Training in progress, epoch 4, checkpoint

checkpoint-324/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5beb1768b76366bc3ac9167ec0bb4f0c2c3246abcde6e385522ad070590d944a
+oid sha256:f2d447c15e7564e31f54f0a9952cc462efe79ec7bd59f194ca4e0b4b6d6cdef1
 size 592330980

checkpoint-324/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:779e2bb085629b66e7229afbe35a1b75bc424b464be5038d01174ea97022e85f
+oid sha256:8f5b41954d25a3bbe0731f805d0a90a4dfc089637cb3ce53f78e1cdcf2300f28
 size 1014670074

checkpoint-324/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f95cf6428a521f225a19b2f71296e812e7a5d112c98910331720e5ff6f1eff62
+oid sha256:bd6dfe6953b9cbab86b93ee2d3f16f1f3460960873f680d5a6b7263a2c839a40
 size 1064

checkpoint-324/trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": null,
-  "best_model_checkpoint": null,
+  "best_metric": 0.276384562253952,
+  "best_model_checkpoint": "longformer-sep_tok_full_labels/checkpoint-324",
   "epoch": 4.0,
   "eval_steps": 500,
   "global_step": 324,
@@ -11,260 +11,260 @@
     {
       "epoch": 1.0,
       "eval_B-Claim": {
-        "f1-score": 0.37160751565762,
-        "precision": 0.42788461538461536,
-        "recall": 0.3284132841328413,
-        "support": 271.0
+        "f1-score": 0.21800947867298578,
+        "precision": 0.3333333333333333,
+        "recall": 0.1619718309859155,
+        "support": 284.0
       },
       "eval_B-MajorClaim": {
-        "f1-score": 0.3578947368421052,
-        "precision": 0.6666666666666666,
-        "recall": 0.2446043165467626,
-        "support": 139.0
+        "f1-score": 0.0,
+        "precision": 0.0,
+        "recall": 0.0,
+        "support": 141.0
       },
       "eval_B-Premise": {
-        "f1-score": 0.8640915593705293,
-        "precision": 0.7895424836601307,
-        "recall": 0.9541864139020537,
-        "support": 633.0
+        "f1-score": 0.8305489260143198,
+        "precision": 0.71900826446281,
+        "recall": 0.9830508474576272,
+        "support": 708.0
       },
       "eval_I-Claim": {
-        "f1-score": 0.5003402749421533,
-        "precision": 0.5493126120741183,
-        "recall": 0.4593851537115721,
-        "support": 4001.0
+        "f1-score": 0.5590492696211935,
+        "precision": 0.5643589102724319,
+        "recall": 0.5538386068187393,
+        "support": 4077.0
       },
       "eval_I-MajorClaim": {
-        "f1-score": 0.7718093699515347,
-        "precision": 0.6502211636611093,
-        "recall": 0.9493293591654247,
-        "support": 2013.0
+        "f1-score": 0.7399038461538461,
+        "precision": 0.7205056179775281,
+        "recall": 0.7603754940711462,
+        "support": 2024.0
       },
       "eval_I-Premise": {
-        "f1-score": 0.875016720916752,
-        "precision": 0.8846812731043188,
-        "recall": 0.865561044460127,
-        "support": 11336.0
+        "f1-score": 0.8867229093291834,
+        "precision": 0.8883965205974068,
+        "recall": 0.8850555918901243,
+        "support": 12232.0
       },
       "eval_O": {
-        "f1-score": 0.9992483530087988,
-        "precision": 0.9995577178239717,
-        "recall": 0.998939179632249,
-        "support": 11312.0
+        "f1-score": 0.9976548035383666,
+        "precision": 0.9961383616794018,
+        "recall": 0.9991758694577221,
+        "support": 12134.0
       },
-      "eval_accuracy": 0.8614038040733883,
-      "eval_loss": 0.31713685393333435,
+      "eval_accuracy": 0.8699050632911393,
+      "eval_loss": 0.32671716809272766,
       "eval_macro avg": {
-        "f1-score": 0.6771440758127848,
-        "precision": 0.7096952189107044,
-        "recall": 0.685774107364433,
-        "support": 29705.0
-      },
-      "eval_runtime": 4.8338,
-      "eval_samples_per_second": 16.55,
-      "eval_steps_per_second": 2.069,
+        "f1-score": 0.6045556047614136,
+        "precision": 0.6031058583318446,
+        "recall": 0.6204954629544678,
+        "support": 31600.0
+      },
+      "eval_runtime": 6.234,
+      "eval_samples_per_second": 12.833,
+      "eval_steps_per_second": 1.604,
       "eval_weighted avg": {
-        "f1-score": 0.8576207231627551,
-        "precision": 0.8601529227027923,
-        "recall": 0.8614038040733883,
-        "support": 29705.0
+        "f1-score": 0.8664142595402331,
+        "precision": 0.8644597558999653,
+        "recall": 0.8699050632911393,
+        "support": 31600.0
       },
       "step": 81
     },
     {
       "epoch": 2.0,
       "eval_B-Claim": {
-        "f1-score": 0.4708624708624709,
-        "precision": 0.6392405063291139,
-        "recall": 0.3726937269372694,
-        "support": 271.0
+        "f1-score": 0.5352697095435685,
+        "precision": 0.6515151515151515,
+        "recall": 0.45422535211267606,
+        "support": 284.0
       },
       "eval_B-MajorClaim": {
-        "f1-score": 0.796875,
-        "precision": 0.8717948717948718,
-        "recall": 0.7338129496402878,
-        "support": 139.0
+        "f1-score": 0.7766990291262136,
+        "precision": 0.7142857142857143,
+        "recall": 0.851063829787234,
+        "support": 141.0
       },
       "eval_B-Premise": {
-        "f1-score": 0.8736616702355461,
-        "precision": 0.796875,
-        "recall": 0.966824644549763,
-        "support": 633.0
+        "f1-score": 0.8802721088435375,
+        "precision": 0.8490813648293963,
+        "recall": 0.9138418079096046,
+        "support": 708.0
       },
       "eval_I-Claim": {
-        "f1-score": 0.5100589925881107,
-        "precision": 0.6459770114942529,
-        "recall": 0.4213946513371657,
-        "support": 4001.0
+        "f1-score": 0.5747557221255521,
+        "precision": 0.6325869180907484,
+        "recall": 0.5266127054206524,
+        "support": 4077.0
       },
       "eval_I-MajorClaim": {
-        "f1-score": 0.8401387776888176,
-        "precision": 0.9077277970011534,
-        "recall": 0.7819175360158966,
-        "support": 2013.0
+        "f1-score": 0.7833260963960051,
+        "precision": 0.6986831913245546,
+        "recall": 0.8913043478260869,
+        "support": 2024.0
       },
       "eval_I-Premise": {
-        "f1-score": 0.8912891699864469,
-        "precision": 0.8338584492430646,
-        "recall": 0.9572159491884262,
-        "support": 11336.0
+        "f1-score": 0.8995515695067264,
+        "precision": 0.8971377459749553,
+        "recall": 0.9019784172661871,
+        "support": 12232.0
       },
       "eval_O": {
-        "f1-score": 0.9996904982977407,
-        "precision": 1.0,
-        "recall": 0.9993811881188119,
-        "support": 11312.0
+        "f1-score": 0.9971231300345225,
+        "precision": 0.9945072962780783,
+        "recall": 0.9997527608373167,
+        "support": 12134.0
       },
-      "eval_accuracy": 0.8830499915839084,
-      "eval_loss": 0.2966194748878479,
+      "eval_accuracy": 0.8864240506329114,
+      "eval_loss": 0.28716403245925903,
       "eval_macro avg": {
-        "f1-score": 0.7689395113798762,
-        "precision": 0.8136390908374939,
-        "recall": 0.7476058065410885,
-        "support": 29705.0
-      },
-      "eval_runtime": 4.8625,
-      "eval_samples_per_second": 16.452,
-      "eval_steps_per_second": 2.057,
+        "f1-score": 0.7781424807965893,
+        "precision": 0.7768281974712284,
+        "recall": 0.791254174451394,
+        "support": 31600.0
+      },
+      "eval_runtime": 6.2816,
+      "eval_samples_per_second": 12.736,
+      "eval_steps_per_second": 1.592,
       "eval_weighted avg": {
-        "f1-score": 0.8731020208182413,
-        "precision": 0.874440834821272,
-        "recall": 0.8830499915839084,
-        "support": 29705.0
+        "f1-score": 0.8834146129726252,
+        "precision": 0.8835831101628018,
+        "recall": 0.8864240506329114,
+        "support": 31600.0
       },
       "step": 162
     },
     {
       "epoch": 3.0,
       "eval_B-Claim": {
-        "f1-score": 0.6085192697768763,
-        "precision": 0.6756756756756757,
-        "recall": 0.5535055350553506,
-        "support": 271.0
+        "f1-score": 0.5580448065173116,
+        "precision": 0.6618357487922706,
+        "recall": 0.4823943661971831,
+        "support": 284.0
       },
       "eval_B-MajorClaim": {
-        "f1-score": 0.8571428571428571,
-        "precision": 0.851063829787234,
-        "recall": 0.8633093525179856,
-        "support": 139.0
+        "f1-score": 0.8046875,
+        "precision": 0.8956521739130435,
+        "recall": 0.7304964539007093,
+        "support": 141.0
       },
       "eval_B-Premise": {
-        "f1-score": 0.8834729626808834,
-        "precision": 0.8529411764705882,
-        "recall": 0.9162717219589257,
-        "support": 633.0
+        "f1-score": 0.8868421052631579,
+        "precision": 0.8300492610837439,
+        "recall": 0.9519774011299436,
+        "support": 708.0
       },
       "eval_I-Claim": {
-        "f1-score": 0.5764474423833614,
-        "precision": 0.6584269662921348,
-        "recall": 0.5126218445388653,
-        "support": 4001.0
+        "f1-score": 0.5488426262920367,
+        "precision": 0.6751432664756447,
+        "recall": 0.4623497669855286,
+        "support": 4077.0
       },
       "eval_I-MajorClaim": {
-        "f1-score": 0.8581151832460733,
-        "precision": 0.9070282235749861,
-        "recall": 0.8142076502732241,
-        "support": 2013.0
+        "f1-score": 0.8110936682365254,
+        "precision": 0.8620689655172413,
+        "recall": 0.7658102766798419,
+        "support": 2024.0
       },
       "eval_I-Premise": {
-        "f1-score": 0.8959744247675935,
-        "precision": 0.8563158317922328,
-        "recall": 0.939484827099506,
-        "support": 11336.0
+        "f1-score": 0.9039491427470623,
+        "precision": 0.854842235662756,
+        "recall": 0.9590418574231524,
+        "support": 12232.0
       },
       "eval_O": {
-        "f1-score": 0.9996020340481981,
-        "precision": 1.0,
-        "recall": 0.9992043847241867,
-        "support": 11312.0
+        "f1-score": 0.9979824597521307,
+        "precision": 0.9972023368715544,
+        "recall": 0.9987638041865832,
+        "support": 12134.0
       },
-      "eval_accuracy": 0.8918700555462044,
-      "eval_loss": 0.2552729547023773,
+      "eval_accuracy": 0.892373417721519,
+      "eval_loss": 0.28271517157554626,
       "eval_macro avg": {
-        "f1-score": 0.811324882006549,
-        "precision": 0.8287788147989789,
-        "recall": 0.7998007594525776,
-        "support": 29705.0
-      },
-      "eval_runtime": 4.8422,
-      "eval_samples_per_second": 16.522,
-      "eval_steps_per_second": 2.065,
+        "f1-score": 0.7873489012583177,
+        "precision": 0.825256284045179,
+        "recall": 0.7644048466432775,
+        "support": 31600.0
+      },
+      "eval_runtime": 6.2071,
+      "eval_samples_per_second": 12.888,
+      "eval_steps_per_second": 1.611,
       "eval_weighted avg": {
-        "f1-score": 0.8867633844066056,
-        "precision": 0.886070631898416,
-        "recall": 0.8918700555462044,
-        "support": 29705.0
+        "f1-score": 0.88435885840807,
+        "precision": 0.8846770016417851,
+        "recall": 0.892373417721519,
+        "support": 31600.0
       },
       "step": 243
     },
     {
       "epoch": 4.0,
       "eval_B-Claim": {
-        "f1-score": 0.6722408026755852,
-        "precision": 0.6146788990825688,
-        "recall": 0.7416974169741697,
-        "support": 271.0
+        "f1-score": 0.659016393442623,
+        "precision": 0.6165644171779141,
+        "recall": 0.7077464788732394,
+        "support": 284.0
       },
       "eval_B-MajorClaim": {
-        "f1-score": 0.8664259927797834,
-        "precision": 0.8695652173913043,
-        "recall": 0.8633093525179856,
-        "support": 139.0
+        "f1-score": 0.8675496688741722,
+        "precision": 0.8136645962732919,
+        "recall": 0.9290780141843972,
+        "support": 141.0
       },
       "eval_B-Premise": {
-        "f1-score": 0.8687035507844755,
-        "precision": 0.9100346020761245,
-        "recall": 0.8309636650868878,
-        "support": 633.0
+        "f1-score": 0.8698224852071005,
+        "precision": 0.9130434782608695,
+        "recall": 0.8305084745762712,
+        "support": 708.0
       },
       "eval_I-Claim": {
-        "f1-score": 0.6483151400094921,
-        "precision": 0.6171222046532641,
-        "recall": 0.6828292926768308,
-        "support": 4001.0
+        "f1-score": 0.6472172351885099,
+        "precision": 0.5964839710444674,
+        "recall": 0.70738287956831,
+        "support": 4077.0
       },
       "eval_I-MajorClaim": {
-        "f1-score": 0.8696993060909791,
-        "precision": 0.9009584664536742,
-        "recall": 0.8405365126676602,
-        "support": 2013.0
+        "f1-score": 0.8549332083430982,
+        "precision": 0.8131966116807846,
+        "recall": 0.9011857707509882,
+        "support": 2024.0
       },
       "eval_I-Premise": {
-        "f1-score": 0.8906159274643798,
-        "precision": 0.9020175517958925,
-        "recall": 0.879498941425547,
-        "support": 11336.0
+        "f1-score": 0.8876102800153433,
+        "precision": 0.9271658801531475,
+        "recall": 0.8512916939175932,
+        "support": 12232.0
       },
       "eval_O": {
-        "f1-score": 0.999557835160948,
-        "precision": 0.9999115357395613,
-        "recall": 0.9992043847241867,
-        "support": 11312.0
+        "f1-score": 0.9981065283609122,
+        "precision": 0.9970394736842105,
+        "recall": 0.9991758694577221,
+        "support": 12134.0
       },
-      "eval_accuracy": 0.8935869382258879,
-      "eval_loss": 0.29465603828430176,
+      "eval_accuracy": 0.8912974683544304,
+      "eval_loss": 0.276384562253952,
       "eval_macro avg": {
-        "f1-score": 0.8307940792808061,
-        "precision": 0.8306126395989127,
-        "recall": 0.8340056522961811,
-        "support": 29705.0
-      },
-      "eval_runtime": 4.8494,
-      "eval_samples_per_second": 16.497,
-      "eval_steps_per_second": 2.062,
+        "f1-score": 0.8263222570616798,
+        "precision": 0.8110226326106693,
+        "recall": 0.8466241687612174,
+        "support": 31600.0
+      },
+      "eval_runtime": 6.2529,
+      "eval_samples_per_second": 12.794,
+      "eval_steps_per_second": 1.599,
       "eval_weighted avg": {
-        "f1-score": 0.8954766464091573,
-        "precision": 0.8982496227307208,
-        "recall": 0.8935869382258879,
-        "support": 29705.0
+        "f1-score": 0.8943886873545748,
+        "precision": 0.9004180663566287,
+        "recall": 0.8912974683544304,
+        "support": 31600.0
       },
       "step": 324
     }
   ],
   "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": 575163250872000.0,
   "train_batch_size": 4,

checkpoint-324/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e3aa7d1dcb1ecb96052506e9849f1d3fa19e688784f8ea37f784dfaca9546785
+oid sha256:4a6eed06a4e3317966121cd80a8275e0b82aa59ebaa0185c3b4676ff92bfed33
 size 4920