End of training

README.md

@@ -4,6 +4,8 @@ license: mit
 base_model: EleutherAI/gpt-neo-1.3B
 tags:
 - generated_from_trainer
+datasets:
+- Ben10x/MedMentions-MTI881-NER
 metrics:
 - precision
 - recall
@@ -11,7 +13,26 @@ metrics:
 - accuracy
 model-index:
 - name: gpt-medmentions
-  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: 0.44823898474262086
+    - name: Recall
+      type: recall
+      value: 0.546458061712299
+    - name: F1
+      type: f1
+      value: 0.4924993145587718
+    - name: Accuracy
+      type: accuracy
+      value: 0.846457800511509
 ---
 
 <!-- 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. -->
 
 # gpt-medmentions
 
-This model is a fine-tuned version of [EleutherAI/gpt-neo-1.3B](https://huggingface.co/EleutherAI/gpt-neo-1.3B) on an unknown dataset.
+This model is a fine-tuned version of [EleutherAI/gpt-neo-1.3B](https://huggingface.co/EleutherAI/gpt-neo-1.3B) on the Ben10x/MedMentions-MTI881-NER dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.0743
-- Precision: 0.4822
-- Recall: 0.5298
-- F1: 0.5049
-- Accuracy: 0.8531
+- Loss: 0.5086
+- Precision: 0.4482
+- Recall: 0.5465
+- F1: 0.4925
+- Accuracy: 0.8465
 
 ## Model description
 

all_results.json

@@ -1,26 +1,26 @@
 {
     "epoch": 5.0,
-    "eval_accuracy": 0.4323785166240409,
-    "eval_f1": 0.5087636106612745,
-    "eval_loss": 1.1892178058624268,
-    "eval_precision": 0.4233535341503,
-    "eval_recall": 0.6373458865725173,
-    "eval_runtime": 11.4272,
+    "eval_accuracy": 0.846457800511509,
+    "eval_f1": 0.4924993145587718,
+    "eval_loss": 0.5086450576782227,
+    "eval_precision": 0.44823898474262086,
+    "eval_recall": 0.546458061712299,
+    "eval_runtime": 33.7698,
     "eval_samples": 2916,
-    "eval_samples_per_second": 255.18,
-    "eval_steps_per_second": 16.014,
-    "predict_accuracy": 0.43058697235747423,
-    "predict_f1": 0.5080028955199871,
-    "predict_loss": 1.1675766706466675,
-    "predict_precision": 0.4175039661554733,
-    "predict_recall": 0.6485931402752105,
-    "predict_runtime": 11.8661,
-    "predict_samples_per_second": 245.995,
-    "predict_steps_per_second": 15.422,
-    "total_flos": 4.924675837027942e+16,
-    "train_loss": 0.9773399162096319,
-    "train_runtime": 7734.8885,
+    "eval_samples_per_second": 86.349,
+    "eval_steps_per_second": 10.808,
+    "predict_accuracy": 0.8543495763524135,
+    "predict_f1": 0.5065875370919882,
+    "predict_loss": 0.49107736349105835,
+    "predict_precision": 0.4661763635039685,
+    "predict_recall": 0.5546699012302894,
+    "predict_runtime": 32.7017,
+    "predict_samples_per_second": 89.262,
+    "predict_steps_per_second": 11.162,
+    "total_flos": 4.62662533103459e+16,
+    "train_loss": 0.2550947672543478,
+    "train_runtime": 8451.3725,
     "train_samples": 23334,
-    "train_samples_per_second": 15.084,
-    "train_steps_per_second": 1.886
+    "train_samples_per_second": 13.805,
+    "train_steps_per_second": 3.452
 }

eval_results.json

@@ -1,12 +1,12 @@
 {
     "epoch": 5.0,
-    "eval_accuracy": 0.4323785166240409,
-    "eval_f1": 0.5087636106612745,
-    "eval_loss": 1.1892178058624268,
-    "eval_precision": 0.4233535341503,
-    "eval_recall": 0.6373458865725173,
-    "eval_runtime": 11.4272,
+    "eval_accuracy": 0.846457800511509,
+    "eval_f1": 0.4924993145587718,
+    "eval_loss": 0.5086450576782227,
+    "eval_precision": 0.44823898474262086,
+    "eval_recall": 0.546458061712299,
+    "eval_runtime": 33.7698,
     "eval_samples": 2916,
-    "eval_samples_per_second": 255.18,
-    "eval_steps_per_second": 16.014
+    "eval_samples_per_second": 86.349,
+    "eval_steps_per_second": 10.808
 }

predict_results.json

@@ -1,10 +1,10 @@
 {
-    "predict_accuracy": 0.43058697235747423,
-    "predict_f1": 0.5080028955199871,
-    "predict_loss": 1.1675766706466675,
-    "predict_precision": 0.4175039661554733,
-    "predict_recall": 0.6485931402752105,
-    "predict_runtime": 11.8661,
-    "predict_samples_per_second": 245.995,
-    "predict_steps_per_second": 15.422
+    "predict_accuracy": 0.8543495763524135,
+    "predict_f1": 0.5065875370919882,
+    "predict_loss": 0.49107736349105835,
+    "predict_precision": 0.4661763635039685,
+    "predict_recall": 0.5546699012302894,
+    "predict_runtime": 32.7017,
+    "predict_samples_per_second": 89.262,
+    "predict_steps_per_second": 11.162
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 5.0,
-    "total_flos": 4.924675837027942e+16,
-    "train_loss": 0.9773399162096319,
-    "train_runtime": 7734.8885,
+    "total_flos": 4.62662533103459e+16,
+    "train_loss": 0.2550947672543478,
+    "train_runtime": 8451.3725,
     "train_samples": 23334,
-    "train_samples_per_second": 15.084,
-    "train_steps_per_second": 1.886
+    "train_samples_per_second": 13.805,
+    "train_steps_per_second": 3.452
 }

trainer_state.json

@@ -1,289 +1,492 @@
 {
-  "best_global_step": 5834,
-  "best_metric": 1.1892178058624268,
-  "best_model_checkpoint": "./output/gpt-medmentions/checkpoint-5834",
+  "best_global_step": 11668,
+  "best_metric": 0.5086450576782227,
+  "best_model_checkpoint": "./output/gpt-medmentions/checkpoint-11668",
   "epoch": 5.0,
   "eval_steps": 500,
-  "global_step": 14585,
+  "global_step": 29170,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.17140898183064793,
-      "grad_norm": 6.046237468719482,
-      "learning_rate": 4.830305107987659e-05,
-      "loss": 1.6106,
+      "epoch": 0.08570449091532396,
+      "grad_norm": 7.563724040985107,
+      "learning_rate": 4.914295509084677e-05,
+      "loss": 0.8639,
       "step": 500
     },
     {
-      "epoch": 0.34281796366129585,
-      "grad_norm": 3.7044713497161865,
-      "learning_rate": 4.658896126157011e-05,
-      "loss": 1.357,
+      "epoch": 0.17140898183064793,
+      "grad_norm": 4.639763832092285,
+      "learning_rate": 4.828591018169352e-05,
+      "loss": 0.674,
       "step": 1000
     },
     {
-      "epoch": 0.5142269454919438,
-      "grad_norm": 4.646540641784668,
-      "learning_rate": 4.487487144326363e-05,
-      "loss": 1.2963,
+      "epoch": 0.2571134727459719,
+      "grad_norm": 4.49621057510376,
+      "learning_rate": 4.7428865272540284e-05,
+      "loss": 0.6427,
       "step": 1500
     },
     {
-      "epoch": 0.6856359273225917,
-      "grad_norm": 3.9113590717315674,
-      "learning_rate": 4.316078162495715e-05,
-      "loss": 1.2696,
+      "epoch": 0.34281796366129585,
+      "grad_norm": 4.740731239318848,
+      "learning_rate": 4.657182036338704e-05,
+      "loss": 0.6284,
       "step": 2000
     },
     {
-      "epoch": 0.8570449091532396,
-      "grad_norm": 3.3376991748809814,
-      "learning_rate": 4.145011998628728e-05,
-      "loss": 1.2514,
+      "epoch": 0.4285224545766198,
+      "grad_norm": 3.477789878845215,
+      "learning_rate": 4.57147754542338e-05,
+      "loss": 0.5904,
       "step": 2500
     },
+    {
+      "epoch": 0.5142269454919438,
+      "grad_norm": 5.801304340362549,
+      "learning_rate": 4.485773054508056e-05,
+      "loss": 0.592,
+      "step": 3000
+    },
+    {
+      "epoch": 0.5999314364072678,
+      "grad_norm": 3.1393306255340576,
+      "learning_rate": 4.4000685635927325e-05,
+      "loss": 0.5751,
+      "step": 3500
+    },
+    {
+      "epoch": 0.6856359273225917,
+      "grad_norm": 2.137924909591675,
+      "learning_rate": 4.314364072677409e-05,
+      "loss": 0.5647,
+      "step": 4000
+    },
+    {
+      "epoch": 0.7713404182379157,
+      "grad_norm": 4.005846977233887,
+      "learning_rate": 4.228659581762084e-05,
+      "loss": 0.5525,
+      "step": 4500
+    },
+    {
+      "epoch": 0.8570449091532396,
+      "grad_norm": 2.6204190254211426,
+      "learning_rate": 4.142955090846761e-05,
+      "loss": 0.5499,
+      "step": 5000
+    },
+    {
+      "epoch": 0.9427494000685636,
+      "grad_norm": 2.9211952686309814,
+      "learning_rate": 4.0572505999314365e-05,
+      "loss": 0.5389,
+      "step": 5500
+    },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.4190537084398977,
-      "eval_f1": 0.3933049606443347,
-      "eval_loss": 1.219216227531433,
-      "eval_precision": 0.40163091733258566,
-      "eval_recall": 0.3853171934543824,
-      "eval_runtime": 11.561,
-      "eval_samples_per_second": 252.228,
-      "eval_steps_per_second": 15.829,
-      "step": 2917
+      "eval_accuracy": 0.8351150895140665,
+      "eval_f1": 0.4352736750651607,
+      "eval_loss": 0.5593962669372559,
+      "eval_precision": 0.435084672166739,
+      "eval_recall": 0.43546284224250326,
+      "eval_runtime": 34.7107,
+      "eval_samples_per_second": 84.009,
+      "eval_steps_per_second": 10.515,
+      "step": 5834
     },
     {
       "epoch": 1.0284538909838876,
-      "grad_norm": 3.143263339996338,
-      "learning_rate": 3.9736030167980805e-05,
-      "loss": 1.2128,
-      "step": 3000
+      "grad_norm": 2.6330394744873047,
+      "learning_rate": 3.9715461090161124e-05,
+      "loss": 0.4835,
+      "step": 6000
+    },
+    {
+      "epoch": 1.1141583818992116,
+      "grad_norm": 7.3211798667907715,
+      "learning_rate": 3.885841618100788e-05,
+      "loss": 0.3699,
+      "step": 6500
     },
     {
       "epoch": 1.1998628728145355,
-      "grad_norm": 3.8394646644592285,
-      "learning_rate": 3.802194034967432e-05,
-      "loss": 1.1019,
-      "step": 3500
+      "grad_norm": 4.461019515991211,
+      "learning_rate": 3.800137127185465e-05,
+      "loss": 0.3823,
+      "step": 7000
+    },
+    {
+      "epoch": 1.2855673637298595,
+      "grad_norm": 3.2148802280426025,
+      "learning_rate": 3.7144326362701406e-05,
+      "loss": 0.3817,
+      "step": 7500
     },
     {
       "epoch": 1.3712718546451834,
-      "grad_norm": 5.761228561401367,
-      "learning_rate": 3.631127871100446e-05,
-      "loss": 1.0983,
-      "step": 4000
+      "grad_norm": 6.220280170440674,
+      "learning_rate": 3.6287281453548164e-05,
+      "loss": 0.3725,
+      "step": 8000
+    },
+    {
+      "epoch": 1.4569763455605074,
+      "grad_norm": 2.2955079078674316,
+      "learning_rate": 3.543023654439493e-05,
+      "loss": 0.38,
+      "step": 8500
     },
     {
       "epoch": 1.5426808364758313,
-      "grad_norm": 5.090176105499268,
-      "learning_rate": 3.459718889269798e-05,
-      "loss": 1.0921,
-      "step": 4500
+      "grad_norm": 4.852652072906494,
+      "learning_rate": 3.457319163524169e-05,
+      "loss": 0.3705,
+      "step": 9000
+    },
+    {
+      "epoch": 1.6283853273911553,
+      "grad_norm": 3.233278751373291,
+      "learning_rate": 3.3716146726088446e-05,
+      "loss": 0.3745,
+      "step": 9500
     },
     {
       "epoch": 1.7140898183064792,
-      "grad_norm": 2.638284683227539,
-      "learning_rate": 3.2883099074391496e-05,
-      "loss": 1.0912,
-      "step": 5000
+      "grad_norm": 3.185788154602051,
+      "learning_rate": 3.285910181693521e-05,
+      "loss": 0.3738,
+      "step": 10000
+    },
+    {
+      "epoch": 1.7997943092218032,
+      "grad_norm": 2.9877729415893555,
+      "learning_rate": 3.200205690778197e-05,
+      "loss": 0.3733,
+      "step": 10500
     },
     {
       "epoch": 1.8854988001371273,
-      "grad_norm": 3.342958927154541,
-      "learning_rate": 3.116900925608502e-05,
-      "loss": 1.0825,
-      "step": 5500
+      "grad_norm": 2.5933837890625,
+      "learning_rate": 3.114501199862873e-05,
+      "loss": 0.3559,
+      "step": 11000
+    },
+    {
+      "epoch": 1.971203291052451,
+      "grad_norm": 2.8704030513763428,
+      "learning_rate": 3.0287967089475487e-05,
+      "loss": 0.3762,
+      "step": 11500
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.4323785166240409,
-      "eval_f1": 0.5087636106612745,
-      "eval_loss": 1.1892178058624268,
-      "eval_precision": 0.4233535341503,
-      "eval_recall": 0.6373458865725173,
-      "eval_runtime": 12.713,
-      "eval_samples_per_second": 229.371,
-      "eval_steps_per_second": 14.395,
-      "step": 5834
+      "eval_accuracy": 0.846457800511509,
+      "eval_f1": 0.4924993145587718,
+      "eval_loss": 0.5086450576782227,
+      "eval_precision": 0.44823898474262086,
+      "eval_recall": 0.546458061712299,
+      "eval_runtime": 34.716,
+      "eval_samples_per_second": 83.996,
+      "eval_steps_per_second": 10.514,
+      "step": 11668
     },
     {
       "epoch": 2.0569077819677752,
-      "grad_norm": 4.2739949226379395,
-      "learning_rate": 2.945491943777854e-05,
-      "loss": 1.0325,
-      "step": 6000
+      "grad_norm": 2.7803287506103516,
+      "learning_rate": 2.9430922180322252e-05,
+      "loss": 0.2413,
+      "step": 12000
+    },
+    {
+      "epoch": 2.142612272883099,
+      "grad_norm": 2.0437328815460205,
+      "learning_rate": 2.8573877271169007e-05,
+      "loss": 0.1788,
+      "step": 12500
     },
     {
       "epoch": 2.228316763798423,
-      "grad_norm": 4.253600120544434,
-      "learning_rate": 2.7740829619472064e-05,
-      "loss": 0.9113,
-      "step": 6500
+      "grad_norm": 1.26553213596344,
+      "learning_rate": 2.7716832362015772e-05,
+      "loss": 0.1809,
+      "step": 13000
+    },
+    {
+      "epoch": 2.314021254713747,
+      "grad_norm": 2.584317922592163,
+      "learning_rate": 2.6859787452862534e-05,
+      "loss": 0.1759,
+      "step": 13500
     },
     {
       "epoch": 2.399725745629071,
-      "grad_norm": 7.081674575805664,
-      "learning_rate": 2.6026739801165584e-05,
-      "loss": 0.9134,
-      "step": 7000
+      "grad_norm": 15.791213989257812,
+      "learning_rate": 2.600274254370929e-05,
+      "loss": 0.1818,
+      "step": 14000
+    },
+    {
+      "epoch": 2.4854302365443948,
+      "grad_norm": 1.3076903820037842,
+      "learning_rate": 2.5145697634556054e-05,
+      "loss": 0.1749,
+      "step": 14500
     },
     {
       "epoch": 2.571134727459719,
-      "grad_norm": 4.441311836242676,
-      "learning_rate": 2.4312649982859104e-05,
-      "loss": 0.9096,
-      "step": 7500
+      "grad_norm": 1.6942801475524902,
+      "learning_rate": 2.4288652725402813e-05,
+      "loss": 0.1767,
+      "step": 15000
+    },
+    {
+      "epoch": 2.656839218375043,
+      "grad_norm": 3.2078938484191895,
+      "learning_rate": 2.3431607816249575e-05,
+      "loss": 0.1737,
+      "step": 15500
     },
     {
       "epoch": 2.742543709290367,
-      "grad_norm": 3.3464174270629883,
-      "learning_rate": 2.2601988344189238e-05,
-      "loss": 0.9105,
-      "step": 8000
+      "grad_norm": 2.383434295654297,
+      "learning_rate": 2.2574562907096333e-05,
+      "loss": 0.1771,
+      "step": 16000
+    },
+    {
+      "epoch": 2.8282482002056906,
+      "grad_norm": 1.1726810932159424,
+      "learning_rate": 2.171751799794309e-05,
+      "loss": 0.1783,
+      "step": 16500
     },
     {
       "epoch": 2.9139526911210147,
-      "grad_norm": 5.6599249839782715,
-      "learning_rate": 2.0887898525882758e-05,
-      "loss": 0.9093,
-      "step": 8500
+      "grad_norm": 8.384498596191406,
+      "learning_rate": 2.0860473088789853e-05,
+      "loss": 0.1775,
+      "step": 17000
+    },
+    {
+      "epoch": 2.999657182036339,
+      "grad_norm": 2.227762222290039,
+      "learning_rate": 2.0003428179636612e-05,
+      "loss": 0.1736,
+      "step": 17500
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.4308056265984655,
-      "eval_f1": 0.4445563583666506,
-      "eval_loss": 1.241865634918213,
-      "eval_precision": 0.4104807637557652,
-      "eval_recall": 0.48480161398789506,
-      "eval_runtime": 12.7077,
-      "eval_samples_per_second": 229.468,
-      "eval_steps_per_second": 14.401,
-      "step": 8751
+      "eval_accuracy": 0.8486956521739131,
+      "eval_f1": 0.4947288251092316,
+      "eval_loss": 0.6310177445411682,
+      "eval_precision": 0.4590834697217676,
+      "eval_recall": 0.5363754889178618,
+      "eval_runtime": 34.7382,
+      "eval_samples_per_second": 83.942,
+      "eval_steps_per_second": 10.507,
+      "step": 17502
     },
     {
       "epoch": 3.0853616729516626,
-      "grad_norm": 2.9584298133850098,
-      "learning_rate": 1.9173808707576275e-05,
-      "loss": 0.8498,
-      "step": 9000
+      "grad_norm": 0.670129120349884,
+      "learning_rate": 1.9146383270483374e-05,
+      "loss": 0.0719,
+      "step": 18000
+    },
+    {
+      "epoch": 3.171066163866987,
+      "grad_norm": 1.9619492292404175,
+      "learning_rate": 1.8289338361330135e-05,
+      "loss": 0.0725,
+      "step": 18500
     },
     {
       "epoch": 3.2567706547823105,
-      "grad_norm": 3.0139501094818115,
-      "learning_rate": 1.74597188892698e-05,
-      "loss": 0.7977,
-      "step": 9500
+      "grad_norm": 0.24141933023929596,
+      "learning_rate": 1.7432293452176894e-05,
+      "loss": 0.0724,
+      "step": 19000
+    },
+    {
+      "epoch": 3.3424751456976347,
+      "grad_norm": 2.3773982524871826,
+      "learning_rate": 1.6575248543023656e-05,
+      "loss": 0.0777,
+      "step": 19500
     },
     {
       "epoch": 3.4281796366129584,
-      "grad_norm": 3.6295053958892822,
-      "learning_rate": 1.5749057250599933e-05,
-      "loss": 0.796,
-      "step": 10000
+      "grad_norm": 1.7786751985549927,
+      "learning_rate": 1.5718203633870414e-05,
+      "loss": 0.0774,
+      "step": 20000
+    },
+    {
+      "epoch": 3.5138841275282826,
+      "grad_norm": 3.449632167816162,
+      "learning_rate": 1.4861158724717176e-05,
+      "loss": 0.0786,
+      "step": 20500
     },
     {
       "epoch": 3.5995886184436063,
-      "grad_norm": 7.977433681488037,
-      "learning_rate": 1.4034967432293453e-05,
-      "loss": 0.7948,
-      "step": 10500
+      "grad_norm": 4.504425048828125,
+      "learning_rate": 1.4004113815563934e-05,
+      "loss": 0.0763,
+      "step": 21000
+    },
+    {
+      "epoch": 3.6852931093589305,
+      "grad_norm": 2.4624173641204834,
+      "learning_rate": 1.3147068906410698e-05,
+      "loss": 0.0726,
+      "step": 21500
     },
     {
       "epoch": 3.7709976002742542,
-      "grad_norm": 2.8351268768310547,
-      "learning_rate": 1.2320877613986975e-05,
-      "loss": 0.7915,
-      "step": 11000
+      "grad_norm": 2.6595826148986816,
+      "learning_rate": 1.2290023997257456e-05,
+      "loss": 0.0766,
+      "step": 22000
+    },
+    {
+      "epoch": 3.8567020911895784,
+      "grad_norm": 0.6136459708213806,
+      "learning_rate": 1.1432979088104218e-05,
+      "loss": 0.072,
+      "step": 22500
     },
     {
       "epoch": 3.942406582104902,
-      "grad_norm": 4.2413434982299805,
-      "learning_rate": 1.0606787795680493e-05,
-      "loss": 0.7938,
-      "step": 11500
+      "grad_norm": 3.4279706478118896,
+      "learning_rate": 1.0575934178950978e-05,
+      "loss": 0.0777,
+      "step": 23000
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.43006393861892583,
-      "eval_f1": 0.35603718836975184,
-      "eval_loss": 1.365502953529358,
-      "eval_precision": 0.37742623979912115,
-      "eval_recall": 0.3369423895987447,
-      "eval_runtime": 12.6489,
-      "eval_samples_per_second": 230.534,
-      "eval_steps_per_second": 14.468,
-      "step": 11668
+      "eval_accuracy": 0.8520971867007673,
+      "eval_f1": 0.49943597242531856,
+      "eval_loss": 0.9267778992652893,
+      "eval_precision": 0.480852775543041,
+      "eval_recall": 0.5195132551064754,
+      "eval_runtime": 34.7136,
+      "eval_samples_per_second": 84.002,
+      "eval_steps_per_second": 10.515,
+      "step": 23336
+    },
+    {
+      "epoch": 4.028111073020226,
+      "grad_norm": 1.2734886407852173,
+      "learning_rate": 9.718889269797737e-06,
+      "loss": 0.0573,
+      "step": 23500
     },
     {
       "epoch": 4.1138155639355505,
-      "grad_norm": 2.2926316261291504,
-      "learning_rate": 8.896126157010627e-06,
-      "loss": 0.7627,
-      "step": 12000
+      "grad_norm": 3.2875871658325195,
+      "learning_rate": 8.861844360644498e-06,
+      "loss": 0.039,
+      "step": 24000
+    },
+    {
+      "epoch": 4.199520054850874,
+      "grad_norm": 13.373528480529785,
+      "learning_rate": 8.004799451491259e-06,
+      "loss": 0.0401,
+      "step": 24500
     },
     {
       "epoch": 4.285224545766198,
-      "grad_norm": 2.511692762374878,
-      "learning_rate": 7.182036338704148e-06,
-      "loss": 0.7492,
-      "step": 12500
+      "grad_norm": 0.31780633330345154,
+      "learning_rate": 7.147754542338019e-06,
+      "loss": 0.0391,
+      "step": 25000
+    },
+    {
+      "epoch": 4.370929036681522,
+      "grad_norm": 1.2009053230285645,
+      "learning_rate": 6.29070963318478e-06,
+      "loss": 0.0416,
+      "step": 25500
     },
     {
       "epoch": 4.456633527596846,
-      "grad_norm": 2.3968722820281982,
-      "learning_rate": 5.467946520397669e-06,
-      "loss": 0.7507,
-      "step": 13000
+      "grad_norm": 1.145856261253357,
+      "learning_rate": 5.43366472403154e-06,
+      "loss": 0.0386,
+      "step": 26000
+    },
+    {
+      "epoch": 4.5423380185121704,
+      "grad_norm": 1.447941780090332,
+      "learning_rate": 4.5766198148783e-06,
+      "loss": 0.0387,
+      "step": 26500
     },
     {
       "epoch": 4.628042509427494,
-      "grad_norm": 3.8909912109375,
-      "learning_rate": 3.75385670209119e-06,
-      "loss": 0.7523,
-      "step": 13500
+      "grad_norm": 2.3324902057647705,
+      "learning_rate": 3.71957490572506e-06,
+      "loss": 0.0434,
+      "step": 27000
+    },
+    {
+      "epoch": 4.713747000342818,
+      "grad_norm": 1.5352897644042969,
+      "learning_rate": 2.8625299965718206e-06,
+      "loss": 0.037,
+      "step": 27500
     },
     {
       "epoch": 4.799451491258142,
-      "grad_norm": 3.9493815898895264,
-      "learning_rate": 2.0431950634213233e-06,
-      "loss": 0.7447,
-      "step": 14000
+      "grad_norm": 0.3736858069896698,
+      "learning_rate": 2.0054850874185807e-06,
+      "loss": 0.0347,
+      "step": 28000
+    },
+    {
+      "epoch": 4.885155982173466,
+      "grad_norm": 0.8886938095092773,
+      "learning_rate": 1.148440178265341e-06,
+      "loss": 0.0379,
+      "step": 28500
     },
     {
       "epoch": 4.9708604730887895,
-      "grad_norm": 4.034415245056152,
-      "learning_rate": 3.2910524511484403e-07,
-      "loss": 0.746,
-      "step": 14500
+      "grad_norm": 1.5339548587799072,
+      "learning_rate": 2.913952691121015e-07,
+      "loss": 0.0394,
+      "step": 29000
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.43336317135549873,
-      "eval_f1": 0.45325477104993267,
-      "eval_loss": 1.451188325881958,
-      "eval_precision": 0.4109656480198381,
-      "eval_recall": 0.5052454606590451,
-      "eval_runtime": 11.736,
-      "eval_samples_per_second": 248.467,
-      "eval_steps_per_second": 15.593,
-      "step": 14585
+      "eval_accuracy": 0.8531202046035805,
+      "eval_f1": 0.5048873426110007,
+      "eval_loss": 1.0743063688278198,
+      "eval_precision": 0.4822375187910436,
+      "eval_recall": 0.5297696653628857,
+      "eval_runtime": 34.8878,
+      "eval_samples_per_second": 83.582,
+      "eval_steps_per_second": 10.462,
+      "step": 29170
     },
     {
       "epoch": 5.0,
-      "step": 14585,
-      "total_flos": 4.924675837027942e+16,
-      "train_loss": 0.9773399162096319,
-      "train_runtime": 7734.8885,
-      "train_samples_per_second": 15.084,
-      "train_steps_per_second": 1.886
+      "step": 29170,
+      "total_flos": 4.62662533103459e+16,
+      "train_loss": 0.2550947672543478,
+      "train_runtime": 8451.3725,
+      "train_samples_per_second": 13.805,
+      "train_steps_per_second": 3.452
     }
   ],
   "logging_steps": 500,
-  "max_steps": 14585,
+  "max_steps": 29170,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 5,
   "save_steps": 500,
@@ -299,7 +502,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 4.924675837027942e+16,
+  "total_flos": 4.62662533103459e+16,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null