End of training

README.md

@@ -4,6 +4,8 @@ license: apache-2.0
 base_model: bert-base-uncased
 tags:
 - generated_from_trainer
+datasets:
+- Ben10x/MedMentions-MTI881-NER
 metrics:
 - precision
 - recall
@@ -11,7 +13,26 @@ metrics:
 - accuracy
 model-index:
 - name: bert-base-mti881
-  results: []
+  results:
+  - task:
+      name: Token Classification
+      type: token-classification
+    dataset:
+      name: Ben10x/MedMentions-MTI881-NER
+      type: Ben10x/MedMentions-MTI881-NER
+    metrics:
+    - name: Precision
+      type: precision
+      value: 0.6301679867699539
+    - name: Recall
+      type: recall
+      value: 0.6139235139489527
+    - name: F1
+      type: f1
+      value: 0.6219396959024139
+    - name: Accuracy
+      type: accuracy
+      value: 0.8788941765196487
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -19,13 +40,13 @@ should probably proofread and complete it, then remove this comment. -->
 
 # bert-base-mti881
 
-This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on an unknown dataset.
+This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the Ben10x/MedMentions-MTI881-NER dataset.
 It achieves the following results on the evaluation set:
-- Loss: 2.3650
-- Precision: 0.6400
-- Recall: 0.6740
-- F1: 0.6566
-- Accuracy: 0.8847
+- Loss: 2.2570
+- Precision: 0.6302
+- Recall: 0.6139
+- F1: 0.6219
+- Accuracy: 0.8789
 
 ## Model description
 

all_results.json

@@ -1,26 +1,26 @@
 {
-    "epoch": 5.0,
-    "eval_accuracy": 0.8675612029632537,
-    "eval_f1": 0.5977588090133096,
-    "eval_loss": 0.41889649629592896,
-    "eval_precision": 0.5715611944917222,
-    "eval_recall": 0.626473331637412,
-    "eval_runtime": 5.6089,
+    "epoch": 15.0,
+    "eval_accuracy": 0.8788941765196487,
+    "eval_f1": 0.6219396959024139,
+    "eval_loss": 2.2569968700408936,
+    "eval_precision": 0.6301679867699539,
+    "eval_recall": 0.6139235139489527,
+    "eval_runtime": 5.5332,
     "eval_samples": 2924,
-    "eval_samples_per_second": 521.316,
-    "eval_steps_per_second": 65.254,
-    "predict_accuracy": 0.8741602672353224,
-    "predict_f1": 0.6073565076798706,
-    "predict_loss": 0.4144425690174103,
-    "predict_precision": 0.5793491671807526,
-    "predict_recall": 0.6382093102276588,
-    "predict_runtime": 5.5766,
-    "predict_samples_per_second": 524.689,
-    "predict_steps_per_second": 65.631,
-    "total_flos": 4651093794825864.0,
-    "train_loss": 0.2857570543859759,
-    "train_runtime": 655.3452,
+    "eval_samples_per_second": 528.446,
+    "eval_steps_per_second": 66.146,
+    "predict_accuracy": 0.8810383719353606,
+    "predict_f1": 0.6222756135232538,
+    "predict_loss": 2.256601333618164,
+    "predict_precision": 0.6286407766990292,
+    "predict_recall": 0.6160380564050288,
+    "predict_runtime": 5.3487,
+    "predict_samples_per_second": 547.044,
+    "predict_steps_per_second": 68.427,
+    "total_flos": 1.39563382170006e+16,
+    "train_loss": 2.07735239021323,
+    "train_runtime": 2398.7609,
     "train_samples": 23399,
-    "train_samples_per_second": 178.524,
-    "train_steps_per_second": 22.316
+    "train_samples_per_second": 146.319,
+    "train_steps_per_second": 18.291
 }

eval_results.json

@@ -1,12 +1,12 @@
 {
-    "epoch": 5.0,
-    "eval_accuracy": 0.8675612029632537,
-    "eval_f1": 0.5977588090133096,
-    "eval_loss": 0.41889649629592896,
-    "eval_precision": 0.5715611944917222,
-    "eval_recall": 0.626473331637412,
-    "eval_runtime": 5.6089,
+    "epoch": 15.0,
+    "eval_accuracy": 0.8788941765196487,
+    "eval_f1": 0.6219396959024139,
+    "eval_loss": 2.2569968700408936,
+    "eval_precision": 0.6301679867699539,
+    "eval_recall": 0.6139235139489527,
+    "eval_runtime": 5.5332,
     "eval_samples": 2924,
-    "eval_samples_per_second": 521.316,
-    "eval_steps_per_second": 65.254
+    "eval_samples_per_second": 528.446,
+    "eval_steps_per_second": 66.146
 }

predict_results.json

@@ -1,10 +1,10 @@
 {
-    "predict_accuracy": 0.8741602672353224,
-    "predict_f1": 0.6073565076798706,
-    "predict_loss": 0.4144425690174103,
-    "predict_precision": 0.5793491671807526,
-    "predict_recall": 0.6382093102276588,
-    "predict_runtime": 5.5766,
-    "predict_samples_per_second": 524.689,
-    "predict_steps_per_second": 65.631
+    "predict_accuracy": 0.8810383719353606,
+    "predict_f1": 0.6222756135232538,
+    "predict_loss": 2.256601333618164,
+    "predict_precision": 0.6286407766990292,
+    "predict_recall": 0.6160380564050288,
+    "predict_runtime": 5.3487,
+    "predict_samples_per_second": 547.044,
+    "predict_steps_per_second": 68.427
 }

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 5.0,
-    "total_flos": 4651093794825864.0,
-    "train_loss": 0.2857570543859759,
-    "train_runtime": 655.3452,
+    "epoch": 15.0,
+    "total_flos": 1.39563382170006e+16,
+    "train_loss": 2.07735239021323,
+    "train_runtime": 2398.7609,
     "train_samples": 23399,
-    "train_samples_per_second": 178.524,
-    "train_steps_per_second": 22.316
+    "train_samples_per_second": 146.319,
+    "train_steps_per_second": 18.291
 }

trainer_state.json

@@ -1,291 +1,817 @@
 {
-  "best_global_step": 5850,
-  "best_metric": 0.41889649629592896,
-  "best_model_checkpoint": "./output/bert-base-mti881/checkpoint-5850",
-  "epoch": 5.0,
+  "best_global_step": 8775,
+  "best_metric": 2.2569968700408936,
+  "best_model_checkpoint": "./output/bert-base-mti881/checkpoint-8775",
+  "epoch": 15.0,
   "eval_steps": 500,
-  "global_step": 14625,
+  "global_step": 43875,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.17094017094017094,
-      "grad_norm": 2.540933132171631,
-      "learning_rate": 4.829059829059829e-05,
-      "loss": 0.8433,
+      "grad_norm": 1.6328529119491577,
+      "learning_rate": 4.943019943019943e-05,
+      "loss": 2.469,
       "step": 500
     },
     {
       "epoch": 0.3418803418803419,
-      "grad_norm": 3.13525390625,
-      "learning_rate": 4.6581196581196586e-05,
-      "loss": 0.5806,
+      "grad_norm": 1.7880568504333496,
+      "learning_rate": 4.886039886039887e-05,
+      "loss": 2.3525,
       "step": 1000
     },
     {
       "epoch": 0.5128205128205128,
-      "grad_norm": 3.402191400527954,
-      "learning_rate": 4.4871794871794874e-05,
-      "loss": 0.5383,
+      "grad_norm": 2.440093994140625,
+      "learning_rate": 4.829059829059829e-05,
+      "loss": 2.331,
       "step": 1500
     },
     {
       "epoch": 0.6837606837606838,
-      "grad_norm": 3.7043747901916504,
-      "learning_rate": 4.316239316239317e-05,
-      "loss": 0.5043,
+      "grad_norm": 2.336617946624756,
+      "learning_rate": 4.772079772079772e-05,
+      "loss": 2.3127,
       "step": 2000
     },
     {
       "epoch": 0.8547008547008547,
-      "grad_norm": 3.2847375869750977,
-      "learning_rate": 4.145299145299146e-05,
-      "loss": 0.4842,
+      "grad_norm": 2.22334885597229,
+      "learning_rate": 4.7150997150997157e-05,
+      "loss": 2.2999,
       "step": 2500
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.8590860030417504,
-      "eval_f1": 0.5612084694700216,
-      "eval_loss": 0.44971752166748047,
-      "eval_precision": 0.5616612875828095,
-      "eval_recall": 0.5607563809039261,
-      "eval_runtime": 6.5198,
-      "eval_samples_per_second": 448.482,
-      "eval_steps_per_second": 56.137,
+      "eval_accuracy": 0.8566329784624442,
+      "eval_f1": 0.5589353612167302,
+      "eval_loss": 2.286189317703247,
+      "eval_precision": 0.5451906796742724,
+      "eval_recall": 0.5733909946578479,
+      "eval_runtime": 6.4563,
+      "eval_samples_per_second": 452.889,
+      "eval_steps_per_second": 56.689,
       "step": 2925
     },
     {
       "epoch": 1.0256410256410255,
-      "grad_norm": 2.163097858428955,
-      "learning_rate": 3.974358974358974e-05,
-      "loss": 0.4497,
+      "grad_norm": 1.5501340627670288,
+      "learning_rate": 4.6581196581196586e-05,
+      "loss": 2.2831,
       "step": 3000
     },
     {
       "epoch": 1.1965811965811965,
-      "grad_norm": 3.154379367828369,
-      "learning_rate": 3.8034188034188035e-05,
-      "loss": 0.3652,
+      "grad_norm": 1.7197738885879517,
+      "learning_rate": 4.6011396011396016e-05,
+      "loss": 2.2335,
       "step": 3500
     },
     {
       "epoch": 1.3675213675213675,
-      "grad_norm": 5.113699913024902,
-      "learning_rate": 3.6324786324786323e-05,
-      "loss": 0.3707,
+      "grad_norm": 2.045734405517578,
+      "learning_rate": 4.544159544159544e-05,
+      "loss": 2.2371,
       "step": 4000
     },
     {
       "epoch": 1.5384615384615383,
-      "grad_norm": 2.192852020263672,
-      "learning_rate": 3.461538461538462e-05,
-      "loss": 0.3606,
+      "grad_norm": 1.6353585720062256,
+      "learning_rate": 4.4871794871794874e-05,
+      "loss": 2.2339,
       "step": 4500
     },
     {
       "epoch": 1.7094017094017095,
-      "grad_norm": 3.2562904357910156,
-      "learning_rate": 3.290598290598291e-05,
-      "loss": 0.359,
+      "grad_norm": 2.460322141647339,
+      "learning_rate": 4.4301994301994304e-05,
+      "loss": 2.233,
       "step": 5000
     },
     {
       "epoch": 1.8803418803418803,
-      "grad_norm": 2.7444751262664795,
-      "learning_rate": 3.1196581196581195e-05,
-      "loss": 0.3492,
+      "grad_norm": 1.5123356580734253,
+      "learning_rate": 4.3732193732193733e-05,
+      "loss": 2.2263,
       "step": 5500
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8675612029632537,
-      "eval_f1": 0.5977588090133096,
-      "eval_loss": 0.41889649629592896,
-      "eval_precision": 0.5715611944917222,
-      "eval_recall": 0.626473331637412,
-      "eval_runtime": 5.6281,
-      "eval_samples_per_second": 519.532,
-      "eval_steps_per_second": 65.03,
+      "eval_accuracy": 0.8603247804543002,
+      "eval_f1": 0.5899178255372945,
+      "eval_loss": 2.275588274002075,
+      "eval_precision": 0.5522597825282936,
+      "eval_recall": 0.6330874247434919,
+      "eval_runtime": 5.6077,
+      "eval_samples_per_second": 521.421,
+      "eval_steps_per_second": 65.267,
       "step": 5850
     },
     {
       "epoch": 2.051282051282051,
-      "grad_norm": 1.9148041009902954,
-      "learning_rate": 2.948717948717949e-05,
-      "loss": 0.3159,
+      "grad_norm": 1.0685631036758423,
+      "learning_rate": 4.316239316239317e-05,
+      "loss": 2.2079,
       "step": 6000
     },
     {
       "epoch": 2.2222222222222223,
-      "grad_norm": 2.1983911991119385,
-      "learning_rate": 2.777777777777778e-05,
-      "loss": 0.2434,
+      "grad_norm": 1.3912978172302246,
+      "learning_rate": 4.259259259259259e-05,
+      "loss": 2.1651,
       "step": 6500
     },
     {
       "epoch": 2.393162393162393,
-      "grad_norm": 3.0160255432128906,
-      "learning_rate": 2.606837606837607e-05,
-      "loss": 0.2473,
+      "grad_norm": 1.7378512620925903,
+      "learning_rate": 4.202279202279202e-05,
+      "loss": 2.1688,
       "step": 7000
     },
     {
       "epoch": 2.564102564102564,
-      "grad_norm": 3.162710666656494,
-      "learning_rate": 2.435897435897436e-05,
-      "loss": 0.2418,
+      "grad_norm": 2.20090913772583,
+      "learning_rate": 4.145299145299146e-05,
+      "loss": 2.1664,
       "step": 7500
     },
     {
       "epoch": 2.735042735042735,
-      "grad_norm": 3.2200865745544434,
-      "learning_rate": 2.264957264957265e-05,
-      "loss": 0.2441,
+      "grad_norm": 1.6386638879776,
+      "learning_rate": 4.088319088319089e-05,
+      "loss": 2.1683,
       "step": 8000
     },
     {
       "epoch": 2.905982905982906,
-      "grad_norm": 1.0021476745605469,
-      "learning_rate": 2.0940170940170943e-05,
-      "loss": 0.2321,
+      "grad_norm": 0.7773854732513428,
+      "learning_rate": 4.031339031339032e-05,
+      "loss": 2.1624,
       "step": 8500
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.8755335328459991,
-      "eval_f1": 0.6145136809287443,
-      "eval_loss": 0.4208410680294037,
-      "eval_precision": 0.5952002542911634,
-      "eval_recall": 0.6351225303145934,
-      "eval_runtime": 5.6691,
-      "eval_samples_per_second": 515.781,
-      "eval_steps_per_second": 64.561,
+      "eval_accuracy": 0.8788941765196487,
+      "eval_f1": 0.6219396959024139,
+      "eval_loss": 2.2569968700408936,
+      "eval_precision": 0.6301679867699539,
+      "eval_recall": 0.6139235139489527,
+      "eval_runtime": 5.624,
+      "eval_samples_per_second": 519.914,
+      "eval_steps_per_second": 65.078,
       "step": 8775
     },
     {
       "epoch": 3.076923076923077,
-      "grad_norm": 4.709245204925537,
-      "learning_rate": 1.923076923076923e-05,
-      "loss": 0.2004,
+      "grad_norm": 2.1296703815460205,
+      "learning_rate": 3.974358974358974e-05,
+      "loss": 2.1407,
       "step": 9000
     },
     {
       "epoch": 3.247863247863248,
-      "grad_norm": 4.517016410827637,
-      "learning_rate": 1.752136752136752e-05,
-      "loss": 0.1575,
+      "grad_norm": 3.029876708984375,
+      "learning_rate": 3.9173789173789176e-05,
+      "loss": 2.1139,
       "step": 9500
     },
     {
       "epoch": 3.4188034188034186,
-      "grad_norm": 2.803234100341797,
-      "learning_rate": 1.581196581196581e-05,
-      "loss": 0.1574,
+      "grad_norm": 2.393371820449829,
+      "learning_rate": 3.8603988603988605e-05,
+      "loss": 2.117,
       "step": 10000
     },
     {
       "epoch": 3.58974358974359,
-      "grad_norm": 4.111881256103516,
-      "learning_rate": 1.4102564102564104e-05,
-      "loss": 0.1511,
+      "grad_norm": 3.3726866245269775,
+      "learning_rate": 3.8034188034188035e-05,
+      "loss": 2.1141,
       "step": 10500
     },
     {
       "epoch": 3.7606837606837606,
-      "grad_norm": 2.1885015964508057,
-      "learning_rate": 1.2393162393162394e-05,
-      "loss": 0.1492,
+      "grad_norm": 1.123772382736206,
+      "learning_rate": 3.746438746438747e-05,
+      "loss": 2.1151,
       "step": 11000
     },
     {
       "epoch": 3.931623931623932,
-      "grad_norm": 3.5671730041503906,
-      "learning_rate": 1.0683760683760684e-05,
-      "loss": 0.1547,
+      "grad_norm": 2.8514039516448975,
+      "learning_rate": 3.6894586894586894e-05,
+      "loss": 2.1192,
       "step": 11500
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.879777265368199,
-      "eval_f1": 0.627274973235609,
-      "eval_loss": 0.4558604061603546,
-      "eval_precision": 0.6097014328023693,
-      "eval_recall": 0.6458916306283389,
-      "eval_runtime": 5.6507,
-      "eval_samples_per_second": 517.462,
-      "eval_steps_per_second": 64.771,
+      "eval_accuracy": 0.8776431339842026,
+      "eval_f1": 0.6255963151834184,
+      "eval_loss": 2.269813299179077,
+      "eval_precision": 0.6073624530863212,
+      "eval_recall": 0.6449588739082507,
+      "eval_runtime": 5.588,
+      "eval_samples_per_second": 523.26,
+      "eval_steps_per_second": 65.497,
       "step": 11700
     },
     {
       "epoch": 4.102564102564102,
-      "grad_norm": 1.0235908031463623,
-      "learning_rate": 8.974358974358976e-06,
-      "loss": 0.1246,
+      "grad_norm": 0.6382957100868225,
+      "learning_rate": 3.6324786324786323e-05,
+      "loss": 2.0942,
       "step": 12000
     },
     {
       "epoch": 4.273504273504273,
-      "grad_norm": 2.5770387649536133,
-      "learning_rate": 7.264957264957266e-06,
-      "loss": 0.0989,
+      "grad_norm": 2.4572439193725586,
+      "learning_rate": 3.575498575498576e-05,
+      "loss": 2.079,
       "step": 12500
     },
     {
       "epoch": 4.444444444444445,
-      "grad_norm": 5.949990272521973,
-      "learning_rate": 5.555555555555556e-06,
-      "loss": 0.1046,
+      "grad_norm": 4.030599117279053,
+      "learning_rate": 3.518518518518519e-05,
+      "loss": 2.0824,
       "step": 13000
     },
     {
       "epoch": 4.615384615384615,
-      "grad_norm": 3.1026697158813477,
-      "learning_rate": 3.846153846153847e-06,
-      "loss": 0.1034,
+      "grad_norm": 1.3176660537719727,
+      "learning_rate": 3.461538461538462e-05,
+      "loss": 2.0844,
       "step": 13500
     },
     {
       "epoch": 4.786324786324786,
-      "grad_norm": 4.661080360412598,
-      "learning_rate": 2.136752136752137e-06,
-      "loss": 0.0989,
+      "grad_norm": 2.164088010787964,
+      "learning_rate": 3.404558404558404e-05,
+      "loss": 2.0804,
       "step": 14000
     },
     {
       "epoch": 4.957264957264957,
-      "grad_norm": 7.092751979827881,
-      "learning_rate": 4.273504273504274e-07,
-      "loss": 0.1037,
+      "grad_norm": 6.9171552658081055,
+      "learning_rate": 3.347578347578348e-05,
+      "loss": 2.0896,
       "step": 14500
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.8808811264288868,
-      "eval_f1": 0.6324561763619196,
-      "eval_loss": 0.5077754855155945,
-      "eval_precision": 0.6130033423523794,
-      "eval_recall": 0.6531840922581192,
-      "eval_runtime": 5.6535,
-      "eval_samples_per_second": 517.199,
-      "eval_steps_per_second": 64.738,
+      "eval_accuracy": 0.8791149487317863,
+      "eval_f1": 0.6324646008618922,
+      "eval_loss": 2.2901737689971924,
+      "eval_precision": 0.612869869551384,
+      "eval_recall": 0.6533536843890444,
+      "eval_runtime": 5.621,
+      "eval_samples_per_second": 520.196,
+      "eval_steps_per_second": 65.113,
       "step": 14625
     },
     {
-      "epoch": 5.0,
-      "step": 14625,
-      "total_flos": 4651093794825864.0,
-      "train_loss": 0.2857570543859759,
-      "train_runtime": 655.3452,
-      "train_samples_per_second": 178.524,
-      "train_steps_per_second": 22.316
+      "epoch": 5.128205128205128,
+      "grad_norm": 2.0550243854522705,
+      "learning_rate": 3.290598290598291e-05,
+      "loss": 2.0631,
+      "step": 15000
+    },
+    {
+      "epoch": 5.299145299145299,
+      "grad_norm": 1.2494322061538696,
+      "learning_rate": 3.2336182336182337e-05,
+      "loss": 2.0561,
+      "step": 15500
+    },
+    {
+      "epoch": 5.47008547008547,
+      "grad_norm": 2.4397966861724854,
+      "learning_rate": 3.176638176638177e-05,
+      "loss": 2.058,
+      "step": 16000
+    },
+    {
+      "epoch": 5.641025641025641,
+      "grad_norm": 2.813675880432129,
+      "learning_rate": 3.1196581196581195e-05,
+      "loss": 2.0611,
+      "step": 16500
+    },
+    {
+      "epoch": 5.811965811965812,
+      "grad_norm": 1.493696928024292,
+      "learning_rate": 3.0626780626780625e-05,
+      "loss": 2.0609,
+      "step": 17000
+    },
+    {
+      "epoch": 5.982905982905983,
+      "grad_norm": 2.580273389816284,
+      "learning_rate": 3.005698005698006e-05,
+      "loss": 2.0621,
+      "step": 17500
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.8811141637639209,
+      "eval_f1": 0.6368790156637131,
+      "eval_loss": 2.2965099811553955,
+      "eval_precision": 0.630865224625624,
+      "eval_recall": 0.6430085644026117,
+      "eval_runtime": 5.5652,
+      "eval_samples_per_second": 525.412,
+      "eval_steps_per_second": 65.766,
+      "step": 17550
+    },
+    {
+      "epoch": 6.153846153846154,
+      "grad_norm": 3.136852741241455,
+      "learning_rate": 2.948717948717949e-05,
+      "loss": 2.0441,
+      "step": 18000
+    },
+    {
+      "epoch": 6.3247863247863245,
+      "grad_norm": 1.3632102012634277,
+      "learning_rate": 2.8917378917378917e-05,
+      "loss": 2.0433,
+      "step": 18500
+    },
+    {
+      "epoch": 6.495726495726496,
+      "grad_norm": 4.941199779510498,
+      "learning_rate": 2.8347578347578346e-05,
+      "loss": 2.0427,
+      "step": 19000
+    },
+    {
+      "epoch": 6.666666666666667,
+      "grad_norm": 2.8133013248443604,
+      "learning_rate": 2.777777777777778e-05,
+      "loss": 2.0436,
+      "step": 19500
+    },
+    {
+      "epoch": 6.837606837606837,
+      "grad_norm": 1.1807732582092285,
+      "learning_rate": 2.720797720797721e-05,
+      "loss": 2.0442,
+      "step": 20000
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.8829661973212971,
+      "eval_f1": 0.6478284496091627,
+      "eval_loss": 2.306105852127075,
+      "eval_precision": 0.6388293487221764,
+      "eval_recall": 0.657084711269397,
+      "eval_runtime": 5.5992,
+      "eval_samples_per_second": 522.219,
+      "eval_steps_per_second": 65.367,
+      "step": 20475
+    },
+    {
+      "epoch": 7.0085470085470085,
+      "grad_norm": 1.7212845087051392,
+      "learning_rate": 2.6638176638176638e-05,
+      "loss": 2.0458,
+      "step": 20500
+    },
+    {
+      "epoch": 7.17948717948718,
+      "grad_norm": 2.134288787841797,
+      "learning_rate": 2.606837606837607e-05,
+      "loss": 2.0322,
+      "step": 21000
+    },
+    {
+      "epoch": 7.35042735042735,
+      "grad_norm": 2.6075599193573,
+      "learning_rate": 2.54985754985755e-05,
+      "loss": 2.033,
+      "step": 21500
+    },
+    {
+      "epoch": 7.521367521367521,
+      "grad_norm": 0.940613329410553,
+      "learning_rate": 2.492877492877493e-05,
+      "loss": 2.0315,
+      "step": 22000
+    },
+    {
+      "epoch": 7.6923076923076925,
+      "grad_norm": 5.997873783111572,
+      "learning_rate": 2.435897435897436e-05,
+      "loss": 2.0317,
+      "step": 22500
+    },
+    {
+      "epoch": 7.863247863247864,
+      "grad_norm": 1.9498519897460938,
+      "learning_rate": 2.3789173789173792e-05,
+      "loss": 2.0301,
+      "step": 23000
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.8818132757690232,
+      "eval_f1": 0.6476386036960986,
+      "eval_loss": 2.3260273933410645,
+      "eval_precision": 0.6279366090626742,
+      "eval_recall": 0.6686169761723056,
+      "eval_runtime": 5.5616,
+      "eval_samples_per_second": 525.751,
+      "eval_steps_per_second": 65.809,
+      "step": 23400
+    },
+    {
+      "epoch": 8.034188034188034,
+      "grad_norm": 3.1696274280548096,
+      "learning_rate": 2.321937321937322e-05,
+      "loss": 2.0325,
+      "step": 23500
+    },
+    {
+      "epoch": 8.205128205128204,
+      "grad_norm": 0.9211856126785278,
+      "learning_rate": 2.264957264957265e-05,
+      "loss": 2.0222,
+      "step": 24000
+    },
+    {
+      "epoch": 8.376068376068377,
+      "grad_norm": 2.332916259765625,
+      "learning_rate": 2.207977207977208e-05,
+      "loss": 2.0244,
+      "step": 24500
+    },
+    {
+      "epoch": 8.547008547008547,
+      "grad_norm": 1.2731038331985474,
+      "learning_rate": 2.150997150997151e-05,
+      "loss": 2.0242,
+      "step": 25000
+    },
+    {
+      "epoch": 8.717948717948717,
+      "grad_norm": 0.8299376964569092,
+      "learning_rate": 2.0940170940170943e-05,
+      "loss": 2.0238,
+      "step": 25500
+    },
+    {
+      "epoch": 8.88888888888889,
+      "grad_norm": 1.503308892250061,
+      "learning_rate": 2.037037037037037e-05,
+      "loss": 2.0242,
+      "step": 26000
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.8830275229357798,
+      "eval_f1": 0.6493926454127109,
+      "eval_loss": 2.3398172855377197,
+      "eval_precision": 0.6353017521090201,
+      "eval_recall": 0.6641227847027897,
+      "eval_runtime": 5.6249,
+      "eval_samples_per_second": 519.833,
+      "eval_steps_per_second": 65.068,
+      "step": 26325
+    },
+    {
+      "epoch": 9.05982905982906,
+      "grad_norm": 1.7587120532989502,
+      "learning_rate": 1.9800569800569802e-05,
+      "loss": 2.0226,
+      "step": 26500
+    },
+    {
+      "epoch": 9.23076923076923,
+      "grad_norm": 0.7542155385017395,
+      "learning_rate": 1.923076923076923e-05,
+      "loss": 2.0177,
+      "step": 27000
+    },
+    {
+      "epoch": 9.401709401709402,
+      "grad_norm": 0.33988329768180847,
+      "learning_rate": 1.866096866096866e-05,
+      "loss": 2.0203,
+      "step": 27500
+    },
+    {
+      "epoch": 9.572649572649572,
+      "grad_norm": 1.8626066446304321,
+      "learning_rate": 1.8091168091168094e-05,
+      "loss": 2.0175,
+      "step": 28000
+    },
+    {
+      "epoch": 9.743589743589745,
+      "grad_norm": 2.40765118598938,
+      "learning_rate": 1.752136752136752e-05,
+      "loss": 2.0183,
+      "step": 28500
+    },
+    {
+      "epoch": 9.914529914529915,
+      "grad_norm": 2.155571222305298,
+      "learning_rate": 1.6951566951566953e-05,
+      "loss": 2.0173,
+      "step": 29000
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.8841559142422607,
+      "eval_f1": 0.652157598499062,
+      "eval_loss": 2.3391082286834717,
+      "eval_precision": 0.641486220472441,
+      "eval_recall": 0.6631900279827017,
+      "eval_runtime": 5.5617,
+      "eval_samples_per_second": 525.742,
+      "eval_steps_per_second": 65.808,
+      "step": 29250
+    },
+    {
+      "epoch": 10.085470085470085,
+      "grad_norm": 1.149816870689392,
+      "learning_rate": 1.6381766381766382e-05,
+      "loss": 2.0171,
+      "step": 29500
+    },
+    {
+      "epoch": 10.256410256410255,
+      "grad_norm": 0.5041487812995911,
+      "learning_rate": 1.581196581196581e-05,
+      "loss": 2.0133,
+      "step": 30000
+    },
+    {
+      "epoch": 10.427350427350428,
+      "grad_norm": 6.211667537689209,
+      "learning_rate": 1.5242165242165243e-05,
+      "loss": 2.0144,
+      "step": 30500
+    },
+    {
+      "epoch": 10.598290598290598,
+      "grad_norm": 0.1538165956735611,
+      "learning_rate": 1.4672364672364672e-05,
+      "loss": 2.0135,
+      "step": 31000
+    },
+    {
+      "epoch": 10.76923076923077,
+      "grad_norm": 1.0518053770065308,
+      "learning_rate": 1.4102564102564104e-05,
+      "loss": 2.0128,
+      "step": 31500
+    },
+    {
+      "epoch": 10.94017094017094,
+      "grad_norm": 1.116525650024414,
+      "learning_rate": 1.3532763532763535e-05,
+      "loss": 2.0132,
+      "step": 32000
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": 0.8832973556395035,
+      "eval_f1": 0.6500785318674052,
+      "eval_loss": 2.3498170375823975,
+      "eval_precision": 0.634142407870333,
+      "eval_recall": 0.6668362587975918,
+      "eval_runtime": 5.7697,
+      "eval_samples_per_second": 506.782,
+      "eval_steps_per_second": 63.434,
+      "step": 32175
+    },
+    {
+      "epoch": 11.11111111111111,
+      "grad_norm": 0.1830213963985443,
+      "learning_rate": 1.2962962962962962e-05,
+      "loss": 2.0121,
+      "step": 32500
+    },
+    {
+      "epoch": 11.282051282051283,
+      "grad_norm": 2.5111734867095947,
+      "learning_rate": 1.2393162393162394e-05,
+      "loss": 2.0103,
+      "step": 33000
+    },
+    {
+      "epoch": 11.452991452991453,
+      "grad_norm": 3.7082180976867676,
+      "learning_rate": 1.1823361823361825e-05,
+      "loss": 2.0103,
+      "step": 33500
+    },
+    {
+      "epoch": 11.623931623931623,
+      "grad_norm": 1.1296755075454712,
+      "learning_rate": 1.1253561253561254e-05,
+      "loss": 2.011,
+      "step": 34000
+    },
+    {
+      "epoch": 11.794871794871796,
+      "grad_norm": 2.4463248252868652,
+      "learning_rate": 1.0683760683760684e-05,
+      "loss": 2.0093,
+      "step": 34500
+    },
+    {
+      "epoch": 11.965811965811966,
+      "grad_norm": 0.03058500401675701,
+      "learning_rate": 1.0113960113960115e-05,
+      "loss": 2.0097,
+      "step": 35000
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 0.8845851935436393,
+      "eval_f1": 0.6505743299483937,
+      "eval_loss": 2.355226993560791,
+      "eval_precision": 0.6388230486309767,
+      "eval_recall": 0.6627660476553888,
+      "eval_runtime": 5.5805,
+      "eval_samples_per_second": 523.964,
+      "eval_steps_per_second": 65.585,
+      "step": 35100
+    },
+    {
+      "epoch": 12.136752136752136,
+      "grad_norm": 1.262992024421692,
+      "learning_rate": 9.544159544159544e-06,
+      "loss": 2.0083,
+      "step": 35500
+    },
+    {
+      "epoch": 12.307692307692308,
+      "grad_norm": 0.350888192653656,
+      "learning_rate": 8.974358974358976e-06,
+      "loss": 2.0082,
+      "step": 36000
+    },
+    {
+      "epoch": 12.478632478632479,
+      "grad_norm": 0.7504994869232178,
+      "learning_rate": 8.404558404558405e-06,
+      "loss": 2.0089,
+      "step": 36500
+    },
+    {
+      "epoch": 12.649572649572649,
+      "grad_norm": 2.052617311477661,
+      "learning_rate": 7.834757834757835e-06,
+      "loss": 2.0072,
+      "step": 37000
+    },
+    {
+      "epoch": 12.820512820512821,
+      "grad_norm": 0.4613409638404846,
+      "learning_rate": 7.264957264957266e-06,
+      "loss": 2.0073,
+      "step": 37500
+    },
+    {
+      "epoch": 12.991452991452991,
+      "grad_norm": 4.136294364929199,
+      "learning_rate": 6.695156695156696e-06,
+      "loss": 2.007,
+      "step": 38000
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": 0.8839228769072266,
+      "eval_f1": 0.6545124566903151,
+      "eval_loss": 2.3634226322174072,
+      "eval_precision": 0.6372178941450486,
+      "eval_recall": 0.6727719833799711,
+      "eval_runtime": 5.622,
+      "eval_samples_per_second": 520.097,
+      "eval_steps_per_second": 65.101,
+      "step": 38025
+    },
+    {
+      "epoch": 13.162393162393162,
+      "grad_norm": 0.16694723069667816,
+      "learning_rate": 6.1253561253561255e-06,
+      "loss": 2.0057,
+      "step": 38500
+    },
+    {
+      "epoch": 13.333333333333334,
+      "grad_norm": 0.8811143636703491,
+      "learning_rate": 5.555555555555556e-06,
+      "loss": 2.0065,
+      "step": 39000
+    },
+    {
+      "epoch": 13.504273504273504,
+      "grad_norm": 0.4992905855178833,
+      "learning_rate": 4.985754985754986e-06,
+      "loss": 2.0068,
+      "step": 39500
+    },
+    {
+      "epoch": 13.675213675213675,
+      "grad_norm": 0.6530119180679321,
+      "learning_rate": 4.415954415954416e-06,
+      "loss": 2.0052,
+      "step": 40000
+    },
+    {
+      "epoch": 13.846153846153847,
+      "grad_norm": 2.222022771835327,
+      "learning_rate": 3.846153846153847e-06,
+      "loss": 2.0062,
+      "step": 40500
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": 0.884462542314674,
+      "eval_f1": 0.6561026065370293,
+      "eval_loss": 2.3629047870635986,
+      "eval_precision": 0.6406237375777653,
+      "eval_recall": 0.6723480030526584,
+      "eval_runtime": 5.6036,
+      "eval_samples_per_second": 521.81,
+      "eval_steps_per_second": 65.316,
+      "step": 40950
+    },
+    {
+      "epoch": 14.017094017094017,
+      "grad_norm": 0.11298029124736786,
+      "learning_rate": 3.2763532763532763e-06,
+      "loss": 2.0064,
+      "step": 41000
+    },
+    {
+      "epoch": 14.188034188034187,
+      "grad_norm": 0.11808889359235764,
+      "learning_rate": 2.7065527065527066e-06,
+      "loss": 2.0048,
+      "step": 41500
+    },
+    {
+      "epoch": 14.35897435897436,
+      "grad_norm": 0.051862556487321854,
+      "learning_rate": 2.136752136752137e-06,
+      "loss": 2.0052,
+      "step": 42000
+    },
+    {
+      "epoch": 14.52991452991453,
+      "grad_norm": 0.021300671622157097,
+      "learning_rate": 1.566951566951567e-06,
+      "loss": 2.0053,
+      "step": 42500
+    },
+    {
+      "epoch": 14.7008547008547,
+      "grad_norm": 0.11307813972234726,
+      "learning_rate": 9.971509971509971e-07,
+      "loss": 2.005,
+      "step": 43000
+    },
+    {
+      "epoch": 14.871794871794872,
+      "grad_norm": 1.3423974514007568,
+      "learning_rate": 4.273504273504274e-07,
+      "loss": 2.0041,
+      "step": 43500
+    },
+    {
+      "epoch": 15.0,
+      "eval_accuracy": 0.8847446401412942,
+      "eval_f1": 0.6565610672834661,
+      "eval_loss": 2.365044116973877,
+      "eval_precision": 0.6400386535674022,
+      "eval_recall": 0.673959128296447,
+      "eval_runtime": 5.6195,
+      "eval_samples_per_second": 520.335,
+      "eval_steps_per_second": 65.131,
+      "step": 43875
+    },
+    {
+      "epoch": 15.0,
+      "step": 43875,
+      "total_flos": 1.39563382170006e+16,
+      "train_loss": 2.07735239021323,
+      "train_runtime": 2398.7609,
+      "train_samples_per_second": 146.319,
+      "train_steps_per_second": 18.291
     }
   ],
   "logging_steps": 500,
-  "max_steps": 14625,
+  "max_steps": 43875,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 5,
+  "num_train_epochs": 15,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -299,7 +825,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 4651093794825864.0,
+  "total_flos": 1.39563382170006e+16,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null