoch": 12.0, "eval_Claim": { "f1-score": 0.5911446066886481, "precision": 0.5933806146572104, "recall": 0.5889253871421868, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7685296234108899, "precision": 0.7994011976047904, "recall": 0.7399538106235566, "support": 2165.0 }, "eval_O": { "f1-score": 0.8974528445777621, "precision": 0.9005203550658096, "recall": 0.8944061613295501, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8881464044346572, "precision": 0.8794646214001053, "recall": 0.8970013037809648, "support": 13039.0 }, "eval_accuracy": 0.8397763687188927, "eval_loss": 0.8434343338012695, "eval_macro avg": { "f1-score": 0.7863183697779894, "precision": 0.7931916971819789, "recall": 0.7800716657190646, "support": 29334.0 }, "eval_runtime": 5.824, "eval_samples_per_second": 13.736, "eval_steps_per_second": 1.717, "eval_weighted avg": { "f1-score": 0.8392967405095946, "precision": 0.8390729472526346, "recall": 0.8397763687188927, "support": 29334.0 }, "step": 972 }, { "epoch": 12.35, "grad_norm": 8.536613464355469, "learning_rate": 7.654320987654322e-06, "loss": 0.0633, "step": 1000 }, { "epoch": 13.0, "eval_Claim": { "f1-score": 0.5875576036866359, "precision": 0.5771842462652784, "recall": 0.5983106522759268, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7726320688852689, "precision": 0.7786116322701688, "recall": 0.766743648960739, "support": 2165.0 }, "eval_O": { "f1-score": 0.8950658744510462, "precision": 0.9130388953304522, "recall": 0.8777867855695176, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8827460510328068, "precision": 0.8741821463488004, "recall": 0.891479407930056, "support": 13039.0 }, "eval_accuracy": 0.835071930183405, "eval_loss": 0.9362213015556335, "eval_macro avg": { "f1-score": 0.7845003995139396, "precision": 0.785754230053675, "recall": 0.7835801236840598, "support": 29334.0 }, "eval_runtime": 5.8402, "eval_samples_per_second": 13.698, "eval_steps_per_second": 1.712, "eval_weighted avg": { "f1-score": 0.83587491458883, "precision": 0.8370485534468686, "recall": 0.835071930183405, "support": 29334.0 }, "step": 1053 }, { "epoch": 14.0, "eval_Claim": { "f1-score": 0.5757274309873166, "precision": 0.6124338624338624, "recall": 0.5431722196152041, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7806691449814126, "precision": 0.7854137447405329, "recall": 0.7759815242494227, "support": 2165.0 }, "eval_O": { "f1-score": 0.894971487817522, "precision": 0.916153682869879, "recall": 0.8747466558573166, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8893903521751998, "precision": 0.8590723933395269, "recall": 0.9219265281079837, "support": 13039.0 }, "eval_accuracy": 0.8402536305993046, "eval_loss": 1.0311453342437744, "eval_macro avg": { "f1-score": 0.7851896039903627, "precision": 0.7932684208459503, "recall": 0.7789567319574817, "support": 29334.0 }, "eval_runtime": 5.8463, "eval_samples_per_second": 13.684, "eval_steps_per_second": 1.71, "eval_weighted avg": { "f1-score": 0.8376709093048489, "precision": 0.8370035916809992, "recall": 0.8402536305993046, "support": 29334.0 }, "step": 1134 }, { "epoch": 15.0, "eval_Claim": { "f1-score": 0.5845799769850402, "precision": 0.5736224028906955, "recall": 0.5959643359924918, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7613293051359518, "precision": 0.8367459878251245, "recall": 0.6983833718244804, "support": 2165.0 }, "eval_O": { "f1-score": 0.8997656860192659, "precision": 0.9253507550605119, "recall": 0.8755573571139035, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8875041976045669, "precision": 0.8641912512716174, "recall": 0.9121098243730348, "support": 13039.0 }, "eval_accuracy": 0.8381059521374514, "eval_loss": 1.0062618255615234, "eval_macro avg": { "f1-score": 0.7832947914362062, "precision": 0.7999775992619873, "recall": 0.7705037223259776, "support": 29334.0 }, "eval_runtime": 5.8597, "eval_samples_per_second": 13.653, "eval_steps_per_second": 1.707, "eval_weighted avg": { "f1-score": 0.8383041122838222, "precision": 0.8405224217982303, "recall": 0.8381059521374514, "support": 29334.0 }, "step": 1215 }, { "epoch": 16.0, "eval_Claim": { "f1-score": 0.587790910198611, "precision": 0.6114068441064638, "recall": 0.5659314875645237, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7780703854440987, "precision": 0.8076540755467196, "recall": 0.7505773672055427, "support": 2165.0 }, "eval_O": { "f1-score": 0.8989037758830695, "precision": 0.9003660024400163, "recall": 0.8974462910417511, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.89104589917231, "precision": 0.8745292075917583, "recall": 0.908198481478641, "support": 13039.0 }, "eval_accuracy": 0.8432194722847208, "eval_loss": 0.9864347577095032, "eval_macro avg": { "f1-score": 0.7889527426745223, "precision": 0.7984890324212395, "recall": 0.7805384068226147, "support": 29334.0 }, "eval_runtime": 5.8915, "eval_samples_per_second": 13.579, "eval_steps_per_second": 1.697, "eval_weighted avg": { "f1-score": 0.8412905564694496, "precision": 0.8400553996388973, "recall": 0.8432194722847208, "support": 29334.0 }, "step": 1296 }, { "epoch": 17.0, "eval_Claim": { "f1-score": 0.5855803726420553, "precision": 0.5777574788764558, "recall": 0.5936180197090568, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7654614652711702, "precision": 0.7891123099558607, "recall": 0.7431870669745958, "support": 2165.0 }, "eval_O": { "f1-score": 0.9005902941022829, "precision": 0.9293800539083558, "recall": 0.8735306039724362, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8890554722638682, "precision": 0.869437724507001, "recall": 0.9095789554413682, "support": 13039.0 }, "eval_accuracy": 0.8392650167041659, "eval_loss": 1.047443151473999, "eval_macro avg": { "f1-score": 0.7851719010698441, "precision": 0.7914218918119184, "recall": 0.7799786615243642, "support": 29334.0 }, "eval_runtime": 5.8682, "eval_samples_per_second": 13.633, "eval_steps_per_second": 1.704, "eval_weighted avg": { "f1-score": 0.8397213794764584, "precision": 0.8412951315142951, "recall": 0.8392650167041659, "support": 29334.0 }, "step": 1377 }, { "epoch": 18.0, "eval_Claim": { "f1-score": 0.5782873522035114, "precision": 0.5889078083191438, "recall": 0.5680431722196152, "support": 4262.0 }, "eval_MajorClaim": { "f1-score": 0.7745774815520113, "precision": 0.7991159135559921, "recall": 0.751501154734411, "support": 2165.0 }, "eval_O": { "f1-score": 0.8985953040090229, "precision": 0.9093172857439302, "recall": 0.8881232265910012, "support": 9868.0 }, "eval_Premise": { "f1-score": 0.8900255754475703, "precision": 0.8732747804265998, "recall": 0.9074315514993481, "support": 13039.0 }, "eval_accuracy": 0.8401172700620441, "eval_loss": 1.0609416961669922, "eval_macro avg": { "f1-score": 0.7853714283030289, "precision": 0.7926539470114164, "recall": 0.7787747762610939, "support": 29334.0 }, "eval_runtime": 5.8605, "eval_samples_per_second": 13.651, "eval_steps_per_second": 1.706, "eval_weighted avg": { "f1-score": 0.8390946642419506, "precision": 0.8386099362381009, "recall": 0.8401172700620441, "support": 29334.0 }, "step": 1458 } ], "logging_steps": 500, "max_steps": 1620, "num_input_tokens_seen": 0, "num_train_epochs": 20, "save_steps": 500, "total_flos": 2588197189082400.0, "train_batch_size": 4, "trial_name": null, "trial_params": null }