End of training

README.md

@@ -1,4 +1,6 @@
 ---
+language:
+- en
 base_model: gokuls/bert_12_layer_model_v1_complete_training_new_wt_init_48
 tags:
 - generated_from_trainer
@@ -14,7 +16,7 @@ model-index:
       name: Text Classification
       type: text-classification
     dataset:
-      name: glue
+      name: GLUE QQP
       type: glue
       config: qqp
       split: validation
@@ -22,10 +24,10 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.7089290131090774
+      value: 0.7601038832550087
     - name: F1
       type: f1
-      value: 0.6289335939963422
+      value: 0.6952012821721505
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -33,12 +35,12 @@ should probably proofread and complete it, then remove this comment. -->
 
 # hBERTv1_new_pretrain_w_init_48_ver2_qqp
 
-This model is a fine-tuned version of [gokuls/bert_12_layer_model_v1_complete_training_new_wt_init_48](https://huggingface.co/gokuls/bert_12_layer_model_v1_complete_training_new_wt_init_48) on the glue dataset.
+This model is a fine-tuned version of [gokuls/bert_12_layer_model_v1_complete_training_new_wt_init_48](https://huggingface.co/gokuls/bert_12_layer_model_v1_complete_training_new_wt_init_48) on the GLUE QQP dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5789
-- Accuracy: 0.7089
-- F1: 0.6289
-- Combined Score: 0.6689
+- Loss: 0.4918
+- Accuracy: 0.7601
+- F1: 0.6952
+- Combined Score: 0.7277
 
 ## Model description
 

all_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 6.0,
+    "eval_accuracy": 0.7601038832550087,
+    "eval_combined_score": 0.7276525827135796,
+    "eval_f1": 0.6952012821721505,
+    "eval_loss": 0.4918229579925537,
+    "eval_runtime": 211.0345,
+    "eval_samples": 40430,
+    "eval_samples_per_second": 191.58,
+    "eval_steps_per_second": 2.995,
+    "train_loss": 0.5074344026286237,
+    "train_runtime": 35514.6787,
+    "train_samples": 363846,
+    "train_samples_per_second": 153.674,
+    "train_steps_per_second": 2.402
+}

eval_results.json

@@ -0,0 +1,11 @@
+{
+    "epoch": 6.0,
+    "eval_accuracy": 0.7601038832550087,
+    "eval_combined_score": 0.7276525827135796,
+    "eval_f1": 0.6952012821721505,
+    "eval_loss": 0.4918229579925537,
+    "eval_runtime": 211.0345,
+    "eval_samples": 40430,
+    "eval_samples_per_second": 191.58,
+    "eval_steps_per_second": 2.995
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 6.0,
+    "train_loss": 0.5074344026286237,
+    "train_runtime": 35514.6787,
+    "train_samples": 363846,
+    "train_samples_per_second": 153.674,
+    "train_steps_per_second": 2.402
+}

trainer_state.json

@@ -0,0 +1,130 @@
+{
+  "best_metric": 0.4918229579925537,
+  "best_model_checkpoint": "hBERTv1_new_pretrain_w_init_48_ver2_qqp/checkpoint-5686",
+  "epoch": 6.0,
+  "eval_steps": 500,
+  "global_step": 34116,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.733333333333334e-05,
+      "loss": 0.5279,
+      "step": 5686
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.7601038832550087,
+      "eval_combined_score": 0.7276525827135796,
+      "eval_f1": 0.6952012821721505,
+      "eval_loss": 0.4918229579925537,
+      "eval_runtime": 213.3541,
+      "eval_samples_per_second": 189.497,
+      "eval_steps_per_second": 2.962,
+      "step": 5686
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.466666666666667e-05,
+      "loss": 0.4826,
+      "step": 11372
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.7643581498886965,
+      "eval_combined_score": 0.7099691042274545,
+      "eval_f1": 0.6555800585662124,
+      "eval_loss": 0.5367019176483154,
+      "eval_runtime": 211.9321,
+      "eval_samples_per_second": 190.769,
+      "eval_steps_per_second": 2.982,
+      "step": 11372
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 3.2000000000000005e-05,
+      "loss": 0.4943,
+      "step": 17058
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.7593618600049469,
+      "eval_combined_score": 0.7016571422623088,
+      "eval_f1": 0.6439524245196707,
+      "eval_loss": 0.5223015546798706,
+      "eval_runtime": 211.9395,
+      "eval_samples_per_second": 190.762,
+      "eval_steps_per_second": 2.982,
+      "step": 17058
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 2.9333333333333333e-05,
+      "loss": 0.492,
+      "step": 22744
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.7599802127133317,
+      "eval_combined_score": 0.7032452425944627,
+      "eval_f1": 0.6465102724755938,
+      "eval_loss": 0.5378891825675964,
+      "eval_runtime": 211.5301,
+      "eval_samples_per_second": 191.131,
+      "eval_steps_per_second": 2.988,
+      "step": 22744
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 2.6666666666666667e-05,
+      "loss": 0.505,
+      "step": 28430
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.74232005936186,
+      "eval_combined_score": 0.6965020361186982,
+      "eval_f1": 0.6506840128755366,
+      "eval_loss": 0.5431071519851685,
+      "eval_runtime": 211.3198,
+      "eval_samples_per_second": 191.321,
+      "eval_steps_per_second": 2.991,
+      "step": 28430
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 2.4e-05,
+      "loss": 0.5428,
+      "step": 34116
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.7089290131090774,
+      "eval_combined_score": 0.6689313035527098,
+      "eval_f1": 0.6289335939963422,
+      "eval_loss": 0.5789130330085754,
+      "eval_runtime": 211.7558,
+      "eval_samples_per_second": 190.927,
+      "eval_steps_per_second": 2.985,
+      "step": 34116
+    },
+    {
+      "epoch": 6.0,
+      "step": 34116,
+      "total_flos": 3.2330651506468454e+17,
+      "train_loss": 0.5074344026286237,
+      "train_runtime": 35514.6787,
+      "train_samples_per_second": 153.674,
+      "train_steps_per_second": 2.402
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 85290,
+  "num_train_epochs": 15,
+  "save_steps": 500,
+  "total_flos": 3.2330651506468454e+17,
+  "trial_name": null,
+  "trial_params": null
+}