End of training

README.md

@@ -1,4 +1,6 @@
 ---
+language:
+- en
 tags:
 - generated_from_trainer
 datasets:
@@ -13,7 +15,7 @@ model-index:
       name: Text Classification
       type: text-classification
     dataset:
-      name: glue
+      name: GLUE QQP
       type: glue
       config: qqp
       split: validation
@@ -21,10 +23,10 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.8134306208261192
+      value: 0.7902547613158546
     - name: F1
       type: f1
-      value: 0.7601361020129106
+      value: 0.7136682874122096
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -32,12 +34,12 @@ should probably proofread and complete it, then remove this comment. -->
 
 # hBERTv2_new_pretrain_qqp
 
-This model is a fine-tuned version of [gokuls/bert_12_layer_model_v2_complete_training_new](https://huggingface.co/gokuls/bert_12_layer_model_v2_complete_training_new) on the glue dataset.
+This model is a fine-tuned version of [gokuls/bert_12_layer_model_v2_complete_training_new](https://huggingface.co/gokuls/bert_12_layer_model_v2_complete_training_new) on the GLUE QQP dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5231
-- Accuracy: 0.8134
-- F1: 0.7601
-- Combined Score: 0.7868
+- Loss: 0.4399
+- Accuracy: 0.7903
+- F1: 0.7137
+- Combined Score: 0.7520
 
 ## Model description
 

all_results.json

@@ -1,16 +1,16 @@
 {
-    "epoch": 11.0,
-    "eval_accuracy": 0.6318327974276527,
-    "eval_combined_score": 0.3159163987138264,
-    "eval_f1": 0.0,
-    "eval_loss": 0.6569251418113708,
-    "eval_runtime": 69.3881,
+    "epoch": 8.0,
+    "eval_accuracy": 0.7902547613158546,
+    "eval_combined_score": 0.7519615243640321,
+    "eval_f1": 0.7136682874122096,
+    "eval_loss": 0.4399268329143524,
+    "eval_runtime": 68.7606,
     "eval_samples": 40430,
-    "eval_samples_per_second": 582.665,
-    "eval_steps_per_second": 4.554,
-    "train_loss": 23.176317952643103,
-    "train_runtime": 20740.4239,
+    "eval_samples_per_second": 587.982,
+    "eval_steps_per_second": 4.596,
+    "train_loss": 0.3639708930574753,
+    "train_runtime": 14995.3156,
     "train_samples": 363846,
-    "train_samples_per_second": 877.142,
-    "train_steps_per_second": 6.854
+    "train_samples_per_second": 1213.199,
+    "train_steps_per_second": 9.48
 }

eval_results.json

@@ -1,11 +1,11 @@
 {
-    "epoch": 11.0,
-    "eval_accuracy": 0.6318327974276527,
-    "eval_combined_score": 0.3159163987138264,
-    "eval_f1": 0.0,
-    "eval_loss": 0.6569251418113708,
-    "eval_runtime": 69.3881,
+    "epoch": 8.0,
+    "eval_accuracy": 0.7902547613158546,
+    "eval_combined_score": 0.7519615243640321,
+    "eval_f1": 0.7136682874122096,
+    "eval_loss": 0.4399268329143524,
+    "eval_runtime": 68.7606,
     "eval_samples": 40430,
-    "eval_samples_per_second": 582.665,
-    "eval_steps_per_second": 4.554
+    "eval_samples_per_second": 587.982,
+    "eval_steps_per_second": 4.596
 }

logs/events.out.tfevents.1686052831.serv-3317.3377083.14

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4000452a817dab9e73ad82fdfae941201a66028c7e2d5e840ead7f1d2b18cde4
+size 475

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 11.0,
-    "train_loss": 23.176317952643103,
-    "train_runtime": 20740.4239,
+    "epoch": 8.0,
+    "train_loss": 0.3639708930574753,
+    "train_runtime": 14995.3156,
     "train_samples": 363846,
-    "train_samples_per_second": 877.142,
-    "train_steps_per_second": 6.854
+    "train_samples_per_second": 1213.199,
+    "train_steps_per_second": 9.48
 }

trainer_state.json

@@ -1,212 +1,161 @@
 {
-  "best_metric": 0.6569251418113708,
-  "best_model_checkpoint": "hBERTv2_new_pretrain_qqp/checkpoint-17058",
-  "epoch": 11.0,
-  "global_step": 31273,
+  "best_metric": 0.4399268329143524,
+  "best_model_checkpoint": "hBERTv2_new_pretrain_qqp/checkpoint-8529",
+  "epoch": 8.0,
+  "global_step": 22744,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 1.0,
-      "learning_rate": 0.00049,
-      "loss": 130.0621,
+      "learning_rate": 3.9200000000000004e-05,
+      "loss": 0.5303,
       "step": 2843
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.36816720257234725,
-      "eval_combined_score": 0.45317878342483403,
-      "eval_f1": 0.5381903642773208,
-      "eval_loss": 125.09819793701172,
-      "eval_runtime": 70.5346,
-      "eval_samples_per_second": 573.194,
-      "eval_steps_per_second": 4.48,
+      "eval_accuracy": 0.7608211723967351,
+      "eval_combined_score": 0.6928911943274398,
+      "eval_f1": 0.6249612162581446,
+      "eval_loss": 0.48931893706321716,
+      "eval_runtime": 69.0476,
+      "eval_samples_per_second": 585.538,
+      "eval_steps_per_second": 4.577,
       "step": 2843
     },
     {
       "epoch": 2.0,
-      "learning_rate": 0.00048,
-      "loss": 63.3796,
+      "learning_rate": 3.8400000000000005e-05,
+      "loss": 0.4677,
       "step": 5686
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.36816720257234725,
-      "eval_combined_score": 0.45317878342483403,
-      "eval_f1": 0.5381903642773208,
-      "eval_loss": 0.8078107833862305,
-      "eval_runtime": 70.4435,
-      "eval_samples_per_second": 573.935,
-      "eval_steps_per_second": 4.486,
+      "eval_accuracy": 0.7773188226564433,
+      "eval_combined_score": 0.7302284114338071,
+      "eval_f1": 0.6831380002111711,
+      "eval_loss": 0.4780506491661072,
+      "eval_runtime": 69.0619,
+      "eval_samples_per_second": 585.417,
+      "eval_steps_per_second": 4.576,
       "step": 5686
     },
     {
       "epoch": 3.0,
-      "learning_rate": 0.00047,
-      "loss": 0.7936,
+      "learning_rate": 3.76e-05,
+      "loss": 0.4229,
       "step": 8529
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.36816720257234725,
-      "eval_combined_score": 0.45317878342483403,
-      "eval_f1": 0.5381903642773208,
-      "eval_loss": 2.612349271774292,
-      "eval_runtime": 70.4376,
-      "eval_samples_per_second": 573.983,
-      "eval_steps_per_second": 4.486,
+      "eval_accuracy": 0.7902547613158546,
+      "eval_combined_score": 0.7519615243640321,
+      "eval_f1": 0.7136682874122096,
+      "eval_loss": 0.4399268329143524,
+      "eval_runtime": 69.0436,
+      "eval_samples_per_second": 585.572,
+      "eval_steps_per_second": 4.577,
       "step": 8529
     },
     {
       "epoch": 4.0,
-      "learning_rate": 0.00046,
-      "loss": 27.5934,
+      "learning_rate": 3.680000000000001e-05,
+      "loss": 0.3712,
       "step": 11372
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.36816720257234725,
-      "eval_combined_score": 0.45317878342483403,
-      "eval_f1": 0.5381903642773208,
-      "eval_loss": 25.9051456451416,
-      "eval_runtime": 69.7747,
-      "eval_samples_per_second": 579.436,
-      "eval_steps_per_second": 4.529,
+      "eval_accuracy": 0.8017808558001484,
+      "eval_combined_score": 0.7590396774044639,
+      "eval_f1": 0.7162984990087793,
+      "eval_loss": 0.4425596296787262,
+      "eval_runtime": 69.1036,
+      "eval_samples_per_second": 585.064,
+      "eval_steps_per_second": 4.573,
       "step": 11372
     },
     {
       "epoch": 5.0,
-      "learning_rate": 0.00045000000000000004,
-      "loss": 20.3073,
+      "learning_rate": 3.6e-05,
+      "loss": 0.3268,
       "step": 14215
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.36816720257234725,
-      "eval_combined_score": 0.45317878342483403,
-      "eval_f1": 0.5381903642773208,
-      "eval_loss": 16.190715789794922,
-      "eval_runtime": 69.466,
-      "eval_samples_per_second": 582.011,
-      "eval_steps_per_second": 4.549,
+      "eval_accuracy": 0.8106851348008904,
+      "eval_combined_score": 0.7727644478133877,
+      "eval_f1": 0.7348437608258851,
+      "eval_loss": 0.4514910578727722,
+      "eval_runtime": 69.0616,
+      "eval_samples_per_second": 585.419,
+      "eval_steps_per_second": 4.576,
       "step": 14215
     },
     {
       "epoch": 6.0,
-      "learning_rate": 0.00044,
-      "loss": 7.6381,
+      "learning_rate": 3.52e-05,
+      "loss": 0.2925,
       "step": 17058
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.6318327974276527,
-      "eval_combined_score": 0.3159163987138264,
-      "eval_f1": 0.0,
-      "eval_loss": 0.6569251418113708,
-      "eval_runtime": 69.8766,
-      "eval_samples_per_second": 578.592,
-      "eval_steps_per_second": 4.522,
+      "eval_accuracy": 0.8118971061093248,
+      "eval_combined_score": 0.7672875750435039,
+      "eval_f1": 0.722678043977683,
+      "eval_loss": 0.5221311450004578,
+      "eval_runtime": 69.0569,
+      "eval_samples_per_second": 585.459,
+      "eval_steps_per_second": 4.576,
       "step": 17058
     },
     {
       "epoch": 7.0,
-      "learning_rate": 0.00043,
-      "loss": 0.6675,
+      "learning_rate": 3.44e-05,
+      "loss": 0.2614,
       "step": 19901
     },
     {
       "epoch": 7.0,
-      "eval_accuracy": 0.6318327974276527,
-      "eval_combined_score": 0.3159163987138264,
-      "eval_f1": 0.0,
-      "eval_loss": 0.6590132713317871,
-      "eval_runtime": 70.184,
-      "eval_samples_per_second": 576.057,
-      "eval_steps_per_second": 4.502,
+      "eval_accuracy": 0.8057630472421469,
+      "eval_combined_score": 0.7792239164658441,
+      "eval_f1": 0.7526847856895412,
+      "eval_loss": 0.45176824927330017,
+      "eval_runtime": 69.1339,
+      "eval_samples_per_second": 584.807,
+      "eval_steps_per_second": 4.571,
       "step": 19901
     },
     {
       "epoch": 8.0,
-      "learning_rate": 0.00042,
-      "loss": 0.6629,
+      "learning_rate": 3.3600000000000004e-05,
+      "loss": 0.2389,
       "step": 22744
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.6318327974276527,
-      "eval_combined_score": 0.3159163987138264,
-      "eval_f1": 0.0,
-      "eval_loss": 0.6588091850280762,
-      "eval_runtime": 69.9082,
-      "eval_samples_per_second": 578.33,
-      "eval_steps_per_second": 4.52,
+      "eval_accuracy": 0.8134306208261192,
+      "eval_combined_score": 0.7867833614195149,
+      "eval_f1": 0.7601361020129106,
+      "eval_loss": 0.5231410264968872,
+      "eval_runtime": 69.0842,
+      "eval_samples_per_second": 585.228,
+      "eval_steps_per_second": 4.574,
       "step": 22744
     },
     {
-      "epoch": 9.0,
-      "learning_rate": 0.00041,
-      "loss": 0.6629,
-      "step": 25587
-    },
-    {
-      "epoch": 9.0,
-      "eval_accuracy": 0.6318327974276527,
-      "eval_combined_score": 0.3159163987138264,
-      "eval_f1": 0.0,
-      "eval_loss": 0.6590038537979126,
-      "eval_runtime": 70.2012,
-      "eval_samples_per_second": 575.916,
-      "eval_steps_per_second": 4.501,
-      "step": 25587
-    },
-    {
-      "epoch": 10.0,
-      "learning_rate": 0.0004,
-      "loss": 0.6713,
-      "step": 28430
-    },
-    {
-      "epoch": 10.0,
-      "eval_accuracy": 0.6318327974276527,
-      "eval_combined_score": 0.3159163987138264,
-      "eval_f1": 0.0,
-      "eval_loss": 0.6693503856658936,
-      "eval_runtime": 69.9698,
-      "eval_samples_per_second": 577.821,
-      "eval_steps_per_second": 4.516,
-      "step": 28430
-    },
-    {
-      "epoch": 11.0,
-      "learning_rate": 0.00039000000000000005,
-      "loss": 2.5009,
-      "step": 31273
-    },
-    {
-      "epoch": 11.0,
-      "eval_accuracy": 0.36816720257234725,
-      "eval_combined_score": 0.45317878342483403,
-      "eval_f1": 0.5381903642773208,
-      "eval_loss": 3.0633163452148438,
-      "eval_runtime": 69.917,
-      "eval_samples_per_second": 578.257,
-      "eval_steps_per_second": 4.52,
-      "step": 31273
-    },
-    {
-      "epoch": 11.0,
-      "step": 31273,
-      "total_flos": 5.856941202410045e+17,
-      "train_loss": 23.176317952643103,
-      "train_runtime": 20740.4239,
-      "train_samples_per_second": 877.142,
-      "train_steps_per_second": 6.854
+      "epoch": 8.0,
+      "step": 22744,
+      "total_flos": 4.25959360175276e+17,
+      "train_loss": 0.3639708930574753,
+      "train_runtime": 14995.3156,
+      "train_samples_per_second": 1213.199,
+      "train_steps_per_second": 9.48
     }
   ],
   "max_steps": 142150,
   "num_train_epochs": 50,
-  "total_flos": 5.856941202410045e+17,
+  "total_flos": 4.25959360175276e+17,
   "trial_name": null,
   "trial_params": null
 }