Training in progress, epoch 5, checkpoint

checkpoint-405/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7f5ec19868d81b157fea4405ccbe022d698701c6b09c24d4ba75cd114dd36bfd
+oid sha256:ca3f1cff9f00f1508294bfc97a6736d552afe16a1e8bede63ac837169d925260
 size 592330980

checkpoint-405/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d92cbf40b8c9d0a7af713da06a48da6cbfb47e6ed358be4c0c78e168ebd4174d
+oid sha256:23f1cd7d75dce7dceaae504dd47d5d10d8f19d3597ec1d9fae06a358b117b6bf
 size 1014670074

checkpoint-405/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c73756f22c58a684dfffbbc966a96d1e3e061222bcc1371ddea16fc12be01725
+oid sha256:112155cc3e2788408d703d89eaa1d2c9a92b4cf9f31734e313b8f50755aa4e0f
 size 1064

checkpoint-405/trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": null,
-  "best_model_checkpoint": null,
+  "best_metric": 0.508861243724823,
+  "best_model_checkpoint": "longformer-full_labels/checkpoint-324",
   "epoch": 5.0,
   "eval_steps": 500,
   "global_step": 405,
@@ -11,323 +11,323 @@
     {
       "epoch": 1.0,
       "eval_B-Claim": {
-        "f1-score": 0.0,
-        "precision": 0.0,
-        "recall": 0.0,
-        "support": 271.0
+        "f1-score": 0.006993006993006993,
+        "precision": 0.5,
+        "recall": 0.0035211267605633804,
+        "support": 284.0
       },
       "eval_B-MajorClaim": {
         "f1-score": 0.0,
         "precision": 0.0,
         "recall": 0.0,
-        "support": 139.0
+        "support": 141.0
       },
       "eval_B-Premise": {
-        "f1-score": 0.6746310611384398,
-        "precision": 0.6075949367088608,
-        "recall": 0.7582938388625592,
-        "support": 633.0
+        "f1-score": 0.6336375488917861,
+        "precision": 0.5883777239709443,
+        "recall": 0.6864406779661016,
+        "support": 708.0
       },
       "eval_I-Claim": {
-        "f1-score": 0.4376506024096386,
-        "precision": 0.550587343690792,
-        "recall": 0.3631592101974506,
-        "support": 4001.0
+        "f1-score": 0.37228850325379614,
+        "precision": 0.41618672324946954,
+        "recall": 0.3367672308069659,
+        "support": 4077.0
       },
       "eval_I-MajorClaim": {
-        "f1-score": 0.620034542314335,
-        "precision": 0.5483008781977854,
-        "recall": 0.7133631395926477,
-        "support": 2013.0
+        "f1-score": 0.49153013051930017,
+        "precision": 0.5611921369689283,
+        "recall": 0.43725296442687744,
+        "support": 2024.0
       },
       "eval_I-Premise": {
-        "f1-score": 0.8620734063103671,
-        "precision": 0.8396186972154862,
-        "recall": 0.8857621736062103,
-        "support": 11336.0
+        "f1-score": 0.846019669697532,
+        "precision": 0.7746890504995582,
+        "recall": 0.9318181818181818,
+        "support": 12232.0
       },
       "eval_O": {
-        "f1-score": 0.8807729058286443,
-        "precision": 0.8630878069080317,
-        "recall": 0.8991979189247779,
-        "support": 9226.0
+        "f1-score": 0.8650519031141869,
+        "precision": 0.9111808904340025,
+        "recall": 0.8233684637211187,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.785908251565951,
-      "eval_loss": 0.5655931830406189,
+      "eval_accuracy": 0.7591191109292971,
+      "eval_loss": 0.6442636251449585,
       "eval_macro avg": {
-        "f1-score": 0.49645178828591785,
-        "precision": 0.48702709467442235,
-        "recall": 0.5171108973119495,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.8252,
-      "eval_samples_per_second": 16.58,
-      "eval_steps_per_second": 2.072,
+        "f1-score": 0.4593601089242298,
+        "precision": 0.535946646446129,
+        "recall": 0.4598812350714013,
+        "support": 29334.0
+      },
+      "eval_runtime": 6.1153,
+      "eval_samples_per_second": 13.082,
+      "eval_steps_per_second": 1.635,
       "eval_weighted avg": {
-        "f1-score": 0.7721020318885458,
-        "precision": 0.766573839857488,
-        "recall": 0.785908251565951,
-        "support": 27619.0
+        "f1-score": 0.7448054609057474,
+        "precision": 0.7451676238152983,
+        "recall": 0.7591191109292971,
+        "support": 29334.0
       },
       "step": 81
     },
     {
       "epoch": 2.0,
       "eval_B-Claim": {
-        "f1-score": 0.36036036036036034,
-        "precision": 0.4624277456647399,
-        "recall": 0.2952029520295203,
-        "support": 271.0
+        "f1-score": 0.47385620915032683,
+        "precision": 0.4420731707317073,
+        "recall": 0.5105633802816901,
+        "support": 284.0
       },
       "eval_B-MajorClaim": {
-        "f1-score": 0.5836909871244635,
-        "precision": 0.723404255319149,
-        "recall": 0.4892086330935252,
-        "support": 139.0
+        "f1-score": 0.5267489711934156,
+        "precision": 0.6274509803921569,
+        "recall": 0.45390070921985815,
+        "support": 141.0
       },
       "eval_B-Premise": {
-        "f1-score": 0.7396491228070177,
-        "precision": 0.6654040404040404,
-        "recall": 0.8325434439178515,
-        "support": 633.0
+        "f1-score": 0.7686170212765958,
+        "precision": 0.7261306532663316,
+        "recall": 0.8163841807909604,
+        "support": 708.0
       },
       "eval_I-Claim": {
-        "f1-score": 0.46704216775764956,
-        "precision": 0.5973520249221184,
-        "recall": 0.3834041489627593,
-        "support": 4001.0
+        "f1-score": 0.5068787075602675,
+        "precision": 0.5221008840353614,
+        "recall": 0.49251900907530044,
+        "support": 4077.0
       },
       "eval_I-MajorClaim": {
-        "f1-score": 0.7326456649010316,
-        "precision": 0.8348157560355781,
-        "recall": 0.6527570789865872,
-        "support": 2013.0
+        "f1-score": 0.6644462947543713,
+        "precision": 0.5741007194244604,
+        "recall": 0.7885375494071146,
+        "support": 2024.0
       },
       "eval_I-Premise": {
-        "f1-score": 0.8679505932065659,
-        "precision": 0.8045344983428744,
-        "recall": 0.9422194777699365,
-        "support": 11336.0
+        "f1-score": 0.8735948241002829,
+        "precision": 0.8642982877260361,
+        "recall": 0.8830935251798561,
+        "support": 12232.0
       },
       "eval_O": {
-        "f1-score": 0.9101698606271778,
-        "precision": 0.9143513454386348,
-        "recall": 0.9060264469976155,
-        "support": 9226.0
+        "f1-score": 0.8825588796944621,
+        "precision": 0.925979519145147,
+        "recall": 0.8430279691933522,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.8169376154096818,
-      "eval_loss": 0.4831336438655853,
+      "eval_accuracy": 0.8015272380173177,
+      "eval_loss": 0.5138605833053589,
       "eval_macro avg": {
-        "f1-score": 0.6659298223977524,
-        "precision": 0.7146128094467336,
-        "recall": 0.6430517402511137,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.8125,
-      "eval_samples_per_second": 16.623,
-      "eval_steps_per_second": 2.078,
+        "f1-score": 0.6709572725328175,
+        "precision": 0.668876316388743,
+        "recall": 0.6840037604497332,
+        "support": 29334.0
+      },
+      "eval_runtime": 6.1482,
+      "eval_samples_per_second": 13.012,
+      "eval_steps_per_second": 1.627,
       "eval_weighted avg": {
-        "f1-score": 0.8047632099274016,
-        "precision": 0.8064582361050344,
-        "recall": 0.8169376154096818,
-        "support": 27619.0
+        "f1-score": 0.8031401896750007,
+        "precision": 0.8089032379475054,
+        "recall": 0.8015272380173177,
+        "support": 29334.0
       },
       "step": 162
     },
     {
       "epoch": 3.0,
       "eval_B-Claim": {
-        "f1-score": 0.5806451612903225,
-        "precision": 0.59765625,
-        "recall": 0.5645756457564576,
-        "support": 271.0
+        "f1-score": 0.4233576642335767,
+        "precision": 0.4393939393939394,
+        "recall": 0.4084507042253521,
+        "support": 284.0
       },
       "eval_B-MajorClaim": {
-        "f1-score": 0.7138047138047138,
-        "precision": 0.6708860759493671,
-        "recall": 0.762589928057554,
-        "support": 139.0
+        "f1-score": 0.6294820717131473,
+        "precision": 0.7181818181818181,
+        "recall": 0.5602836879432624,
+        "support": 141.0
       },
       "eval_B-Premise": {
-        "f1-score": 0.7739984882842026,
-        "precision": 0.7420289855072464,
-        "recall": 0.8088467614533965,
-        "support": 633.0
+        "f1-score": 0.7608562691131499,
+        "precision": 0.6709816612729234,
+        "recall": 0.8785310734463276,
+        "support": 708.0
       },
       "eval_I-Claim": {
-        "f1-score": 0.6128912532475566,
-        "precision": 0.6068103870651641,
-        "recall": 0.6190952261934516,
-        "support": 4001.0
+        "f1-score": 0.46000546000546,
+        "precision": 0.5186211141889813,
+        "recall": 0.41329408879077756,
+        "support": 4077.0
       },
       "eval_I-MajorClaim": {
-        "f1-score": 0.8072942336126171,
-        "precision": 0.8009779951100244,
-        "recall": 0.8137108792846498,
-        "support": 2013.0
+        "f1-score": 0.6850351161534306,
+        "precision": 0.7556615017878426,
+        "recall": 0.6264822134387352,
+        "support": 2024.0
       },
       "eval_I-Premise": {
-        "f1-score": 0.8797559224694903,
-        "precision": 0.8951789627465303,
-        "recall": 0.8648553281580804,
-        "support": 11336.0
+        "f1-score": 0.8689253296477533,
+        "precision": 0.7999862438957287,
+        "recall": 0.9508665794637018,
+        "support": 12232.0
       },
       "eval_O": {
-        "f1-score": 0.9226235130211126,
-        "precision": 0.912356930902925,
-        "recall": 0.933123780619987,
-        "support": 9226.0
+        "f1-score": 0.8740981828044481,
+        "precision": 0.9404692424419283,
+        "recall": 0.8164775030401297,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.8435859372171332,
-      "eval_loss": 0.4283505082130432,
+      "eval_accuracy": 0.7996863707643008,
+      "eval_loss": 0.5769864320755005,
       "eval_macro avg": {
-        "f1-score": 0.7558590408185736,
-        "precision": 0.7465565124687511,
-        "recall": 0.7666853642176539,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.8554,
-      "eval_samples_per_second": 16.476,
-      "eval_steps_per_second": 2.06,
+        "f1-score": 0.6716800133815666,
+        "precision": 0.6918993601661659,
+        "recall": 0.6649122643354695,
+        "support": 29334.0
+      },
+      "eval_runtime": 6.0693,
+      "eval_samples_per_second": 13.181,
+      "eval_steps_per_second": 1.648,
       "eval_weighted avg": {
-        "f1-score": 0.8439412578936614,
-        "precision": 0.8447189682878564,
-        "recall": 0.8435859372171332,
-        "support": 27619.0
+        "f1-score": 0.7930703491848509,
+        "precision": 0.7980829724295806,
+        "recall": 0.7996863707643008,
+        "support": 29334.0
       },
       "step": 243
     },
     {
       "epoch": 4.0,
       "eval_B-Claim": {
-        "f1-score": 0.6033333333333333,
-        "precision": 0.5501519756838906,
-        "recall": 0.6678966789667896,
-        "support": 271.0
+        "f1-score": 0.5486211901306242,
+        "precision": 0.4666666666666667,
+        "recall": 0.6654929577464789,
+        "support": 284.0
       },
       "eval_B-MajorClaim": {
-        "f1-score": 0.6847457627118644,
-        "precision": 0.6474358974358975,
-        "recall": 0.7266187050359713,
-        "support": 139.0
+        "f1-score": 0.6866666666666668,
+        "precision": 0.6477987421383647,
+        "recall": 0.7304964539007093,
+        "support": 141.0
       },
       "eval_B-Premise": {
-        "f1-score": 0.7818471337579618,
-        "precision": 0.7881219903691814,
-        "recall": 0.7756714060031595,
-        "support": 633.0
+        "f1-score": 0.7603550295857988,
+        "precision": 0.7981366459627329,
+        "recall": 0.7259887005649718,
+        "support": 708.0
       },
       "eval_I-Claim": {
-        "f1-score": 0.6216657593903103,
-        "precision": 0.550733024691358,
-        "recall": 0.7135716070982254,
-        "support": 4001.0
+        "f1-score": 0.5798045602605862,
+        "precision": 0.5201636469900643,
+        "recall": 0.6548933038999264,
+        "support": 4077.0
       },
       "eval_I-MajorClaim": {
-        "f1-score": 0.7652218782249742,
-        "precision": 0.7960279119699409,
-        "recall": 0.7367113760556383,
-        "support": 2013.0
+        "f1-score": 0.7716646989374262,
+        "precision": 0.7390321121664405,
+        "recall": 0.8073122529644269,
+        "support": 2024.0
       },
       "eval_I-Premise": {
-        "f1-score": 0.8739165400768475,
-        "precision": 0.8853883758826725,
-        "recall": 0.8627381792519407,
-        "support": 11336.0
+        "f1-score": 0.8704094554664417,
+        "precision": 0.899807994414383,
+        "recall": 0.842871157619359,
+        "support": 12232.0
       },
       "eval_O": {
-        "f1-score": 0.9062570845613239,
-        "precision": 0.9497505345687812,
-        "recall": 0.8665727292434424,
-        "support": 9226.0
+        "f1-score": 0.8969362234264276,
+        "precision": 0.9231016731016731,
+        "recall": 0.8722132144304824,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.8286324631594192,
-      "eval_loss": 0.4826153814792633,
+      "eval_accuracy": 0.8190836571896093,
+      "eval_loss": 0.508861243724823,
       "eval_macro avg": {
-        "f1-score": 0.7481410702938023,
-        "precision": 0.7382299586573888,
-        "recall": 0.7642543830935953,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.8516,
-      "eval_samples_per_second": 16.49,
-      "eval_steps_per_second": 2.061,
+        "f1-score": 0.7306368320677102,
+        "precision": 0.7135296402057607,
+        "recall": 0.7570382915894793,
+        "support": 29334.0
+      },
+      "eval_runtime": 6.0931,
+      "eval_samples_per_second": 13.13,
+      "eval_steps_per_second": 1.641,
       "eval_weighted avg": {
-        "f1-score": 0.8345383371838407,
-        "precision": 0.8451795530099128,
-        "recall": 0.8286324631594192,
-        "support": 27619.0
+        "f1-score": 0.825475128990697,
+        "precision": 0.835926930625345,
+        "recall": 0.8190836571896093,
+        "support": 29334.0
       },
       "step": 324
     },
     {
       "epoch": 5.0,
       "eval_B-Claim": {
-        "f1-score": 0.5825242718446603,
-        "precision": 0.6147540983606558,
-        "recall": 0.5535055350553506,
-        "support": 271.0
+        "f1-score": 0.5466237942122186,
+        "precision": 0.5029585798816568,
+        "recall": 0.5985915492957746,
+        "support": 284.0
       },
       "eval_B-MajorClaim": {
-        "f1-score": 0.7317073170731707,
-        "precision": 0.7094594594594594,
-        "recall": 0.7553956834532374,
-        "support": 139.0
+        "f1-score": 0.6842105263157895,
+        "precision": 0.728,
+        "recall": 0.6453900709219859,
+        "support": 141.0
       },
       "eval_B-Premise": {
-        "f1-score": 0.7840735068912711,
-        "precision": 0.7607726597325408,
-        "recall": 0.8088467614533965,
-        "support": 633.0
+        "f1-score": 0.775623268698061,
+        "precision": 0.7608695652173914,
+        "recall": 0.7909604519774012,
+        "support": 708.0
       },
       "eval_I-Claim": {
-        "f1-score": 0.5830055074744296,
-        "precision": 0.6132413793103448,
-        "recall": 0.5556110972256936,
-        "support": 4001.0
+        "f1-score": 0.5628513335725392,
+        "precision": 0.5492530345471522,
+        "recall": 0.577140053961246,
+        "support": 4077.0
       },
       "eval_I-MajorClaim": {
-        "f1-score": 0.803728638011393,
-        "precision": 0.8393726338561385,
-        "recall": 0.7709885742672627,
-        "support": 2013.0
+        "f1-score": 0.7028394714647174,
+        "precision": 0.8153946510110893,
+        "recall": 0.6175889328063241,
+        "support": 2024.0
       },
       "eval_I-Premise": {
-        "f1-score": 0.8823829787234042,
-        "precision": 0.8523512002630713,
-        "recall": 0.9146083274523642,
-        "support": 11336.0
+        "f1-score": 0.8796072750684049,
+        "precision": 0.8660855784469097,
+        "recall": 0.8935578809679529,
+        "support": 12232.0
       },
       "eval_O": {
-        "f1-score": 0.9254767941792524,
-        "precision": 0.9415657245401525,
-        "recall": 0.9099284630392369,
-        "support": 9226.0
+        "f1-score": 0.8964530307676581,
+        "precision": 0.9043101670447515,
+        "recall": 0.8887312525334414,
+        "support": 9868.0
       },
-      "eval_accuracy": 0.8438031789709982,
-      "eval_loss": 0.513404905796051,
+      "eval_accuracy": 0.8223904002181769,
+      "eval_loss": 0.5749920606613159,
       "eval_macro avg": {
-        "f1-score": 0.7561284305996544,
-        "precision": 0.7616453079317661,
-        "recall": 0.7526977774209346,
-        "support": 27619.0
-      },
-      "eval_runtime": 4.8184,
-      "eval_samples_per_second": 16.603,
-      "eval_steps_per_second": 2.075,
+        "f1-score": 0.7211726714427698,
+        "precision": 0.7324102251641358,
+        "recall": 0.715994313209161,
+        "support": 29334.0
+      },
+      "eval_runtime": 6.0798,
+      "eval_samples_per_second": 13.158,
+      "eval_steps_per_second": 1.645,
       "eval_weighted avg": {
-        "f1-score": 0.8417228378374912,
-        "precision": 0.8414191958613282,
-        "recall": 0.8438031789709982,
-        "support": 27619.0
+        "f1-score": 0.8223802682715222,
+        "precision": 0.8246928072651426,
+        "recall": 0.8223904002181769,
+        "support": 29334.0
       },
       "step": 405
     }
   ],
   "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": 718954063590000.0,
   "train_batch_size": 4,

checkpoint-405/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:96f940fbb595414fcc9335f58119ba24a5ec23e0d6daa1a329588d596e481421
-size 4856
+oid sha256:5653e8d83a92b07eb1df6cde8969d97342166ccb896f2bdca8c3721b3a61a616
+size 4920