Training in progress, epoch 17, checkpoint

checkpoint-1377/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f67efc4ff75de9fdc1e63fcf974b5a915819f3ac88169eb9cee1269907c9380c
+oid sha256:f62a81053649ee2a238cbc272d3ef5659b43d59921fe3f035373b5e0e8d741c8
 size 592318676

checkpoint-1377/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c65890a8b9b9632a5fd21e73e3721bbc75ed89aba2dab05232f8c0f16e87f10f
+oid sha256:7509803ac2f20674e0d196042545603fe048ec41373bb0b9573f5224f97fef1a
 size 1014645498

checkpoint-1377/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:032f609362cfbe4a4e5efe180a3007cc3751761c4055f83fe66b2c2d7b6093ac
+oid sha256:f07a894ccb1b17e16344c8e9343bb0267b1fc15fa8de24c4c96e2ca314e74fb1
 size 1064

checkpoint-1377/trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": null,
-  "best_model_checkpoint": null,
+  "best_metric": 0.2252955436706543,
+  "best_model_checkpoint": "longformer-spans/checkpoint-162",
   "epoch": 17.0,
   "eval_steps": 500,
   "global_step": 1377,
@@ -11,685 +11,685 @@
     {
       "epoch": 1.0,
       "eval_B": {
-        "f1-score": 0.8331822302810515,
-        "precision": 0.7901977644024075,
-        "recall": 0.8811121764141898,
-        "support": 1043.0
+        "f1-score": 0.816,
+        "precision": 0.7461594732991953,
+        "recall": 0.9002647837599294,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9471294753175105,
-        "precision": 0.9443616777446711,
-        "recall": 0.9499135446685879,
-        "support": 17350.0
+        "f1-score": 0.9321376763813793,
+        "precision": 0.9024103768767235,
+        "recall": 0.9638902525500463,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.8992868897421833,
-        "precision": 0.9103731674811195,
-        "recall": 0.8884673748103187,
-        "support": 9226.0
+        "f1-score": 0.8560784528570645,
+        "precision": 0.931782945736434,
+        "recall": 0.7917511147142278,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9267895289474637,
-      "eval_loss": 0.20478281378746033,
+      "eval_accuracy": 0.9035249198881844,
+      "eval_loss": 0.2619660496711731,
       "eval_macro avg": {
-        "f1-score": 0.8931995317802484,
-        "precision": 0.8816442032093993,
-        "recall": 0.9064976986310321,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6128,
-      "eval_samples_per_second": 17.343,
-      "eval_steps_per_second": 2.168,
+        "f1-score": 0.8680720430794812,
+        "precision": 0.860117598637451,
+        "recall": 0.8853020503414012,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.5032,
+      "eval_samples_per_second": 14.537,
+      "eval_steps_per_second": 1.817,
       "eval_weighted avg": {
-        "f1-score": 0.9268447919078652,
-        "precision": 0.9271861479533134,
-        "recall": 0.9267895289474637,
-        "support": 27619.0
+        "f1-score": 0.9020655278480035,
+        "precision": 0.9062562975065145,
+        "recall": 0.9035249198881844,
+        "support": 29334.0
       },
       "step": 81
     },
     {
       "epoch": 2.0,
       "eval_B": {
-        "f1-score": 0.8682547399124939,
-        "precision": 0.8806706114398422,
-        "recall": 0.8561840843720039,
-        "support": 1043.0
+        "f1-score": 0.8512256973795435,
+        "precision": 0.8167072181670721,
+        "recall": 0.8887908208296558,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9590983771942466,
-        "precision": 0.9592919333448654,
-        "recall": 0.9589048991354466,
-        "support": 17350.0
+        "f1-score": 0.9416781438711729,
+        "precision": 0.9152551099212274,
+        "recall": 0.9696721758577429,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9226525313717006,
-        "precision": 0.9208594256100194,
-        "recall": 0.9244526338608281,
-        "support": 9226.0
+        "f1-score": 0.8778173190984578,
+        "precision": 0.9380041484212952,
+        "recall": 0.8248885285772193,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9435171439950758,
-      "eval_loss": 0.16574698686599731,
+      "eval_accuracy": 0.9178427763005387,
+      "eval_loss": 0.2252955436706543,
       "eval_macro avg": {
-        "f1-score": 0.9166685494928136,
-        "precision": 0.9202739901315756,
-        "recall": 0.9131805391227594,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6238,
-      "eval_samples_per_second": 17.302,
-      "eval_steps_per_second": 2.163,
+        "f1-score": 0.8902403867830581,
+        "precision": 0.8899888255031981,
+        "recall": 0.8944505084215394,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.4972,
+      "eval_samples_per_second": 14.553,
+      "eval_steps_per_second": 1.819,
       "eval_weighted avg": {
-        "f1-score": 0.9434932036816766,
-        "precision": 0.9434846863370583,
-        "recall": 0.9435171439950758,
-        "support": 27619.0
+        "f1-score": 0.9167016237671239,
+        "precision": 0.9191015935430046,
+        "recall": 0.9178427763005387,
+        "support": 29334.0
       },
       "step": 162
     },
     {
       "epoch": 3.0,
       "eval_B": {
-        "f1-score": 0.8791725434884815,
-        "precision": 0.8625461254612546,
-        "recall": 0.8964525407478428,
-        "support": 1043.0
+        "f1-score": 0.8531120331950207,
+        "precision": 0.8050117462803446,
+        "recall": 0.9073256840247131,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9588372159825352,
-        "precision": 0.9557349825345015,
-        "recall": 0.9619596541786744,
-        "support": 17350.0
+        "f1-score": 0.9469915571230095,
+        "precision": 0.9280963603037444,
+        "recall": 0.9666721213112965,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9204284621270084,
-        "precision": 0.9282407407407407,
-        "recall": 0.9127465857359636,
-        "support": 9226.0
+        "f1-score": 0.8903876792352629,
+        "precision": 0.9353938852934612,
+        "recall": 0.8495135792460479,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9430464535283681,
-      "eval_loss": 0.15394224226474762,
+      "eval_accuracy": 0.9249676143724006,
+      "eval_loss": 0.22786012291908264,
       "eval_macro avg": {
-        "f1-score": 0.9194794071993417,
-        "precision": 0.9155072829121655,
-        "recall": 0.9237195935541602,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6233,
-      "eval_samples_per_second": 17.303,
-      "eval_steps_per_second": 2.163,
+        "f1-score": 0.8968304231844311,
+        "precision": 0.8895006639591835,
+        "recall": 0.9078371281940192,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.5036,
+      "eval_samples_per_second": 14.536,
+      "eval_steps_per_second": 1.817,
       "eval_weighted avg": {
-        "f1-score": 0.9429985029052194,
-        "precision": 0.9430314866542512,
-        "recall": 0.9430464535283681,
-        "support": 27619.0
+        "f1-score": 0.9243239165827936,
+        "precision": 0.9257972230878861,
+        "recall": 0.9249676143724006,
+        "support": 29334.0
       },
       "step": 243
     },
     {
       "epoch": 4.0,
       "eval_B": {
-        "f1-score": 0.8800746616892207,
-        "precision": 0.8572727272727273,
-        "recall": 0.9041227229146692,
-        "support": 1043.0
+        "f1-score": 0.8567807351077312,
+        "precision": 0.8217179902755267,
+        "recall": 0.8949691085613416,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9537844036697248,
-        "precision": 0.9488876212207644,
-        "recall": 0.9587319884726225,
-        "support": 17350.0
+        "f1-score": 0.9472597903427299,
+        "precision": 0.9432635621180161,
+        "recall": 0.9512900234549719,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9100192149327477,
-        "precision": 0.9220157970853265,
-        "recall": 0.8983308042488619,
-        "support": 9226.0
+        "f1-score": 0.8979927100980543,
+        "precision": 0.9099989595255437,
+        "recall": 0.8862991487636805,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9364929939534379,
-      "eval_loss": 0.19801080226898193,
+      "eval_accuracy": 0.9272516533715143,
+      "eval_loss": 0.23897655308246613,
       "eval_macro avg": {
-        "f1-score": 0.9146260934305644,
-        "precision": 0.9093920485262728,
-        "recall": 0.9203951718787179,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6132,
-      "eval_samples_per_second": 17.342,
-      "eval_steps_per_second": 2.168,
+        "f1-score": 0.9006777451828384,
+        "precision": 0.8916601706396955,
+        "recall": 0.910852760259998,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.4904,
+      "eval_samples_per_second": 14.571,
+      "eval_steps_per_second": 1.821,
       "eval_weighted avg": {
-        "f1-score": 0.9363812792925563,
-        "precision": 0.9364514800186444,
-        "recall": 0.9364929939534379,
-        "support": 27619.0
+        "f1-score": 0.9271915992526735,
+        "precision": 0.9273787107073643,
+        "recall": 0.9272516533715143,
+        "support": 29334.0
       },
       "step": 324
     },
     {
       "epoch": 5.0,
       "eval_B": {
-        "f1-score": 0.883877159309021,
-        "precision": 0.8847262247838616,
-        "recall": 0.8830297219558965,
-        "support": 1043.0
+        "f1-score": 0.8624407072013798,
+        "precision": 0.8431703204047217,
+        "recall": 0.8826125330979699,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9591500482927107,
-        "precision": 0.9456643513331839,
-        "recall": 0.9730259365994236,
-        "support": 17350.0
+        "f1-score": 0.948190701170407,
+        "precision": 0.9335059992600032,
+        "recall": 0.9633447880870561,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9191176470588236,
-        "precision": 0.9454503781801513,
-        "recall": 0.8942120095382614,
-        "support": 9226.0
+        "f1-score": 0.8955333298423835,
+        "precision": 0.9265359193845487,
+        "recall": 0.8665383056343737,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9432999022412107,
-      "eval_loss": 0.20074255764484406,
+      "eval_accuracy": 0.9276607349832958,
+      "eval_loss": 0.25390708446502686,
       "eval_macro avg": {
-        "f1-score": 0.9207149515535183,
-        "precision": 0.9252803180990656,
-        "recall": 0.9167558893645271,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6092,
-      "eval_samples_per_second": 17.356,
-      "eval_steps_per_second": 2.17,
+        "f1-score": 0.9020549127380568,
+        "precision": 0.9010707463497579,
+        "recall": 0.9041652089397999,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.4919,
+      "eval_samples_per_second": 14.567,
+      "eval_steps_per_second": 1.821,
       "eval_weighted avg": {
-        "f1-score": 0.9429348139614955,
-        "precision": 0.9432916158141273,
-        "recall": 0.9432999022412107,
-        "support": 27619.0
+        "f1-score": 0.9271646670996412,
+        "precision": 0.9276721180179627,
+        "recall": 0.9276607349832958,
+        "support": 29334.0
       },
       "step": 405
     },
     {
       "epoch": 6.0,
       "eval_B": {
-        "f1-score": 0.8897338403041826,
-        "precision": 0.882186616399623,
-        "recall": 0.8974113135186961,
-        "support": 1043.0
+        "f1-score": 0.8601036269430052,
+        "precision": 0.841927303465765,
+        "recall": 0.8790820829655781,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9624160683661308,
-        "precision": 0.9707969272268809,
-        "recall": 0.954178674351585,
-        "support": 17350.0
+        "f1-score": 0.9474285714285714,
+        "precision": 0.9452679589509693,
+        "recall": 0.9495990836197021,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9297955261331482,
-        "precision": 0.9161493950552341,
-        "recall": 0.9438543247344461,
-        "support": 9226.0
+        "f1-score": 0.8983777165595348,
+        "precision": 0.9045613314156564,
+        "recall": 0.8922780705310093,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9485861182519281,
-      "eval_loss": 0.16565027832984924,
+      "eval_accuracy": 0.9275925547146656,
+      "eval_loss": 0.29299432039260864,
       "eval_macro avg": {
-        "f1-score": 0.9273151449344872,
-        "precision": 0.9230443128939126,
-        "recall": 0.9318147708682424,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6219,
-      "eval_samples_per_second": 17.309,
-      "eval_steps_per_second": 2.164,
+        "f1-score": 0.9019699716437038,
+        "precision": 0.8972521979441302,
+        "recall": 0.9069864123720964,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.4971,
+      "eval_samples_per_second": 14.553,
+      "eval_steps_per_second": 1.819,
       "eval_weighted avg": {
-        "f1-score": 0.9487745648174827,
-        "precision": 0.9491959030765336,
-        "recall": 0.9485861182519281,
-        "support": 27619.0
+        "f1-score": 0.9275549436263691,
+        "precision": 0.9275827485063247,
+        "recall": 0.9275925547146656,
+        "support": 29334.0
       },
       "step": 486
     },
     {
       "epoch": 6.17,
-      "grad_norm": 4.767655849456787,
-      "learning_rate": 1.7530864197530865e-05,
-      "loss": 0.1585,
+      "grad_norm": 2.853372573852539,
+      "learning_rate": 1.3827160493827162e-05,
+      "loss": 0.1621,
       "step": 500
     },
     {
       "epoch": 7.0,
       "eval_B": {
-        "f1-score": 0.8802588996763754,
-        "precision": 0.85,
-        "recall": 0.912751677852349,
-        "support": 1043.0
+        "f1-score": 0.8665526090675792,
+        "precision": 0.8406639004149378,
+        "recall": 0.8940864960282436,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.958423523458803,
-        "precision": 0.9354661691269275,
-        "recall": 0.9825360230547551,
-        "support": 17350.0
+        "f1-score": 0.9492722371967655,
+        "precision": 0.9382959450098577,
+        "recall": 0.9605083728795069,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9139526911210146,
-        "precision": 0.9664088931851136,
-        "recall": 0.8668978972469109,
-        "support": 9226.0
+        "f1-score": 0.8984919396775871,
+        "precision": 0.9227729117709891,
+        "recall": 0.8754560194568302,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9412723125384699,
-      "eval_loss": 0.2591131329536438,
+      "eval_accuracy": 0.9293311515647371,
+      "eval_loss": 0.3148973286151886,
       "eval_macro avg": {
-        "f1-score": 0.9175450380853977,
-        "precision": 0.917291687437347,
-        "recall": 0.9207285327180049,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.617,
-      "eval_samples_per_second": 17.327,
-      "eval_steps_per_second": 2.166,
+        "f1-score": 0.9047722619806439,
+        "precision": 0.9005775857319281,
+        "recall": 0.9100169627881934,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.4821,
+      "eval_samples_per_second": 14.593,
+      "eval_steps_per_second": 1.824,
       "eval_weighted avg": {
-        "f1-score": 0.9406164485555297,
-        "precision": 0.9425749115781907,
-        "recall": 0.9412723125384699,
-        "support": 27619.0
+        "f1-score": 0.9289946986889036,
+        "precision": 0.9293030221719495,
+        "recall": 0.9293311515647371,
+        "support": 29334.0
       },
       "step": 567
     },
     {
       "epoch": 8.0,
       "eval_B": {
-        "f1-score": 0.8687910028116214,
-        "precision": 0.84967919340055,
-        "recall": 0.8887823585810163,
-        "support": 1043.0
+        "f1-score": 0.8624629707998307,
+        "precision": 0.8284552845528456,
+        "recall": 0.8993821712268314,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9595010203788336,
-        "precision": 0.956997878562009,
-        "recall": 0.9620172910662824,
-        "support": 17350.0
+        "f1-score": 0.9467755410030451,
+        "precision": 0.9356556940449557,
+        "recall": 0.9581628756886489,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9212035166275324,
-        "precision": 0.9282491471332673,
-        "recall": 0.9142640364188164,
-        "support": 9226.0
+        "f1-score": 0.8934263985831857,
+        "precision": 0.9191854233654877,
+        "recall": 0.8690717470612079,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9432999022412107,
-      "eval_loss": 0.2693409323692322,
+      "eval_accuracy": 0.9259221381332242,
+      "eval_loss": 0.34766319394111633,
       "eval_macro avg": {
-        "f1-score": 0.9164985132726624,
-        "precision": 0.9116420730319422,
-        "recall": 0.9216878953553717,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6153,
-      "eval_samples_per_second": 17.333,
-      "eval_steps_per_second": 2.167,
+        "f1-score": 0.9008883034620205,
+        "precision": 0.8944321339877629,
+        "recall": 0.9088722646588961,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.5202,
+      "eval_samples_per_second": 14.492,
+      "eval_steps_per_second": 1.812,
       "eval_weighted avg": {
-        "f1-score": 0.9432823550422136,
-        "precision": 0.9433417293609165,
-        "recall": 0.9432999022412107,
-        "support": 27619.0
+        "f1-score": 0.9255723133682386,
+        "precision": 0.9259745494680297,
+        "recall": 0.9259221381332242,
+        "support": 29334.0
       },
       "step": 648
     },
     {
       "epoch": 9.0,
       "eval_B": {
-        "f1-score": 0.8645343367826904,
-        "precision": 0.8485687903970452,
-        "recall": 0.8811121764141898,
-        "support": 1043.0
+        "f1-score": 0.8654012079378774,
+        "precision": 0.8464135021097047,
+        "recall": 0.8852603706972639,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9567003348057282,
-        "precision": 0.934535260814599,
-        "recall": 0.9799423631123919,
-        "support": 17350.0
+        "f1-score": 0.9474813007694164,
+        "precision": 0.9316216786166175,
+        "recall": 0.9638902525500463,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9104673003585861,
-        "precision": 0.9586479683567062,
-        "recall": 0.8668978972469109,
-        "support": 9226.0
+        "f1-score": 0.8933802299333298,
+        "precision": 0.9268053588933667,
+        "recall": 0.8622821240372922,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9384481697382236,
-      "eval_loss": 0.3087053894996643,
+      "eval_accuracy": 0.9266721210881571,
+      "eval_loss": 0.3807723820209503,
       "eval_macro avg": {
-        "f1-score": 0.910567323982335,
-        "precision": 0.9139173398561168,
-        "recall": 0.9093174789244975,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6156,
-      "eval_samples_per_second": 17.333,
-      "eval_steps_per_second": 2.167,
+        "f1-score": 0.9020875795468745,
+        "precision": 0.9016135132065629,
+        "recall": 0.9038109157615342,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.532,
+      "eval_samples_per_second": 14.461,
+      "eval_steps_per_second": 1.808,
       "eval_weighted avg": {
-        "f1-score": 0.9377758584761231,
-        "precision": 0.9393435743356523,
-        "recall": 0.9384481697382236,
-        "support": 27619.0
+        "f1-score": 0.9261113508073029,
+        "precision": 0.9267103706800465,
+        "recall": 0.9266721210881571,
+        "support": 29334.0
       },
       "step": 729
     },
     {
       "epoch": 10.0,
       "eval_B": {
-        "f1-score": 0.8803126526624329,
-        "precision": 0.897410358565737,
-        "recall": 0.8638542665388304,
-        "support": 1043.0
+        "f1-score": 0.8593548387096774,
+        "precision": 0.8380872483221476,
+        "recall": 0.881729920564872,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9590983771942466,
-        "precision": 0.9592919333448654,
-        "recall": 0.9589048991354466,
-        "support": 17350.0
+        "f1-score": 0.9447995351539802,
+        "precision": 0.9158687080751703,
+        "recall": 0.9756177385043364,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9220456265542221,
-        "precision": 0.9197584124245038,
-        "recall": 0.9243442445263386,
-        "support": 9226.0
+        "f1-score": 0.8826362209837131,
+        "precision": 0.9469406710786021,
+        "recall": 0.8265099310903932,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9437705927079184,
-      "eval_loss": 0.2796257436275482,
+      "eval_accuracy": 0.9218313220154087,
+      "eval_loss": 0.46634167432785034,
       "eval_macro avg": {
-        "f1-score": 0.9204855521369671,
-        "precision": 0.9254869014450354,
-        "recall": 0.9157011367335386,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.641,
-      "eval_samples_per_second": 17.238,
-      "eval_steps_per_second": 2.155,
+        "f1-score": 0.8955968649491236,
+        "precision": 0.9002988758253068,
+        "recall": 0.8946191967198672,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.5106,
+      "eval_samples_per_second": 14.517,
+      "eval_steps_per_second": 1.815,
       "eval_weighted avg": {
-        "f1-score": 0.9437458232244597,
-        "precision": 0.9437490553802075,
-        "recall": 0.9437705927079184,
-        "support": 27619.0
+        "f1-score": 0.9205874800198834,
+        "precision": 0.9233171207368492,
+        "recall": 0.9218313220154087,
+        "support": 29334.0
       },
       "step": 810
     },
     {
       "epoch": 11.0,
       "eval_B": {
-        "f1-score": 0.8830917874396135,
-        "precision": 0.8899707887049659,
-        "recall": 0.8763183125599233,
-        "support": 1043.0
+        "f1-score": 0.8585028126352229,
+        "precision": 0.8421052631578947,
+        "recall": 0.8755516328331863,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9574634882843123,
-        "precision": 0.9647182727751448,
-        "recall": 0.9503170028818444,
-        "support": 17350.0
+        "f1-score": 0.9467715680954965,
+        "precision": 0.941814648890808,
+        "recall": 0.9517809414716631,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9200619426496501,
-        "precision": 0.9067466582465004,
-        "recall": 0.9337741166269239,
-        "support": 9226.0
+        "f1-score": 0.8955223880597014,
+        "precision": 0.906636203136359,
+        "recall": 0.8846777462505067,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9419964517180202,
-      "eval_loss": 0.3198166787624359,
+      "eval_accuracy": 0.9262630394763756,
+      "eval_loss": 0.39979249238967896,
       "eval_macro avg": {
-        "f1-score": 0.9202057394578587,
-        "precision": 0.9204785732422037,
-        "recall": 0.9201364773562305,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6345,
-      "eval_samples_per_second": 17.262,
-      "eval_steps_per_second": 2.158,
+        "f1-score": 0.9002655895968069,
+        "precision": 0.8968520383950206,
+        "recall": 0.9040034401851186,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.5098,
+      "eval_samples_per_second": 14.52,
+      "eval_steps_per_second": 1.815,
       "eval_weighted avg": {
-        "f1-score": 0.942161111514465,
-        "precision": 0.9425303680165918,
-        "recall": 0.9419964517180202,
-        "support": 27619.0
+        "f1-score": 0.9261219666592889,
+        "precision": 0.9261293813943774,
+        "recall": 0.9262630394763756,
+        "support": 29334.0
       },
       "step": 891
     },
     {
       "epoch": 12.0,
       "eval_B": {
-        "f1-score": 0.871306005719733,
-        "precision": 0.8663507109004739,
-        "recall": 0.8763183125599233,
-        "support": 1043.0
+        "f1-score": 0.8661758336942399,
+        "precision": 0.8503401360544217,
+        "recall": 0.8826125330979699,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9576855173984852,
-        "precision": 0.9464205312922107,
-        "recall": 0.9692219020172911,
-        "support": 17350.0
+        "f1-score": 0.9483851819874267,
+        "precision": 0.9383342231713828,
+        "recall": 0.9586537937053401,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9162135168127844,
-        "precision": 0.9386084583901774,
-        "recall": 0.8948623455451984,
-        "support": 9226.0
+        "f1-score": 0.8972844112769486,
+        "precision": 0.9182223165040305,
+        "recall": 0.8772800972841508,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9408740359897172,
-      "eval_loss": 0.35264310240745544,
+      "eval_accuracy": 0.9283425376695984,
+      "eval_loss": 0.4524174630641937,
       "eval_macro avg": {
-        "f1-score": 0.9150683466436677,
-        "precision": 0.917126566860954,
-        "recall": 0.9134675200408043,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6326,
-      "eval_samples_per_second": 17.269,
-      "eval_steps_per_second": 2.159,
+        "f1-score": 0.9039484756528716,
+        "precision": 0.9022988919099451,
+        "recall": 0.906182141362487,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.5194,
+      "eval_samples_per_second": 14.494,
+      "eval_steps_per_second": 1.812,
       "eval_weighted avg": {
-        "f1-score": 0.9405699625961891,
-        "precision": 0.9407871989028143,
-        "recall": 0.9408740359897172,
-        "support": 27619.0
+        "f1-score": 0.9280195449455239,
+        "precision": 0.9281698543264606,
+        "recall": 0.9283425376695984,
+        "support": 29334.0
       },
       "step": 972
     },
     {
       "epoch": 12.35,
-      "grad_norm": 0.3602512776851654,
-      "learning_rate": 1.506172839506173e-05,
-      "loss": 0.021,
+      "grad_norm": 1.163489818572998,
+      "learning_rate": 7.654320987654322e-06,
+      "loss": 0.0212,
       "step": 1000
     },
     {
       "epoch": 13.0,
       "eval_B": {
-        "f1-score": 0.877092300334768,
-        "precision": 0.875,
-        "recall": 0.8791946308724832,
-        "support": 1043.0
+        "f1-score": 0.8624407072013798,
+        "precision": 0.8431703204047217,
+        "recall": 0.8826125330979699,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9597242793665263,
-        "precision": 0.9486991778353419,
-        "recall": 0.9710086455331413,
-        "support": 17350.0
+        "f1-score": 0.94720379658092,
+        "precision": 0.9365968111768783,
+        "recall": 0.9580537827960508,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9203392649259936,
-        "precision": 0.9419040054464994,
-        "recall": 0.8997398655972252,
-        "support": 9226.0
+        "f1-score": 0.8942532315838654,
+        "precision": 0.9167642362959021,
+        "recall": 0.8728212403729225,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.943734385748941,
-      "eval_loss": 0.3594338297843933,
+      "eval_accuracy": 0.9264675802822663,
+      "eval_loss": 0.45370540022850037,
       "eval_macro avg": {
-        "f1-score": 0.9190519482090961,
-        "precision": 0.9218677277606138,
-        "recall": 0.9166477140009498,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6042,
-      "eval_samples_per_second": 17.376,
-      "eval_steps_per_second": 2.172,
+        "f1-score": 0.9012992451220551,
+        "precision": 0.8988437892925006,
+        "recall": 0.9044958520889811,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.5225,
+      "eval_samples_per_second": 14.486,
+      "eval_steps_per_second": 1.811,
       "eval_weighted avg": {
-        "f1-score": 0.943447393984779,
-        "precision": 0.9436461164304495,
-        "recall": 0.943734385748941,
-        "support": 27619.0
+        "f1-score": 0.9261172500595471,
+        "precision": 0.926316588126141,
+        "recall": 0.9264675802822663,
+        "support": 29334.0
       },
       "step": 1053
     },
     {
       "epoch": 14.0,
       "eval_B": {
-        "f1-score": 0.8754052802223252,
-        "precision": 0.8467741935483871,
-        "recall": 0.9060402684563759,
-        "support": 1043.0
+        "f1-score": 0.8689595124074879,
+        "precision": 0.8573883161512027,
+        "recall": 0.880847308031774,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9574973153224439,
-        "precision": 0.9392880904856953,
-        "recall": 0.9764265129682997,
-        "support": 17350.0
+        "f1-score": 0.9480582004921365,
+        "precision": 0.9300970873786408,
+        "recall": 0.9667266677575956,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9130164471825015,
-        "precision": 0.9539388213062477,
-        "recall": 0.8754606546715803,
-        "support": 9226.0
+        "f1-score": 0.8934309645472265,
+        "precision": 0.9303346132748217,
+        "recall": 0.8593433319821646,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9400412759332344,
-      "eval_loss": 0.3840242028236389,
+      "eval_accuracy": 0.9272857435058294,
+      "eval_loss": 0.49018412828445435,
       "eval_macro avg": {
-        "f1-score": 0.9153063475757568,
-        "precision": 0.9133337017801101,
-        "recall": 0.9193091453654186,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6182,
-      "eval_samples_per_second": 17.323,
-      "eval_steps_per_second": 2.165,
+        "f1-score": 0.9034828924822836,
+        "precision": 0.9059400056015551,
+        "recall": 0.9023057692571781,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.5116,
+      "eval_samples_per_second": 14.515,
+      "eval_steps_per_second": 1.814,
       "eval_weighted avg": {
-        "f1-score": 0.9395385738014427,
-        "precision": 0.9406884180878825,
-        "recall": 0.9400412759332344,
-        "support": 27619.0
+        "f1-score": 0.9266264019680934,
+        "precision": 0.9273686789700647,
+        "recall": 0.9272857435058294,
+        "support": 29334.0
       },
       "step": 1134
     },
     {
       "epoch": 15.0,
       "eval_B": {
-        "f1-score": 0.8797709923664121,
-        "precision": 0.8755935422602089,
-        "recall": 0.8839884947267498,
-        "support": 1043.0
+        "f1-score": 0.865451388888889,
+        "precision": 0.8514090520922288,
+        "recall": 0.8799646954986761,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9603688219460511,
-        "precision": 0.954302299112224,
-        "recall": 0.9665129682997118,
-        "support": 17350.0
+        "f1-score": 0.9478711568207105,
+        "precision": 0.943136407819419,
+        "recall": 0.9526536846124475,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.922722283205269,
-        "precision": 0.9346230820547031,
-        "recall": 0.9111207457186213,
-        "support": 9226.0
+        "f1-score": 0.8980679546968688,
+        "precision": 0.9084499740798341,
+        "recall": 0.8879205512768544,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9448930084362215,
-      "eval_loss": 0.3762665390968323,
+      "eval_accuracy": 0.9280698165950774,
+      "eval_loss": 0.463104248046875,
       "eval_macro avg": {
-        "f1-score": 0.9209540325059108,
-        "precision": 0.9215063078090453,
-        "recall": 0.920540736248361,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.586,
-      "eval_samples_per_second": 17.444,
-      "eval_steps_per_second": 2.181,
+        "f1-score": 0.9037968334688228,
+        "precision": 0.9009984779971606,
+        "recall": 0.9068463104626593,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.477,
+      "eval_samples_per_second": 14.607,
+      "eval_steps_per_second": 1.826,
       "eval_weighted avg": {
-        "f1-score": 0.944749483712443,
-        "precision": 0.9447562007752335,
-        "recall": 0.9448930084362215,
-        "support": 27619.0
+        "f1-score": 0.9279338964530544,
+        "precision": 0.9279249527781314,
+        "recall": 0.9280698165950774,
+        "support": 29334.0
       },
       "step": 1215
     },
     {
       "epoch": 16.0,
       "eval_B": {
-        "f1-score": 0.8858669256920835,
-        "precision": 0.8976377952755905,
-        "recall": 0.8744007670182167,
-        "support": 1043.0
+        "f1-score": 0.8670469504168494,
+        "precision": 0.8621291448516579,
+        "recall": 0.8720211827007943,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9586212929752427,
-        "precision": 0.9701351590627397,
-        "recall": 0.9473775216138328,
-        "support": 17350.0
+        "f1-score": 0.949639510706667,
+        "precision": 0.9403208556149732,
+        "recall": 0.9591447117220313,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9235412474849094,
-        "precision": 0.9027950310559006,
-        "recall": 0.9452633860828095,
-        "support": 9226.0
+        "f1-score": 0.8996693531721429,
+        "precision": 0.917685497470489,
+        "recall": 0.8823469801378192,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9439154205438285,
-      "eval_loss": 0.36898088455200195,
+      "eval_accuracy": 0.9299447739824095,
+      "eval_loss": 0.4684809744358063,
       "eval_macro avg": {
-        "f1-score": 0.9226764887174118,
-        "precision": 0.923522661798077,
-        "recall": 0.922347224904953,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6024,
-      "eval_samples_per_second": 17.382,
-      "eval_steps_per_second": 2.173,
+        "f1-score": 0.9054519380985532,
+        "precision": 0.9067118326457067,
+        "recall": 0.9045042915202149,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.5034,
+      "eval_samples_per_second": 14.536,
+      "eval_steps_per_second": 1.817,
       "eval_weighted avg": {
-        "f1-score": 0.9441554794131966,
-        "precision": 0.9449027186622512,
-        "recall": 0.9439154205438285,
-        "support": 27619.0
+        "f1-score": 0.9296394123443894,
+        "precision": 0.9296862022276206,
+        "recall": 0.9299447739824095,
+        "support": 29334.0
       },
       "step": 1296
     },
     {
       "epoch": 17.0,
       "eval_B": {
-        "f1-score": 0.8863309352517986,
-        "precision": 0.8867562380038387,
-        "recall": 0.8859060402684564,
-        "support": 1043.0
+        "f1-score": 0.8695278969957082,
+        "precision": 0.8462823725981621,
+        "recall": 0.8940864960282436,
+        "support": 1133.0
       },
       "eval_I": {
-        "f1-score": 0.9630614486386496,
-        "precision": 0.9659090909090909,
-        "recall": 0.9602305475504322,
-        "support": 17350.0
+        "f1-score": 0.9479728646973986,
+        "precision": 0.9287246847035429,
+        "recall": 0.9680357824687722,
+        "support": 18333.0
       },
       "eval_O": {
-        "f1-score": 0.9296685529506872,
-        "precision": 0.9245363918962375,
-        "recall": 0.9348580099718188,
-        "support": 9226.0
+        "f1-score": 0.892887383573243,
+        "precision": 0.9344262295081968,
+        "recall": 0.8548844750709363,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.9489481878417032,
-      "eval_loss": 0.3569599986076355,
+      "eval_accuracy": 0.9271152928342538,
+      "eval_loss": 0.5304692983627319,
       "eval_macro avg": {
-        "f1-score": 0.9263536456137119,
-        "precision": 0.9257339069363891,
-        "recall": 0.9269981992635691,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.6143,
-      "eval_samples_per_second": 17.337,
-      "eval_steps_per_second": 2.167,
+        "f1-score": 0.9034627150887832,
+        "precision": 0.9031444289366339,
+        "recall": 0.9056689178559841,
+        "support": 29334.0
+      },
+      "eval_runtime": 5.4944,
+      "eval_samples_per_second": 14.56,
+      "eval_steps_per_second": 1.82,
       "eval_weighted avg": {
-        "f1-score": 0.9490090650954502,
-        "precision": 0.9490996138580476,
-        "recall": 0.9489481878417032,
-        "support": 27619.0
+        "f1-score": 0.9264121612086422,
+        "precision": 0.927458430681484,
+        "recall": 0.9271152928342538,
+        "support": 29334.0
       },
       "step": 1377
     }
   ],
   "logging_steps": 500,
-  "max_steps": 4050,
+  "max_steps": 1620,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 50,
+  "num_train_epochs": 20,
   "save_steps": 500,
   "total_flos": 2444373096505200.0,
   "train_batch_size": 4,

checkpoint-1377/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:821ca433db9ddb761291ab61fc06d651ad248cd9d560cdb25f7f905e5aaff14e
+oid sha256:1094850123635c13193a80398a7bb1e351dd871416d3ce265431b259bf603290
 size 4856