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: 0.8271827850605986
+      value: 0.8117487014593124
     - name: F1
       type: f1
-      value: 0.7702476077735031
+      value: 0.7245086328591595
 ---
 
 <!-- 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. -->
 
 # hBERTv1_new_pretrain_48_KD_qqp
 
-This model is a fine-tuned version of [gokuls/bert_12_layer_model_v1_complete_training_new_48_KD](https://huggingface.co/gokuls/bert_12_layer_model_v1_complete_training_new_48_KD) on the glue dataset.
+This model is a fine-tuned version of [gokuls/bert_12_layer_model_v1_complete_training_new_48_KD](https://huggingface.co/gokuls/bert_12_layer_model_v1_complete_training_new_48_KD) on the GLUE QQP dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4820
-- Accuracy: 0.8272
-- F1: 0.7702
-- Combined Score: 0.7987
+- Loss: 0.4106
+- Accuracy: 0.8117
+- F1: 0.7245
+- Combined Score: 0.7681
 
 ## Model description
 

all_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 8.0,
+    "eval_accuracy": 0.8117487014593124,
+    "eval_combined_score": 0.768128667159236,
+    "eval_f1": 0.7245086328591595,
+    "eval_loss": 0.41062286496162415,
+    "eval_runtime": 83.7982,
+    "eval_samples": 40430,
+    "eval_samples_per_second": 482.469,
+    "eval_steps_per_second": 3.771,
+    "train_loss": 0.3166344636265677,
+    "train_runtime": 19509.5891,
+    "train_samples": 363846,
+    "train_samples_per_second": 932.48,
+    "train_steps_per_second": 7.286
+}

eval_results.json

@@ -0,0 +1,11 @@
+{
+    "epoch": 8.0,
+    "eval_accuracy": 0.8117487014593124,
+    "eval_combined_score": 0.768128667159236,
+    "eval_f1": 0.7245086328591595,
+    "eval_loss": 0.41062286496162415,
+    "eval_runtime": 83.7982,
+    "eval_samples": 40430,
+    "eval_samples_per_second": 482.469,
+    "eval_steps_per_second": 3.771
+}

logs/events.out.tfevents.1686791631.serv-9207.3372160.9

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b2a8de4ca1cf97b58228956e5bb973e2282476886fa0e6fae1c03b17f9c35776
+size 475

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 8.0,
+    "train_loss": 0.3166344636265677,
+    "train_runtime": 19509.5891,
+    "train_samples": 363846,
+    "train_samples_per_second": 932.48,
+    "train_steps_per_second": 7.286
+}

trainer_state.json

@@ -0,0 +1,161 @@
+{
+  "best_metric": 0.41062286496162415,
+  "best_model_checkpoint": "hBERTv1_new_pretrain_48_KD_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": 3.9200000000000004e-05,
+      "loss": 0.5016,
+      "step": 2843
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.777615631956468,
+      "eval_combined_score": 0.7369584662062361,
+      "eval_f1": 0.696301300456004,
+      "eval_loss": 0.4533442258834839,
+      "eval_runtime": 83.5598,
+      "eval_samples_per_second": 483.845,
+      "eval_steps_per_second": 3.782,
+      "step": 2843
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.8400000000000005e-05,
+      "loss": 0.4125,
+      "step": 5686
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.8023002720751917,
+      "eval_combined_score": 0.764615682141248,
+      "eval_f1": 0.7269310922073042,
+      "eval_loss": 0.44330382347106934,
+      "eval_runtime": 83.4995,
+      "eval_samples_per_second": 484.195,
+      "eval_steps_per_second": 3.784,
+      "step": 5686
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 3.76e-05,
+      "loss": 0.3574,
+      "step": 8529
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.8117487014593124,
+      "eval_combined_score": 0.768128667159236,
+      "eval_f1": 0.7245086328591595,
+      "eval_loss": 0.41062286496162415,
+      "eval_runtime": 84.5872,
+      "eval_samples_per_second": 477.968,
+      "eval_steps_per_second": 3.736,
+      "step": 8529
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 3.680000000000001e-05,
+      "loss": 0.3134,
+      "step": 11372
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.8207766510017314,
+      "eval_combined_score": 0.7834347898641708,
+      "eval_f1": 0.7460929287266103,
+      "eval_loss": 0.43951016664505005,
+      "eval_runtime": 84.745,
+      "eval_samples_per_second": 477.078,
+      "eval_steps_per_second": 3.729,
+      "step": 11372
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 3.6e-05,
+      "loss": 0.279,
+      "step": 14215
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.8235716052436309,
+      "eval_combined_score": 0.793112126116974,
+      "eval_f1": 0.7626526469903171,
+      "eval_loss": 0.4975138306617737,
+      "eval_runtime": 84.8847,
+      "eval_samples_per_second": 476.293,
+      "eval_steps_per_second": 3.723,
+      "step": 14215
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 3.52e-05,
+      "loss": 0.248,
+      "step": 17058
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.8129359386594114,
+      "eval_combined_score": 0.7597674223304428,
+      "eval_f1": 0.7065989060014742,
+      "eval_loss": 0.5526914000511169,
+      "eval_runtime": 85.0627,
+      "eval_samples_per_second": 475.297,
+      "eval_steps_per_second": 3.715,
+      "step": 17058
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 3.44e-05,
+      "loss": 0.2215,
+      "step": 19901
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.8209497897600792,
+      "eval_combined_score": 0.7953398629113403,
+      "eval_f1": 0.7697299360626014,
+      "eval_loss": 0.4813808798789978,
+      "eval_runtime": 84.9371,
+      "eval_samples_per_second": 475.999,
+      "eval_steps_per_second": 3.72,
+      "step": 19901
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 3.3600000000000004e-05,
+      "loss": 0.1998,
+      "step": 22744
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.8271827850605986,
+      "eval_combined_score": 0.7987151964170509,
+      "eval_f1": 0.7702476077735031,
+      "eval_loss": 0.48196226358413696,
+      "eval_runtime": 85.1204,
+      "eval_samples_per_second": 474.974,
+      "eval_steps_per_second": 3.712,
+      "step": 22744
+    },
+    {
+      "epoch": 8.0,
+      "step": 22744,
+      "total_flos": 4.310753534195794e+17,
+      "train_loss": 0.3166344636265677,
+      "train_runtime": 19509.5891,
+      "train_samples_per_second": 932.48,
+      "train_steps_per_second": 7.286
+    }
+  ],
+  "max_steps": 142150,
+  "num_train_epochs": 50,
+  "total_flos": 4.310753534195794e+17,
+  "trial_name": null,
+  "trial_params": null
+}