Training in progress, epoch 32, checkpoint

checkpoint-2592/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2d90b605a76c19e3ced65a2968ebeae2852269393ba3d8c7f8183784058c2b75
+oid sha256:58604eadc144fad6d9c61b56bf21a6d4c0d898130bc08779a178df76189004a3
 size 592318676

checkpoint-2592/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:902af4187308b4ad408181de2677ec1a986dfd56595ecf8e3dc1177b85f9d1d8
+oid sha256:3c0341eac477accd8a88106b1366d69b43d0a714ff4f3aed5dadeccac16acac3
 size 1014645498

checkpoint-2592/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c36c35f4797e013c49557271e05367bd8b03f280228194db7a22f1cc4558e45a
+oid sha256:41f918bac4bb60118f27fd9b0d48aed0bee4a4e46dd9e93ab31eeb9022fe1c15
 size 14244

checkpoint-2592/trainer_state.json

@@ -11,1283 +11,1283 @@
     {
       "epoch": 1.0,
       "eval_B": {
-        "f1-score": 0.8266764382557715,
-        "precision": 0.7866108786610879,
-        "recall": 0.871042471042471,
-        "support": 1295.0
+        "f1-score": 0.7905866302864938,
+        "precision": 0.7768096514745308,
+        "recall": 0.8048611111111111,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9410658617447671,
-        "precision": 0.9344077040239768,
-        "recall": 0.9478195863443808,
-        "support": 20065.0
+        "f1-score": 0.9005331227169513,
+        "precision": 0.9637593111099372,
+        "recall": 0.8450919534905267,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.8562443302086484,
-        "precision": 0.8789421405512788,
-        "recall": 0.8346893055064262,
-        "support": 8481.0
+        "f1-score": 0.8288892663043479,
+        "precision": 0.7463108800367001,
+        "recall": 0.9320156593144275,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.912335377500754,
-      "eval_loss": 0.2450982630252838,
+      "eval_accuracy": 0.8705373134328358,
+      "eval_loss": 0.32930415868759155,
       "eval_macro avg": {
-        "f1-score": 0.8746622100697291,
-        "precision": 0.8666535744121145,
-        "recall": 0.8845171209644261,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6269,
-      "eval_samples_per_second": 17.29,
-      "eval_steps_per_second": 2.161,
+        "f1-score": 0.8400030064359308,
+        "precision": 0.8289599475403894,
+        "recall": 0.8606562413053552,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7349,
+      "eval_samples_per_second": 17.107,
+      "eval_steps_per_second": 2.323,
       "eval_weighted avg": {
-        "f1-score": 0.9119949287205363,
-        "precision": 0.9122301519092053,
-        "recall": 0.912335377500754,
-        "support": 29841.0
+        "f1-score": 0.8734092702599646,
+        "precision": 0.8877430445874,
+        "recall": 0.8705373134328358,
+        "support": 33500.0
       },
       "step": 81
     },
     {
       "epoch": 2.0,
       "eval_B": {
-        "f1-score": 0.8833395452851286,
-        "precision": 0.8537463976945245,
-        "recall": 0.915057915057915,
-        "support": 1295.0
+        "f1-score": 0.8656218739579861,
+        "precision": 0.8325849903784477,
+        "recall": 0.9013888888888889,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9549518227873249,
-        "precision": 0.9516903430183636,
-        "recall": 0.958235733864939,
-        "support": 20065.0
+        "f1-score": 0.9547168948727912,
+        "precision": 0.9423736462093862,
+        "recall": 0.9673877796822161,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.88924750463212,
-        "precision": 0.9016969696969697,
-        "recall": 0.877137130055418,
-        "support": 8481.0
+        "f1-score": 0.9013528191962081,
+        "precision": 0.9332379102341274,
+        "recall": 0.8715745249689678,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9333132267685399,
-      "eval_loss": 0.20076116919517517,
+      "eval_accuracy": 0.9345970149253732,
+      "eval_loss": 0.2162298709154129,
       "eval_macro avg": {
-        "f1-score": 0.9091796242348579,
-        "precision": 0.9023779034699526,
-        "recall": 0.916810259659424,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6524,
-      "eval_samples_per_second": 17.195,
-      "eval_steps_per_second": 2.149,
+        "f1-score": 0.9072305293423284,
+        "precision": 0.902732182273987,
+        "recall": 0.9134503978466909,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.753,
+      "eval_samples_per_second": 17.042,
+      "eval_steps_per_second": 2.314,
       "eval_weighted avg": {
-        "f1-score": 0.9331705077630417,
-        "precision": 0.9332314707173981,
-        "recall": 0.9333132267685399,
-        "support": 29841.0
+        "f1-score": 0.9342040950316517,
+        "precision": 0.9347982961417612,
+        "recall": 0.9345970149253732,
+        "support": 33500.0
       },
       "step": 162
     },
     {
       "epoch": 3.0,
       "eval_B": {
-        "f1-score": 0.8927335640138409,
-        "precision": 0.8889739663093414,
-        "recall": 0.8965250965250965,
-        "support": 1295.0
+        "f1-score": 0.8603587759409076,
+        "precision": 0.8717034925160371,
+        "recall": 0.8493055555555555,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9590235396687009,
-        "precision": 0.9586653386454184,
-        "recall": 0.9593820084724645,
-        "support": 20065.0
+        "f1-score": 0.9434195218082959,
+        "precision": 0.9677057963955188,
+        "recall": 0.9203224162690509,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9006849315068494,
-        "precision": 0.9020697811945594,
-        "recall": 0.8993043273198915,
-        "support": 8481.0
+        "f1-score": 0.8879310344827586,
+        "precision": 0.8459410391631365,
+        "recall": 0.9343072663038289,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9395797727958178,
-      "eval_loss": 0.19863653182983398,
+      "eval_accuracy": 0.9216417910447762,
+      "eval_loss": 0.24082203209400177,
       "eval_macro avg": {
-        "f1-score": 0.9174806783964637,
-        "precision": 0.9165696953831063,
-        "recall": 0.9184038107724842,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.655,
-      "eval_samples_per_second": 17.186,
-      "eval_steps_per_second": 2.148,
+        "f1-score": 0.8972364440773206,
+        "precision": 0.8951167760248975,
+        "recall": 0.9013117460428117,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7554,
+      "eval_samples_per_second": 17.033,
+      "eval_steps_per_second": 2.313,
       "eval_weighted avg": {
-        "f1-score": 0.9395665759512081,
-        "precision": 0.9395561516236713,
-        "recall": 0.9395797727958178,
-        "support": 29841.0
+        "f1-score": 0.9225019575751797,
+        "precision": 0.9255121957960801,
+        "recall": 0.9216417910447762,
+        "support": 33500.0
       },
       "step": 243
     },
     {
       "epoch": 4.0,
       "eval_B": {
-        "f1-score": 0.8999616711383672,
-        "precision": 0.893455098934551,
-        "recall": 0.9065637065637066,
-        "support": 1295.0
+        "f1-score": 0.8804020100502512,
+        "precision": 0.8504854368932039,
+        "recall": 0.9125,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9612657113512985,
-        "precision": 0.9436362763454799,
-        "recall": 0.9795664091701969,
-        "support": 20065.0
+        "f1-score": 0.9589184992350025,
+        "precision": 0.945595388218339,
+        "recall": 0.9726224116366332,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.8996847765622102,
-        "precision": 0.9454403741231489,
-        "recall": 0.8581535196321188,
-        "support": 8481.0
+        "f1-score": 0.9091178797468354,
+        "precision": 0.9427750999897446,
+        "recall": 0.877780960565263,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9418920277470594,
-      "eval_loss": 0.19743724167346954,
+      "eval_accuracy": 0.9403880597014925,
+      "eval_loss": 0.22083505988121033,
       "eval_macro avg": {
-        "f1-score": 0.920304053017292,
-        "precision": 0.9275105831343932,
-        "recall": 0.9147612117886741,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6615,
-      "eval_samples_per_second": 17.162,
-      "eval_steps_per_second": 2.145,
+        "f1-score": 0.9161461296773631,
+        "precision": 0.9129519750337626,
+        "recall": 0.9209677907339654,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7449,
+      "eval_samples_per_second": 17.071,
+      "eval_steps_per_second": 2.318,
       "eval_weighted avg": {
-        "f1-score": 0.9411036309913238,
-        "precision": 0.9419713163409645,
-        "recall": 0.9418920277470594,
-        "support": 29841.0
+        "f1-score": 0.9399744505088647,
+        "precision": 0.9406253819936745,
+        "recall": 0.9403880597014925,
+        "support": 33500.0
       },
       "step": 324
     },
     {
       "epoch": 5.0,
       "eval_B": {
-        "f1-score": 0.8953846153846153,
-        "precision": 0.8919540229885058,
-        "recall": 0.8988416988416988,
-        "support": 1295.0
+        "f1-score": 0.8879781420765027,
+        "precision": 0.8736559139784946,
+        "recall": 0.9027777777777778,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9602426713742261,
-        "precision": 0.9581224570804803,
-        "recall": 0.9623722900573137,
-        "support": 20065.0
+        "f1-score": 0.9613201259451634,
+        "precision": 0.9539317642765919,
+        "recall": 0.9688238291564367,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9022119433078337,
-        "precision": 0.9075399665950847,
-        "recall": 0.8969461148449476,
-        "support": 8481.0
+        "f1-score": 0.916006030835076,
+        "precision": 0.9334853291038858,
+        "recall": 0.899169292466342,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9410207432726785,
-      "eval_loss": 0.2340461015701294,
+      "eval_accuracy": 0.9442089552238806,
+      "eval_loss": 0.2257838249206543,
       "eval_macro avg": {
-        "f1-score": 0.9192797433555584,
-        "precision": 0.9192054822213569,
-        "recall": 0.9193867012479867,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6719,
-      "eval_samples_per_second": 17.124,
-      "eval_steps_per_second": 2.14,
+        "f1-score": 0.921768099618914,
+        "precision": 0.9203576691196576,
+        "recall": 0.9235902998001855,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7845,
+      "eval_samples_per_second": 16.93,
+      "eval_steps_per_second": 2.299,
       "eval_weighted avg": {
-        "f1-score": 0.9409353496612266,
-        "precision": 0.9408751053176122,
-        "recall": 0.9410207432726785,
-        "support": 29841.0
+        "f1-score": 0.9440011117702735,
+        "precision": 0.9440889959185916,
+        "recall": 0.9442089552238806,
+        "support": 33500.0
       },
       "step": 405
     },
     {
       "epoch": 6.0,
       "eval_B": {
-        "f1-score": 0.8929831438651511,
-        "precision": 0.9068471337579618,
-        "recall": 0.8795366795366796,
-        "support": 1295.0
+        "f1-score": 0.8873579056148813,
+        "precision": 0.8803827751196173,
+        "recall": 0.8944444444444445,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9522293798155866,
-        "precision": 0.9655225409836066,
-        "recall": 0.9392972838275604,
-        "support": 20065.0
+        "f1-score": 0.9622060984570168,
+        "precision": 0.9539267015706806,
+        "recall": 0.9706304720433594,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.8885216003647556,
-        "precision": 0.8599007170435742,
-        "recall": 0.919113312109421,
-        "support": 8481.0
+        "f1-score": 0.9182769530299342,
+        "precision": 0.9365567911040509,
+        "recall": 0.9006970304592762,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.930967460875976,
-      "eval_loss": 0.21774999797344208,
+      "eval_accuracy": 0.9454925373134329,
+      "eval_loss": 0.2242102175951004,
       "eval_macro avg": {
-        "f1-score": 0.9112447080151643,
-        "precision": 0.9107567972617142,
-        "recall": 0.9126490918245537,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.647,
-      "eval_samples_per_second": 17.215,
-      "eval_steps_per_second": 2.152,
+        "f1-score": 0.9226136523672773,
+        "precision": 0.923622089264783,
+        "recall": 0.9219239823156933,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7715,
+      "eval_samples_per_second": 16.976,
+      "eval_steps_per_second": 2.305,
       "eval_weighted avg": {
-        "f1-score": 0.9315521386682287,
-        "precision": 0.9329578366776978,
-        "recall": 0.930967460875976,
-        "support": 29841.0
+        "f1-score": 0.945255312255509,
+        "precision": 0.9453351097376493,
+        "recall": 0.9454925373134329,
+        "support": 33500.0
       },
       "step": 486
     },
     {
       "epoch": 6.17,
-      "grad_norm": 9.310327529907227,
+      "grad_norm": 1.8705081939697266,
       "learning_rate": 1.7530864197530865e-05,
-      "loss": 0.1627,
+      "loss": 0.1601,
       "step": 500
     },
     {
       "epoch": 7.0,
       "eval_B": {
-        "f1-score": 0.8936490041337843,
-        "precision": 0.8704245973645681,
-        "recall": 0.9181467181467181,
-        "support": 1295.0
+        "f1-score": 0.8876749743939911,
+        "precision": 0.8730691739422431,
+        "recall": 0.9027777777777778,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9599271850630982,
-        "precision": 0.9477800446905664,
-        "recall": 0.9723897333665587,
-        "support": 20065.0
+        "f1-score": 0.9611415113943341,
+        "precision": 0.9581108451482232,
+        "recall": 0.9641914114976606,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.8979841172877214,
-        "precision": 0.9316770186335404,
-        "recall": 0.8666430845419172,
-        "support": 8481.0
+        "f1-score": 0.9164739884393064,
+        "precision": 0.9247594050743657,
+        "recall": 0.9083357204239473,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9399819040916859,
-      "eval_loss": 0.27376389503479004,
+      "eval_accuracy": 0.944089552238806,
+      "eval_loss": 0.22517083585262299,
       "eval_macro avg": {
-        "f1-score": 0.9171867688282013,
-        "precision": 0.9166272202295582,
-        "recall": 0.9190598453517312,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6671,
-      "eval_samples_per_second": 17.141,
-      "eval_steps_per_second": 2.143,
+        "f1-score": 0.9217634914092105,
+        "precision": 0.9186464747216107,
+        "recall": 0.9251016365664619,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7926,
+      "eval_samples_per_second": 16.901,
+      "eval_steps_per_second": 2.295,
       "eval_weighted avg": {
-        "f1-score": 0.9394463230910989,
-        "precision": 0.9398464945991887,
-        "recall": 0.9399819040916859,
-        "support": 29841.0
+        "f1-score": 0.9440192791200506,
+        "precision": 0.9440287663891153,
+        "recall": 0.944089552238806,
+        "support": 33500.0
       },
       "step": 567
     },
     {
       "epoch": 8.0,
       "eval_B": {
-        "f1-score": 0.8948545861297539,
-        "precision": 0.8651766402307137,
-        "recall": 0.9266409266409267,
-        "support": 1295.0
+        "f1-score": 0.894540522210919,
+        "precision": 0.8740888005301524,
+        "recall": 0.9159722222222222,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9590139880661254,
-        "precision": 0.9414702069429106,
-        "recall": 0.9772240219287316,
-        "support": 20065.0
+        "f1-score": 0.9638018385639943,
+        "precision": 0.9540255968049378,
+        "recall": 0.9737805160513272,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.8927241122423641,
-        "precision": 0.9427035531663825,
-        "recall": 0.8477773847423653,
-        "support": 8481.0
+        "f1-score": 0.9203132648066569,
+        "precision": 0.9441598875163202,
+        "recall": 0.8976415544734078,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9382393351429241,
-      "eval_loss": 0.3103167414665222,
+      "eval_accuracy": 0.9474925373134329,
+      "eval_loss": 0.24869142472743988,
       "eval_macro avg": {
-        "f1-score": 0.9155308954794145,
-        "precision": 0.916450133446669,
-        "recall": 0.9172141111040079,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6339,
-      "eval_samples_per_second": 17.264,
-      "eval_steps_per_second": 2.158,
+        "f1-score": 0.9262185418605234,
+        "precision": 0.9240914282838034,
+        "recall": 0.9291314309156524,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7743,
+      "eval_samples_per_second": 16.966,
+      "eval_steps_per_second": 2.304,
       "eval_weighted avg": {
-        "f1-score": 0.9373896838414373,
-        "precision": 0.9385098450391195,
-        "recall": 0.9382393351429241,
-        "support": 29841.0
+        "f1-score": 0.947228939205516,
+        "precision": 0.9475052218791054,
+        "recall": 0.9474925373134329,
+        "support": 33500.0
       },
       "step": 648
     },
     {
       "epoch": 9.0,
       "eval_B": {
-        "f1-score": 0.9011235955056179,
-        "precision": 0.8749090909090909,
-        "recall": 0.9289575289575289,
-        "support": 1295.0
+        "f1-score": 0.8851236868858015,
+        "precision": 0.8643282594308405,
+        "recall": 0.9069444444444444,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9615119937082186,
-        "precision": 0.9485040973670175,
-        "recall": 0.9748816346872664,
-        "support": 20065.0
+        "f1-score": 0.9617001723147618,
+        "precision": 0.9540067462849849,
+        "recall": 0.9695186918052532,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9013722126929674,
-        "precision": 0.9380339155935229,
-        "recall": 0.8674684589081476,
-        "support": 8481.0
+        "f1-score": 0.9165854609237966,
+        "precision": 0.9358272808675754,
+        "recall": 0.8981189725961998,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9423611809255722,
-      "eval_loss": 0.32883673906326294,
+      "eval_accuracy": 0.9445074626865672,
+      "eval_loss": 0.3045330345630646,
       "eval_macro avg": {
-        "f1-score": 0.9213359339689347,
-        "precision": 0.9204823679565438,
-        "recall": 0.9237692075176476,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6714,
-      "eval_samples_per_second": 17.126,
-      "eval_steps_per_second": 2.141,
+        "f1-score": 0.9211364400414532,
+        "precision": 0.9180540955278002,
+        "recall": 0.9248607029486324,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7599,
+      "eval_samples_per_second": 17.017,
+      "eval_steps_per_second": 2.311,
       "eval_weighted avg": {
-        "f1-score": 0.9417992341337168,
-        "precision": 0.9423346276781992,
-        "recall": 0.9423611809255722,
-        "support": 29841.0
+        "f1-score": 0.9443044555560967,
+        "precision": 0.9444685205421044,
+        "recall": 0.9445074626865672,
+        "support": 33500.0
       },
       "step": 729
     },
     {
       "epoch": 10.0,
       "eval_B": {
-        "f1-score": 0.8935361216730039,
-        "precision": 0.8801498127340824,
-        "recall": 0.9073359073359073,
-        "support": 1295.0
+        "f1-score": 0.8815068493150685,
+        "precision": 0.8695945945945946,
+        "recall": 0.89375,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9632332861646931,
-        "precision": 0.9596359319351009,
-        "recall": 0.9668577124345876,
-        "support": 20065.0
+        "f1-score": 0.956297959947595,
+        "precision": 0.9660159757999717,
+        "recall": 0.9467735211006625,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9095462405342556,
-        "precision": 0.9200241254523522,
-        "recall": 0.8993043273198915,
-        "support": 8481.0
+        "f1-score": 0.9101049868766404,
+        "precision": 0.8937678357728068,
+        "recall": 0.9270505108373914,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9450755671726819,
-      "eval_loss": 0.3041447699069977,
+      "eval_accuracy": 0.9383283582089552,
+      "eval_loss": 0.32627996802330017,
       "eval_macro avg": {
-        "f1-score": 0.9221052161239841,
-        "precision": 0.9199366233738452,
-        "recall": 0.9244993156967954,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6532,
-      "eval_samples_per_second": 17.193,
-      "eval_steps_per_second": 2.149,
+        "f1-score": 0.9159699320464346,
+        "precision": 0.909792802055791,
+        "recall": 0.9225246773126846,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7534,
+      "eval_samples_per_second": 17.04,
+      "eval_steps_per_second": 2.314,
       "eval_weighted avg": {
-        "f1-score": 0.9449504651463465,
-        "precision": 0.9449285744355027,
-        "recall": 0.9450755671726819,
-        "support": 29841.0
+        "f1-score": 0.9386418940884026,
+        "precision": 0.9392845859659347,
+        "recall": 0.9383283582089552,
+        "support": 33500.0
       },
       "step": 810
     },
     {
       "epoch": 11.0,
       "eval_B": {
-        "f1-score": 0.8924122310305775,
-        "precision": 0.8729689807976366,
-        "recall": 0.9127413127413128,
-        "support": 1295.0
+        "f1-score": 0.8901136755080951,
+        "precision": 0.8831168831168831,
+        "recall": 0.8972222222222223,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9588723786993913,
-        "precision": 0.9483330083837005,
-        "recall": 0.9696486419137802,
-        "support": 20065.0
+        "f1-score": 0.9620073001078947,
+        "precision": 0.9535359970874671,
+        "recall": 0.9706304720433594,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.8956965718453683,
-        "precision": 0.9243507715468574,
-        "recall": 0.8687654757693668,
-        "support": 8481.0
+        "f1-score": 0.917413322945072,
+        "precision": 0.9361025539103647,
+        "recall": 0.8994557433400172,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9385074226735028,
-      "eval_loss": 0.3531416654586792,
+      "eval_accuracy": 0.9452238805970149,
+      "eval_loss": 0.3123624622821808,
       "eval_macro avg": {
-        "f1-score": 0.9156603938584458,
-        "precision": 0.9152175869093981,
-        "recall": 0.9170518101414866,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6508,
-      "eval_samples_per_second": 17.201,
-      "eval_steps_per_second": 2.15,
+        "f1-score": 0.9231780995203539,
+        "precision": 0.9242518113715716,
+        "recall": 0.922436145868533,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7539,
+      "eval_samples_per_second": 17.039,
+      "eval_steps_per_second": 2.314,
       "eval_weighted avg": {
-        "f1-score": 0.938033267772811,
-        "precision": 0.9382465579853486,
-        "recall": 0.9385074226735028,
-        "support": 29841.0
+        "f1-score": 0.9449756719810304,
+        "precision": 0.9450588635199616,
+        "recall": 0.9452238805970149,
+        "support": 33500.0
       },
       "step": 891
     },
     {
       "epoch": 12.0,
       "eval_B": {
-        "f1-score": 0.8970199924556771,
-        "precision": 0.8768436578171092,
-        "recall": 0.9181467181467181,
-        "support": 1295.0
+        "f1-score": 0.8800819952169457,
+        "precision": 0.8661735036987223,
+        "recall": 0.8944444444444445,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9595004897159647,
-        "precision": 0.9431046931407943,
-        "recall": 0.9764764515325193,
-        "support": 20065.0
+        "f1-score": 0.9595463311598861,
+        "precision": 0.9512039692293687,
+        "recall": 0.9680363181544448,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.8953122104873078,
-        "precision": 0.9400778210116731,
-        "recall": 0.8546162009197029,
-        "support": 8481.0
+        "f1-score": 0.9115367743822197,
+        "precision": 0.9310035842293907,
+        "recall": 0.8928673732454884,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9393116852652391,
-      "eval_loss": 0.4237186312675476,
+      "eval_accuracy": 0.9413731343283582,
+      "eval_loss": 0.3427491784095764,
       "eval_macro avg": {
-        "f1-score": 0.9172775642196499,
-        "precision": 0.9200087239898588,
-        "recall": 0.9164131235329801,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6388,
-      "eval_samples_per_second": 17.246,
-      "eval_steps_per_second": 2.156,
+        "f1-score": 0.9170550335863505,
+        "precision": 0.9161270190524938,
+        "recall": 0.9184493786147926,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7455,
+      "eval_samples_per_second": 17.069,
+      "eval_steps_per_second": 2.318,
       "eval_weighted avg": {
-        "f1-score": 0.9385463313402296,
-        "precision": 0.9393689288141548,
-        "recall": 0.9393116852652391,
-        "support": 29841.0
+        "f1-score": 0.9411214734915179,
+        "precision": 0.9412337452750731,
+        "recall": 0.9413731343283582,
+        "support": 33500.0
       },
       "step": 972
     },
     {
       "epoch": 12.35,
-      "grad_norm": 0.5223535895347595,
+      "grad_norm": 0.47551417350769043,
       "learning_rate": 1.506172839506173e-05,
-      "loss": 0.0196,
+      "loss": 0.0222,
       "step": 1000
     },
     {
       "epoch": 13.0,
       "eval_B": {
-        "f1-score": 0.8891411275066211,
-        "precision": 0.8716617210682492,
-        "recall": 0.9073359073359073,
-        "support": 1295.0
+        "f1-score": 0.8896434634974534,
+        "precision": 0.8704318936877077,
+        "recall": 0.9097222222222222,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9610692322951795,
-        "precision": 0.9563736860287223,
-        "recall": 0.9658111138798904,
-        "support": 20065.0
+        "f1-score": 0.9620136667970457,
+        "precision": 0.9556591698665204,
+        "recall": 0.9684532357437347,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9043145233678416,
-        "precision": 0.918104495747266,
-        "recall": 0.8909326730338404,
-        "support": 8481.0
+        "f1-score": 0.9183177933177933,
+        "precision": 0.9343808676746714,
+        "recall": 0.9027976701995608,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9419925605710264,
-      "eval_loss": 0.4308999180793762,
+      "eval_accuracy": 0.9454029850746268,
+      "eval_loss": 0.3816670775413513,
       "eval_macro avg": {
-        "f1-score": 0.9181749610565474,
-        "precision": 0.9153799676147458,
-        "recall": 0.9213598980832126,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6555,
-      "eval_samples_per_second": 17.184,
-      "eval_steps_per_second": 2.148,
+        "f1-score": 0.9233249745374308,
+        "precision": 0.9201573104096331,
+        "recall": 0.9269910427218392,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7247,
+      "eval_samples_per_second": 17.144,
+      "eval_steps_per_second": 2.328,
       "eval_weighted avg": {
-        "f1-score": 0.9418177466843106,
-        "precision": 0.9418211242043584,
-        "recall": 0.9419925605710264,
-        "support": 29841.0
+        "f1-score": 0.945242324238848,
+        "precision": 0.9453435001186778,
+        "recall": 0.9454029850746268,
+        "support": 33500.0
       },
       "step": 1053
     },
     {
       "epoch": 14.0,
       "eval_B": {
-        "f1-score": 0.9025601834161253,
-        "precision": 0.8933434190620272,
-        "recall": 0.911969111969112,
-        "support": 1295.0
+        "f1-score": 0.8865272540281112,
+        "precision": 0.8754231550440081,
+        "recall": 0.8979166666666667,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.960806215404473,
-        "precision": 0.9570764513895758,
-        "recall": 0.9645651632195366,
-        "support": 20065.0
+        "f1-score": 0.9599206715247671,
+        "precision": 0.9557331129172981,
+        "recall": 0.9641450873210728,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9030873763261413,
-        "precision": 0.9131011208870676,
-        "recall": 0.8932908855087843,
-        "support": 8481.0
+        "f1-score": 0.9141367826632559,
+        "precision": 0.9242631270739801,
+        "recall": 0.9042299245679366,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9420260715123487,
-      "eval_loss": 0.4198700487613678,
+      "eval_accuracy": 0.9425671641791045,
+      "eval_loss": 0.3827010989189148,
       "eval_macro avg": {
-        "f1-score": 0.9221512583822467,
-        "precision": 0.9211736637795568,
-        "recall": 0.9232750535658111,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6834,
-      "eval_samples_per_second": 17.082,
-      "eval_steps_per_second": 2.135,
+        "f1-score": 0.9201949027387114,
+        "precision": 0.9184731316784287,
+        "recall": 0.9220972261852255,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7319,
+      "eval_samples_per_second": 17.118,
+      "eval_steps_per_second": 2.325,
       "eval_weighted avg": {
-        "f1-score": 0.9418744743217934,
-        "precision": 0.9418125843993292,
-        "recall": 0.9420260715123487,
-        "support": 29841.0
+        "f1-score": 0.9424525763235196,
+        "precision": 0.9424426203479068,
+        "recall": 0.9425671641791045,
+        "support": 33500.0
       },
       "step": 1134
     },
     {
       "epoch": 15.0,
       "eval_B": {
-        "f1-score": 0.8935681470137825,
-        "precision": 0.8861047835990888,
-        "recall": 0.9011583011583012,
-        "support": 1295.0
+        "f1-score": 0.8871919472405415,
+        "precision": 0.8868841082581541,
+        "recall": 0.8875,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9633398694432723,
-        "precision": 0.9595530063291139,
-        "recall": 0.9671567405930726,
-        "support": 20065.0
+        "f1-score": 0.9616856471022177,
+        "precision": 0.9571841585975862,
+        "recall": 0.9662296752675221,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9100768726535964,
-        "precision": 0.92,
-        "recall": 0.9003655229336163,
-        "support": 8481.0
+        "f1-score": 0.9181813798756087,
+        "precision": 0.9273470977795092,
+        "recall": 0.9091950730449728,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9453101437619382,
-      "eval_loss": 0.39179423451423645,
+      "eval_accuracy": 0.9450149253731344,
+      "eval_loss": 0.39016807079315186,
       "eval_macro avg": {
-        "f1-score": 0.9223282963702171,
-        "precision": 0.9218859299760677,
-        "recall": 0.9228935215616634,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6645,
-      "eval_samples_per_second": 17.151,
-      "eval_steps_per_second": 2.144,
+        "f1-score": 0.9223529914061226,
+        "precision": 0.9238051215450831,
+        "recall": 0.920974916104165,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7376,
+      "eval_samples_per_second": 17.097,
+      "eval_steps_per_second": 2.322,
       "eval_weighted avg": {
-        "f1-score": 0.945174330208011,
-        "precision": 0.9451243847979119,
-        "recall": 0.9453101437619382,
-        "support": 29841.0
+        "f1-score": 0.944882927148036,
+        "precision": 0.9448344388830637,
+        "recall": 0.9450149253731344,
+        "support": 33500.0
       },
       "step": 1215
     },
     {
       "epoch": 16.0,
       "eval_B": {
-        "f1-score": 0.9022614028363357,
-        "precision": 0.895738203957382,
-        "recall": 0.9088803088803089,
-        "support": 1295.0
+        "f1-score": 0.8897338403041825,
+        "precision": 0.8857536132140399,
+        "recall": 0.89375,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9632045864531594,
-        "precision": 0.9552494853445741,
-        "recall": 0.9712932967854473,
-        "support": 20065.0
+        "f1-score": 0.9608010520729992,
+        "precision": 0.9570693142121713,
+        "recall": 0.9645620049103627,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9083463808262073,
-        "precision": 0.9282461538461538,
-        "recall": 0.8892819243013795,
-        "support": 8481.0
+        "f1-score": 0.9159121556540166,
+        "precision": 0.9240112719852298,
+        "recall": 0.9079537859257137,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9452766328206159,
-      "eval_loss": 0.44052162766456604,
+      "eval_accuracy": 0.943820895522388,
+      "eval_loss": 0.40087994933128357,
       "eval_macro avg": {
-        "f1-score": 0.9246041233719008,
-        "precision": 0.92641128104937,
-        "recall": 0.9231518433223785,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6376,
-      "eval_samples_per_second": 17.25,
-      "eval_steps_per_second": 2.156,
+        "f1-score": 0.9221490160103993,
+        "precision": 0.9222780664704805,
+        "recall": 0.9220885969453588,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.8145,
+      "eval_samples_per_second": 16.824,
+      "eval_steps_per_second": 2.285,
       "eval_weighted avg": {
-        "f1-score": 0.9449688080038458,
-        "precision": 0.944992377210312,
-        "recall": 0.9452766328206159,
-        "support": 29841.0
+        "f1-score": 0.9437127476806678,
+        "precision": 0.9436689713560499,
+        "recall": 0.943820895522388,
+        "support": 33500.0
       },
       "step": 1296
     },
     {
       "epoch": 17.0,
       "eval_B": {
-        "f1-score": 0.8923649906890131,
-        "precision": 0.8618705035971223,
-        "recall": 0.9250965250965251,
-        "support": 1295.0
+        "f1-score": 0.8904347826086957,
+        "precision": 0.89198606271777,
+        "recall": 0.8888888888888888,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9599605522682445,
-        "precision": 0.9498902171261283,
-        "recall": 0.97024669823075,
-        "support": 20065.0
+        "f1-score": 0.9615171394252979,
+        "precision": 0.9589034784611841,
+        "recall": 0.9641450873210728,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9001642635517431,
-        "precision": 0.9298642533936652,
-        "recall": 0.8723027944817828,
-        "support": 8481.0
+        "f1-score": 0.9182546920750733,
+        "precision": 0.9232625482625483,
+        "recall": 0.9133008689009835,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9404510572701987,
-      "eval_loss": 0.4657233655452728,
+      "eval_accuracy": 0.9450149253731344,
+      "eval_loss": 0.42787015438079834,
       "eval_macro avg": {
-        "f1-score": 0.9174966021696669,
-        "precision": 0.9138749913723053,
-        "recall": 0.9225486726030193,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6396,
-      "eval_samples_per_second": 17.243,
-      "eval_steps_per_second": 2.155,
+        "f1-score": 0.9234022047030224,
+        "precision": 0.9247173631471673,
+        "recall": 0.9221116150369818,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7453,
+      "eval_samples_per_second": 17.07,
+      "eval_steps_per_second": 2.318,
       "eval_weighted avg": {
-        "f1-score": 0.940032648483192,
-        "precision": 0.9403789498282803,
-        "recall": 0.9404510572701987,
-        "support": 29841.0
+        "f1-score": 0.9449366556964975,
+        "precision": 0.9448847160539952,
+        "recall": 0.9450149253731344,
+        "support": 33500.0
       },
       "step": 1377
     },
     {
       "epoch": 18.0,
       "eval_B": {
-        "f1-score": 0.9014302280633939,
-        "precision": 0.9024767801857585,
-        "recall": 0.9003861003861003,
-        "support": 1295.0
+        "f1-score": 0.8895027624309393,
+        "precision": 0.8846153846153846,
+        "recall": 0.8944444444444445,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9638021281301453,
-        "precision": 0.9506496024820632,
-        "recall": 0.9773236979815599,
-        "support": 20065.0
+        "f1-score": 0.9607206210979795,
+        "precision": 0.9555912007332722,
+        "recall": 0.9659054060314078,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9083038653822705,
-        "precision": 0.9404115641964399,
-        "recall": 0.87831623629289,
-        "support": 8481.0
+        "f1-score": 0.9159781610861478,
+        "precision": 0.9271322378716745,
+        "recall": 0.905089277188962,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9458463188230958,
-      "eval_loss": 0.4731404185295105,
+      "eval_accuracy": 0.943820895522388,
+      "eval_loss": 0.422620564699173,
       "eval_macro avg": {
-        "f1-score": 0.9245120738586032,
-        "precision": 0.9311793156214204,
-        "recall": 0.9186753448868501,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6405,
-      "eval_samples_per_second": 17.24,
-      "eval_steps_per_second": 2.155,
+        "f1-score": 0.9220671815383555,
+        "precision": 0.9224462744067772,
+        "recall": 0.9218130425549381,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7703,
+      "eval_samples_per_second": 16.98,
+      "eval_steps_per_second": 2.306,
       "eval_weighted avg": {
-        "f1-score": 0.9453224398840687,
-        "precision": 0.9456493475450943,
-        "recall": 0.9458463188230958,
-        "support": 29841.0
+        "f1-score": 0.9436716210924735,
+        "precision": 0.9436432636210552,
+        "recall": 0.943820895522388,
+        "support": 33500.0
       },
       "step": 1458
     },
     {
       "epoch": 18.52,
-      "grad_norm": 1.548362135887146,
+      "grad_norm": 0.1027584820985794,
       "learning_rate": 1.2592592592592593e-05,
-      "loss": 0.0038,
+      "loss": 0.0054,
       "step": 1500
     },
     {
       "epoch": 19.0,
       "eval_B": {
-        "f1-score": 0.8891402714932127,
-        "precision": 0.868828297715549,
-        "recall": 0.9104247104247104,
-        "support": 1295.0
+        "f1-score": 0.8924247665167762,
+        "precision": 0.8890420399724328,
+        "recall": 0.8958333333333334,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9603099696984749,
-        "precision": 0.9571718572065158,
-        "recall": 0.9634687266384251,
-        "support": 20065.0
+        "f1-score": 0.9599648245857633,
+        "precision": 0.959121387283237,
+        "recall": 0.9608097466067541,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9027910305343512,
-        "precision": 0.913358271992277,
-        "recall": 0.8924655111425539,
-        "support": 8481.0
+        "f1-score": 0.9158252380724505,
+        "precision": 0.9179777436684574,
+        "recall": 0.913682803399217,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9409872323313562,
-      "eval_loss": 0.49517783522605896,
+      "eval_accuracy": 0.9432835820895522,
+      "eval_loss": 0.4375981390476227,
       "eval_macro avg": {
-        "f1-score": 0.9174137572420129,
-        "precision": 0.9131194756381139,
-        "recall": 0.9221196494018965,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6377,
-      "eval_samples_per_second": 17.25,
-      "eval_steps_per_second": 2.156,
+        "f1-score": 0.9227382763916633,
+        "precision": 0.9220470569747091,
+        "recall": 0.9234419611131015,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7487,
+      "eval_samples_per_second": 17.057,
+      "eval_steps_per_second": 2.316,
       "eval_weighted avg": {
-        "f1-score": 0.9408741973642117,
-        "precision": 0.9408859443435835,
-        "recall": 0.9409872323313562,
-        "support": 29841.0
+        "f1-score": 0.9432623895656658,
+        "precision": 0.9432464129636805,
+        "recall": 0.9432835820895522,
+        "support": 33500.0
       },
       "step": 1539
     },
     {
       "epoch": 20.0,
       "eval_B": {
-        "f1-score": 0.899884925201381,
-        "precision": 0.8940548780487805,
-        "recall": 0.9057915057915058,
-        "support": 1295.0
+        "f1-score": 0.8907273354015857,
+        "precision": 0.8843258042436687,
+        "recall": 0.8972222222222223,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9634953846153845,
-        "precision": 0.951896887159533,
-        "recall": 0.9753800149514079,
-        "support": 20065.0
+        "f1-score": 0.9620682542229149,
+        "precision": 0.9520958083832335,
+        "recall": 0.972251818223931,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9073556231003039,
-        "precision": 0.9365039528171665,
-        "recall": 0.8799669850253508,
-        "support": 8481.0
+        "f1-score": 0.9177252296267345,
+        "precision": 0.9396698349174587,
+        "recall": 0.8967822018523823,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9452431218792936,
-      "eval_loss": 0.45907479524612427,
+      "eval_accuracy": 0.9454328358208955,
+      "eval_loss": 0.44444534182548523,
       "eval_macro avg": {
-        "f1-score": 0.9235786443056897,
-        "precision": 0.9274852393418267,
-        "recall": 0.9203795019227549,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6356,
-      "eval_samples_per_second": 17.258,
-      "eval_steps_per_second": 2.157,
+        "f1-score": 0.9235069397504118,
+        "precision": 0.9253638158481202,
+        "recall": 0.922085414099512,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7631,
+      "eval_samples_per_second": 17.006,
+      "eval_steps_per_second": 2.309,
       "eval_weighted avg": {
-        "f1-score": 0.9447796290324438,
-        "precision": 0.9450119678218422,
-        "recall": 0.9452431218792936,
-        "support": 29841.0
+        "f1-score": 0.9451388387095264,
+        "precision": 0.9452980165603072,
+        "recall": 0.9454328358208955,
+        "support": 33500.0
       },
       "step": 1620
     },
     {
       "epoch": 21.0,
       "eval_B": {
-        "f1-score": 0.8944337811900192,
-        "precision": 0.8893129770992366,
-        "recall": 0.8996138996138996,
-        "support": 1295.0
+        "f1-score": 0.890790825059911,
+        "precision": 0.8784604996623903,
+        "recall": 0.9034722222222222,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9624703322784811,
-        "precision": 0.9549624687239366,
-        "recall": 0.9700971841515076,
-        "support": 20065.0
+        "f1-score": 0.9615375783600082,
+        "precision": 0.9533284764593388,
+        "recall": 0.9698892852179553,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9066089361958025,
-        "precision": 0.9251350024545901,
-        "recall": 0.8888102818063908,
-        "support": 8481.0
+        "f1-score": 0.9162201656113005,
+        "precision": 0.9351695336581486,
+        "recall": 0.8980234889716414,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9439361951677223,
-      "eval_loss": 0.4614206850528717,
+      "eval_accuracy": 0.9445671641791045,
+      "eval_loss": 0.4363422691822052,
       "eval_macro avg": {
-        "f1-score": 0.9211710165547675,
-        "precision": 0.9231368160925877,
-        "recall": 0.9195071218572659,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6636,
-      "eval_samples_per_second": 17.154,
-      "eval_steps_per_second": 2.144,
+        "f1-score": 0.9228495230104065,
+        "precision": 0.9223195032599593,
+        "recall": 0.9237949988039397,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7902,
+      "eval_samples_per_second": 16.91,
+      "eval_steps_per_second": 2.296,
       "eval_weighted avg": {
-        "f1-score": 0.9436416122678664,
-        "precision": 0.9436363458364894,
-        "recall": 0.9439361951677223,
-        "support": 29841.0
+        "f1-score": 0.9443290831818184,
+        "precision": 0.9444332975177129,
+        "recall": 0.9445671641791045,
+        "support": 33500.0
       },
       "step": 1701
     },
     {
       "epoch": 22.0,
       "eval_B": {
-        "f1-score": 0.9011072928598702,
-        "precision": 0.8912386706948641,
-        "recall": 0.9111969111969112,
-        "support": 1295.0
+        "f1-score": 0.8907216494845361,
+        "precision": 0.8816326530612245,
+        "recall": 0.9,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9635141147970535,
-        "precision": 0.9558585511795576,
-        "recall": 0.9712932967854473,
-        "support": 20065.0
+        "f1-score": 0.961758728985775,
+        "precision": 0.9588801399825022,
+        "recall": 0.9646546532635383,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9095068938527304,
-        "precision": 0.9292568897637795,
-        "recall": 0.8905789411625987,
-        "support": 8481.0
+        "f1-score": 0.9186952756663139,
+        "precision": 0.9258217783380199,
+        "recall": 0.9116776472834909,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9457457859991287,
-      "eval_loss": 0.4724768102169037,
+      "eval_accuracy": 0.9453134328358209,
+      "eval_loss": 0.3998025953769684,
       "eval_macro avg": {
-        "f1-score": 0.9247094338365515,
-        "precision": 0.9254513705460671,
-        "recall": 0.9243563830483191,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6333,
-      "eval_samples_per_second": 17.266,
-      "eval_steps_per_second": 2.158,
+        "f1-score": 0.9237252180455416,
+        "precision": 0.9221115237939155,
+        "recall": 0.925444100182343,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7313,
+      "eval_samples_per_second": 17.12,
+      "eval_steps_per_second": 2.325,
       "eval_weighted avg": {
-        "f1-score": 0.9454566745223489,
-        "precision": 0.9454939040264833,
-        "recall": 0.9457457859991287,
-        "support": 29841.0
+        "f1-score": 0.9452424023858794,
+        "precision": 0.9452247190072991,
+        "recall": 0.9453134328358209,
+        "support": 33500.0
       },
       "step": 1782
     },
     {
       "epoch": 23.0,
       "eval_B": {
-        "f1-score": 0.8956916099773242,
-        "precision": 0.8771280532938565,
-        "recall": 0.915057915057915,
-        "support": 1295.0
+        "f1-score": 0.8938053097345132,
+        "precision": 0.8765020026702269,
+        "recall": 0.9118055555555555,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9624542808091362,
-        "precision": 0.9609957269204015,
-        "recall": 0.9639172688761525,
-        "support": 20065.0
+        "f1-score": 0.9620159689850926,
+        "precision": 0.9584770313146641,
+        "recall": 0.9655811367952934,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9084001187295934,
-        "precision": 0.914753706360593,
-        "recall": 0.9021341822898243,
-        "support": 8481.0
+        "f1-score": 0.918660748745658,
+        "precision": 0.9284251584592882,
+        "recall": 0.9090995894204144,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9442377936396233,
-      "eval_loss": 0.47468113899230957,
+      "eval_accuracy": 0.9456119402985075,
+      "eval_loss": 0.4420088231563568,
       "eval_macro avg": {
-        "f1-score": 0.9221820031720179,
-        "precision": 0.9176258288582838,
-        "recall": 0.927036455407964,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6476,
-      "eval_samples_per_second": 17.213,
-      "eval_steps_per_second": 2.152,
+        "f1-score": 0.9248273424884212,
+        "precision": 0.9211347308147264,
+        "recall": 0.9288287605904211,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7375,
+      "eval_samples_per_second": 17.098,
+      "eval_steps_per_second": 2.322,
       "eval_weighted avg": {
-        "f1-score": 0.9441944702356367,
-        "precision": 0.9442138759866489,
-        "recall": 0.9442377936396233,
-        "support": 29841.0
+        "f1-score": 0.9455299220929005,
+        "precision": 0.9455583057725045,
+        "recall": 0.9456119402985075,
+        "support": 33500.0
       },
       "step": 1863
     },
     {
       "epoch": 24.0,
       "eval_B": {
-        "f1-score": 0.9008390541571321,
-        "precision": 0.8899773926149209,
-        "recall": 0.911969111969112,
-        "support": 1295.0
+        "f1-score": 0.89482699554642,
+        "precision": 0.8830290736984449,
+        "recall": 0.9069444444444444,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9639726399486381,
-        "precision": 0.9553151918559123,
-        "recall": 0.9727884375778719,
-        "support": 20065.0
+        "f1-score": 0.962135833103575,
+        "precision": 0.9544605005242285,
+        "recall": 0.969935609394543,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9101008271448409,
-        "precision": 0.9325661964860184,
-        "recall": 0.8886923711826435,
-        "support": 8481.0
+        "f1-score": 0.9185192391885975,
+        "precision": 0.9362356207854026,
+        "recall": 0.9014608994557434,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9462484501189639,
-      "eval_loss": 0.4712923467159271,
+      "eval_accuracy": 0.945820895522388,
+      "eval_loss": 0.4822453558444977,
       "eval_macro avg": {
-        "f1-score": 0.9249708404168704,
-        "precision": 0.9259529269856172,
-        "recall": 0.9244833069098758,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6509,
-      "eval_samples_per_second": 17.201,
-      "eval_steps_per_second": 2.15,
+        "f1-score": 0.9251606892795309,
+        "precision": 0.924575065002692,
+        "recall": 0.9261136510982436,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7345,
+      "eval_samples_per_second": 17.108,
+      "eval_steps_per_second": 2.323,
       "eval_weighted avg": {
-        "f1-score": 0.9459221443892064,
-        "precision": 0.9460143413566277,
-        "recall": 0.9462484501189639,
-        "support": 29841.0
+        "f1-score": 0.9456068386512209,
+        "precision": 0.945692428281427,
+        "recall": 0.945820895522388,
+        "support": 33500.0
       },
       "step": 1944
     },
     {
       "epoch": 24.69,
-      "grad_norm": 0.04072243347764015,
+      "grad_norm": 0.10262901335954666,
       "learning_rate": 1.0123456790123458e-05,
-      "loss": 0.0014,
+      "loss": 0.0037,
       "step": 2000
     },
     {
       "epoch": 25.0,
       "eval_B": {
-        "f1-score": 0.8943768996960487,
-        "precision": 0.8803290949887809,
-        "recall": 0.9088803088803089,
-        "support": 1295.0
+        "f1-score": 0.8864342423196411,
+        "precision": 0.8812628689087165,
+        "recall": 0.8916666666666667,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9629187495336865,
-        "precision": 0.9610305798252582,
-        "recall": 0.9648143533516073,
-        "support": 20065.0
+        "f1-score": 0.9595285234126801,
+        "precision": 0.9537320946409775,
+        "recall": 0.9653958400889424,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9089721512974288,
-        "precision": 0.9155502392344498,
-        "recall": 0.9024879141610659,
-        "support": 8481.0
+        "f1-score": 0.9129445923058311,
+        "precision": 0.9255298273155416,
+        "recall": 0.9006970304592762,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9446734358768137,
-      "eval_loss": 0.48565673828125,
+      "eval_accuracy": 0.942,
+      "eval_loss": 0.46240004897117615,
       "eval_macro avg": {
-        "f1-score": 0.922089266842388,
-        "precision": 0.9189699713494962,
-        "recall": 0.925394192130994,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6314,
-      "eval_samples_per_second": 17.274,
-      "eval_steps_per_second": 2.159,
+        "f1-score": 0.9196357860127174,
+        "precision": 0.9201749302884119,
+        "recall": 0.9192531790716284,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7408,
+      "eval_samples_per_second": 17.086,
+      "eval_steps_per_second": 2.32,
       "eval_weighted avg": {
-        "f1-score": 0.9446122988389564,
-        "precision": 0.944602605179171,
-        "recall": 0.9446734358768137,
-        "support": 29841.0
+        "f1-score": 0.941823171912501,
+        "precision": 0.9418002131259403,
+        "recall": 0.942,
+        "support": 33500.0
       },
       "step": 2025
     },
     {
       "epoch": 26.0,
       "eval_B": {
-        "f1-score": 0.897679726131609,
-        "precision": 0.8845577211394303,
-        "recall": 0.9111969111969112,
-        "support": 1295.0
+        "f1-score": 0.892392049348869,
+        "precision": 0.8809201623815968,
+        "recall": 0.9041666666666667,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.962990450744644,
-        "precision": 0.9560839023431743,
-        "recall": 0.9699975080986792,
-        "support": 20065.0
+        "f1-score": 0.9613188432492171,
+        "precision": 0.9556420233463035,
+        "recall": 0.9670635104461018,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9077024833142927,
-        "precision": 0.9261349693251534,
-        "recall": 0.8899893880438627,
-        "support": 8481.0
+        "f1-score": 0.9172881355932202,
+        "precision": 0.9306278864105335,
+        "recall": 0.904325408192495,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9447069468181362,
-      "eval_loss": 0.4857862889766693,
+      "eval_accuracy": 0.9447462686567164,
+      "eval_loss": 0.4685353934764862,
       "eval_macro avg": {
-        "f1-score": 0.9227908867301818,
-        "precision": 0.9222588642692527,
-        "recall": 0.9237279357798177,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.7085,
-      "eval_samples_per_second": 16.991,
-      "eval_steps_per_second": 2.124,
+        "f1-score": 0.9236663427304355,
+        "precision": 0.9223966907128114,
+        "recall": 0.9251851951017546,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7522,
+      "eval_samples_per_second": 17.045,
+      "eval_steps_per_second": 2.315,
       "eval_weighted avg": {
-        "f1-score": 0.94444299455515,
-        "precision": 0.9444682290887698,
-        "recall": 0.9447069468181362,
-        "support": 29841.0
+        "f1-score": 0.9445908377418215,
+        "precision": 0.9446100073487664,
+        "recall": 0.9447462686567164,
+        "support": 33500.0
       },
       "step": 2106
     },
     {
       "epoch": 27.0,
       "eval_B": {
-        "f1-score": 0.8986280487804879,
-        "precision": 0.8871331828442438,
-        "recall": 0.9104247104247104,
-        "support": 1295.0
+        "f1-score": 0.8878923766816142,
+        "precision": 0.882111034955449,
+        "recall": 0.89375,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9620497586334943,
-        "precision": 0.9557796360059027,
-        "recall": 0.9684026912534264,
-        "support": 20065.0
+        "f1-score": 0.9603469923524712,
+        "precision": 0.9467098748762265,
+        "recall": 0.9743827303469681,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9053591790193843,
-        "precision": 0.9219017355169885,
-        "recall": 0.8893998349251268,
-        "support": 8481.0
+        "f1-score": 0.9131848640126133,
+        "precision": 0.943398147205538,
+        "recall": 0.8848467487825837,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9434335310478872,
-      "eval_loss": 0.4996240735054016,
+      "eval_accuracy": 0.9429253731343283,
+      "eval_loss": 0.5011060833930969,
       "eval_macro avg": {
-        "f1-score": 0.9220123288111223,
-        "precision": 0.9216048514557116,
-        "recall": 0.9227424122010879,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6631,
-      "eval_samples_per_second": 17.156,
-      "eval_steps_per_second": 2.145,
+        "f1-score": 0.9204747443488995,
+        "precision": 0.9240730190124045,
+        "recall": 0.9176598263765173,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7877,
+      "eval_samples_per_second": 16.918,
+      "eval_steps_per_second": 2.298,
       "eval_weighted avg": {
-        "f1-score": 0.9431856481825407,
-        "precision": 0.9431722960745722,
-        "recall": 0.9434335310478872,
-        "support": 29841.0
+        "f1-score": 0.942488376929505,
+        "precision": 0.9428977538797776,
+        "recall": 0.9429253731343283,
+        "support": 33500.0
       },
       "step": 2187
     },
     {
       "epoch": 28.0,
       "eval_B": {
-        "f1-score": 0.8919434898816342,
-        "precision": 0.8821752265861027,
-        "recall": 0.901930501930502,
-        "support": 1295.0
+        "f1-score": 0.8914835164835164,
+        "precision": 0.8817934782608695,
+        "recall": 0.9013888888888889,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9604717405427016,
-        "precision": 0.9630706017938568,
-        "recall": 0.9578868676800398,
-        "support": 20065.0
+        "f1-score": 0.9599201652393882,
+        "precision": 0.9618175053483397,
+        "recall": 0.9580302960114884,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9045243823719267,
-        "precision": 0.9003504672897197,
-        "recall": 0.9087371772196675,
-        "support": 8481.0
+        "f1-score": 0.9164245916472212,
+        "precision": 0.9141174235227056,
+        "recall": 0.9187434355008116,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9414898964511913,
-      "eval_loss": 0.4974842071533203,
+      "eval_accuracy": 0.9433134328358209,
+      "eval_loss": 0.4657354950904846,
       "eval_macro avg": {
-        "f1-score": 0.9189798709320876,
-        "precision": 0.9151987652232264,
-        "recall": 0.9228515156100698,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6318,
-      "eval_samples_per_second": 17.272,
-      "eval_steps_per_second": 2.159,
+        "f1-score": 0.9226094244567086,
+        "precision": 0.919242802377305,
+        "recall": 0.9260542068003962,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7715,
+      "eval_samples_per_second": 16.976,
+      "eval_steps_per_second": 2.305,
       "eval_weighted avg": {
-        "f1-score": 0.941597251442054,
-        "precision": 0.9417345550251954,
-        "recall": 0.9414898964511913,
-        "support": 29841.0
+        "f1-score": 0.9433805259426951,
+        "precision": 0.9434653394986137,
+        "recall": 0.9433134328358209,
+        "support": 33500.0
       },
       "step": 2268
     },
     {
       "epoch": 29.0,
       "eval_B": {
-        "f1-score": 0.9026415094339624,
-        "precision": 0.8826568265682657,
-        "recall": 0.9235521235521236,
-        "support": 1295.0
+        "f1-score": 0.8932038834951457,
+        "precision": 0.8623141564318035,
+        "recall": 0.9263888888888889,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.963459587838005,
-        "precision": 0.9565730005895068,
-        "recall": 0.9704460503364066,
-        "support": 20065.0
+        "f1-score": 0.9610192139335576,
+        "precision": 0.9559099867088318,
+        "recall": 0.9661833510909343,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9098789958461262,
-        "precision": 0.929520295202952,
-        "recall": 0.8910505836575876,
-        "support": 8481.0
+        "f1-score": 0.9164846896685592,
+        "precision": 0.9318136964673377,
+        "recall": 0.9016518667048601,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9458463188230958,
-      "eval_loss": 0.4864692687988281,
+      "eval_accuracy": 0.9442985074626866,
+      "eval_loss": 0.45431333780288696,
       "eval_macro avg": {
-        "f1-score": 0.9253266977060312,
-        "precision": 0.9229167074535748,
-        "recall": 0.9283495858487059,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6138,
-      "eval_samples_per_second": 17.339,
-      "eval_steps_per_second": 2.167,
+        "f1-score": 0.9235692623657542,
+        "precision": 0.9166792798693243,
+        "recall": 0.931408035561561,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7759,
+      "eval_samples_per_second": 16.96,
+      "eval_steps_per_second": 2.303,
       "eval_weighted avg": {
-        "f1-score": 0.9455923443737659,
-        "precision": 0.9456767357277099,
-        "recall": 0.9458463188230958,
-        "support": 29841.0
+        "f1-score": 0.9441814781586727,
+        "precision": 0.9443536152670977,
+        "recall": 0.9442985074626866,
+        "support": 33500.0
       },
       "step": 2349
     },
     {
       "epoch": 30.0,
       "eval_B": {
-        "f1-score": 0.8974650018917897,
-        "precision": 0.8798219584569733,
-        "recall": 0.9158301158301159,
-        "support": 1295.0
+        "f1-score": 0.8953290146607569,
+        "precision": 0.8794373744139317,
+        "recall": 0.9118055555555555,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9618674357267865,
-        "precision": 0.9515704252828716,
-        "recall": 0.9723897333665587,
-        "support": 20065.0
+        "f1-score": 0.9620589989448088,
+        "precision": 0.9528786295269687,
+        "recall": 0.9714179830453513,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9045537340619307,
-        "precision": 0.9324070597070973,
-        "recall": 0.87831623629289,
-        "support": 8481.0
+        "f1-score": 0.9176964782884776,
+        "precision": 0.9394,
+        "recall": 0.896973169101499,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9431989544586308,
-      "eval_loss": 0.5332404375076294,
+      "eval_accuracy": 0.9455820895522388,
+      "eval_loss": 0.4872027635574341,
       "eval_macro avg": {
-        "f1-score": 0.921295390560169,
-        "precision": 0.9212664811489807,
-        "recall": 0.9221786951631882,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6482,
-      "eval_samples_per_second": 17.211,
-      "eval_steps_per_second": 2.151,
+        "f1-score": 0.9250281639646811,
+        "precision": 0.9239053346469669,
+        "recall": 0.926732235900802,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7685,
+      "eval_samples_per_second": 16.987,
+      "eval_steps_per_second": 2.307,
       "eval_weighted avg": {
-        "f1-score": 0.9427836699134436,
-        "precision": 0.9430104317173852,
-        "recall": 0.9431989544586308,
-        "support": 29841.0
+        "f1-score": 0.945321689804427,
+        "precision": 0.9455079699926786,
+        "recall": 0.9455820895522388,
+        "support": 33500.0
       },
       "step": 2430
     },
     {
       "epoch": 30.86,
-      "grad_norm": 0.05446575954556465,
+      "grad_norm": 8.135757446289062,
       "learning_rate": 7.654320987654322e-06,
-      "loss": 0.0008,
+      "loss": 0.0025,
       "step": 2500
     },
     {
       "epoch": 31.0,
       "eval_B": {
-        "f1-score": 0.9002674818494459,
-        "precision": 0.8910741301059002,
-        "recall": 0.9096525096525097,
-        "support": 1295.0
+        "f1-score": 0.8891177480260899,
+        "precision": 0.8791581805838425,
+        "recall": 0.8993055555555556,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9637251998815517,
-        "precision": 0.9544454763184906,
-        "recall": 0.9731871417891852,
-        "support": 20065.0
+        "f1-score": 0.9598950783461034,
+        "precision": 0.9535978787601719,
+        "recall": 0.9662759994441099,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.9096185236684602,
-        "precision": 0.9333746898263028,
-        "recall": 0.8870416224501828,
-        "support": 8481.0
+        "f1-score": 0.9137011538834482,
+        "precision": 0.9281000689451394,
+        "recall": 0.8997421942136924,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9459468516470627,
-      "eval_loss": 0.512050986289978,
+      "eval_accuracy": 0.9425970149253732,
+      "eval_loss": 0.4678420424461365,
       "eval_macro avg": {
-        "f1-score": 0.9245370684664859,
-        "precision": 0.9262980987502312,
-        "recall": 0.9232937579639593,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6554,
-      "eval_samples_per_second": 17.184,
-      "eval_steps_per_second": 2.148,
+        "f1-score": 0.9209046600852139,
+        "precision": 0.9202853760963846,
+        "recall": 0.9217745830711194,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.7618,
+      "eval_samples_per_second": 17.01,
+      "eval_steps_per_second": 2.31,
       "eval_weighted avg": {
-        "f1-score": 0.9455938884035581,
-        "precision": 0.945706920855016,
-        "recall": 0.9459468516470627,
-        "support": 29841.0
+        "f1-score": 0.9424112477025748,
+        "precision": 0.9424267824148957,
+        "recall": 0.9425970149253732,
+        "support": 33500.0
       },
       "step": 2511
     },
     {
       "epoch": 32.0,
       "eval_B": {
-        "f1-score": 0.8938120702826584,
-        "precision": 0.8843537414965986,
-        "recall": 0.9034749034749034,
-        "support": 1295.0
+        "f1-score": 0.8871355060034306,
+        "precision": 0.8766101694915254,
+        "recall": 0.8979166666666667,
+        "support": 1440.0
       },
       "eval_I": {
-        "f1-score": 0.9597342259079783,
-        "precision": 0.962041163803896,
-        "recall": 0.9574383254423124,
-        "support": 20065.0
+        "f1-score": 0.9603147215128719,
+        "precision": 0.9538848263254114,
+        "recall": 0.966831889563163,
+        "support": 21587.0
       },
       "eval_O": {
-        "f1-score": 0.903112155020552,
-        "precision": 0.8995204117440636,
-        "recall": 0.9067326966159651,
-        "support": 8481.0
+        "f1-score": 0.9144436899796294,
+        "precision": 0.9292262198127156,
+        "recall": 0.9001241287119259,
+        "support": 10473.0
       },
-      "eval_accuracy": 0.9406856338594551,
-      "eval_loss": 0.5488592982292175,
+      "eval_accuracy": 0.9430149253731344,
+      "eval_loss": 0.4877029359340668,
       "eval_macro avg": {
-        "f1-score": 0.9188861504037296,
-        "precision": 0.9153051056815195,
-        "recall": 0.9225486418443937,
-        "support": 29841.0
-      },
-      "eval_runtime": 4.6459,
-      "eval_samples_per_second": 17.22,
-      "eval_steps_per_second": 2.152,
+        "f1-score": 0.9206313058319773,
+        "precision": 0.9199070718765507,
+        "recall": 0.9216242283139185,
+        "support": 33500.0
+      },
+      "eval_runtime": 4.759,
+      "eval_samples_per_second": 17.02,
+      "eval_steps_per_second": 2.311,
       "eval_weighted avg": {
-        "f1-score": 0.9407810750507332,
-        "precision": 0.9409009972509189,
-        "recall": 0.9406856338594551,
-        "support": 29841.0
+        "f1-score": 0.9428285906597003,
+        "precision": 0.942854226568747,
+        "recall": 0.9430149253731344,
+        "support": 33500.0
       },
       "step": 2592
     }
@@ -1297,7 +1297,7 @@
   "num_input_tokens_seen": 0,
   "num_train_epochs": 50,
   "save_steps": 500,
-  "total_flos": 4601172887539200.0,
+  "total_flos": 4586883530745600.0,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null

checkpoint-2592/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d229d7bc9127d222c880edfd3fa9418a9ccede656fe27f7f0aa04990c9605be4
+oid sha256:b8ca363922308e3892528dd031742ffedc372734b464a042d09444af9700901e
 size 4856