End of training

README.md

@@ -1,4 +1,6 @@
 ---
+language:
+- en
 tags:
 - generated_from_trainer
 datasets:
@@ -12,7 +14,7 @@ model-index:
       name: Text Classification
       type: text-classification
     dataset:
-      name: glue
+      name: GLUE QNLI
       type: glue
       config: qnli
       split: validation
@@ -20,7 +22,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.7752150832875709
+      value: 0.8112758557569101
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -28,10 +30,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # hBERTv1_qnli
 
-This model is a fine-tuned version of [gokuls/bert_12_layer_model_v1](https://huggingface.co/gokuls/bert_12_layer_model_v1) on the glue dataset.
+This model is a fine-tuned version of [gokuls/bert_12_layer_model_v1](https://huggingface.co/gokuls/bert_12_layer_model_v1) on the GLUE QNLI dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.7668
-- Accuracy: 0.7752
+- Loss: 0.4198
+- Accuracy: 0.8113
 
 ## Model description
 

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 8.0,
+    "eval_accuracy": 0.8112758557569101,
+    "eval_loss": 0.4198276698589325,
+    "eval_runtime": 6.7804,
+    "eval_samples": 5463,
+    "eval_samples_per_second": 805.699,
+    "eval_steps_per_second": 3.245,
+    "train_loss": 0.3110694268854653,
+    "train_runtime": 3241.717,
+    "train_samples": 104743,
+    "train_samples_per_second": 1615.548,
+    "train_steps_per_second": 6.324
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 8.0,
+    "eval_accuracy": 0.8112758557569101,
+    "eval_loss": 0.4198276698589325,
+    "eval_runtime": 6.7804,
+    "eval_samples": 5463,
+    "eval_samples_per_second": 805.699,
+    "eval_steps_per_second": 3.245
+}

logs/events.out.tfevents.1677684770.serv-9220.1178704.8

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2033be8d7a3e737be96839feb22048425d6600f71fb1d91b12037029d4c737f9
+size 363

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 8.0,
+    "train_loss": 0.3110694268854653,
+    "train_runtime": 3241.717,
+    "train_samples": 104743,
+    "train_samples_per_second": 1615.548,
+    "train_steps_per_second": 6.324
+}

trainer_state.json

@@ -0,0 +1,145 @@
+{
+  "best_metric": 0.4198276698589325,
+  "best_model_checkpoint": "hBERTv1_qnli/checkpoint-1230",
+  "epoch": 8.0,
+  "global_step": 3280,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.9e-05,
+      "loss": 0.6667,
+      "step": 410
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.6873512721947648,
+      "eval_loss": 0.5954663157463074,
+      "eval_runtime": 6.817,
+      "eval_samples_per_second": 801.373,
+      "eval_steps_per_second": 3.227,
+      "step": 410
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 4.8e-05,
+      "loss": 0.4998,
+      "step": 820
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.794801391177009,
+      "eval_loss": 0.44860708713531494,
+      "eval_runtime": 6.81,
+      "eval_samples_per_second": 802.199,
+      "eval_steps_per_second": 3.231,
+      "step": 820
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 4.7e-05,
+      "loss": 0.3985,
+      "step": 1230
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.8112758557569101,
+      "eval_loss": 0.4198276698589325,
+      "eval_runtime": 6.8165,
+      "eval_samples_per_second": 801.434,
+      "eval_steps_per_second": 3.227,
+      "step": 1230
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.600000000000001e-05,
+      "loss": 0.3106,
+      "step": 1640
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.7865641588870584,
+      "eval_loss": 0.48407182097435,
+      "eval_runtime": 6.8451,
+      "eval_samples_per_second": 798.094,
+      "eval_steps_per_second": 3.214,
+      "step": 1640
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 4.5e-05,
+      "loss": 0.2286,
+      "step": 2050
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.790591250228812,
+      "eval_loss": 0.5340340733528137,
+      "eval_runtime": 6.828,
+      "eval_samples_per_second": 800.093,
+      "eval_steps_per_second": 3.222,
+      "step": 2050
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.4000000000000006e-05,
+      "loss": 0.1662,
+      "step": 2460
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.7728354384038074,
+      "eval_loss": 0.6281628608703613,
+      "eval_runtime": 6.835,
+      "eval_samples_per_second": 799.265,
+      "eval_steps_per_second": 3.219,
+      "step": 2460
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 4.3e-05,
+      "loss": 0.1237,
+      "step": 2870
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.7752150832875709,
+      "eval_loss": 0.6678447723388672,
+      "eval_runtime": 6.8476,
+      "eval_samples_per_second": 797.793,
+      "eval_steps_per_second": 3.213,
+      "step": 2870
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.2e-05,
+      "loss": 0.0945,
+      "step": 3280
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.7752150832875709,
+      "eval_loss": 0.7668098211288452,
+      "eval_runtime": 6.8357,
+      "eval_samples_per_second": 799.183,
+      "eval_steps_per_second": 3.218,
+      "step": 3280
+    },
+    {
+      "epoch": 8.0,
+      "step": 3280,
+      "total_flos": 1.0584157486002995e+17,
+      "train_loss": 0.3110694268854653,
+      "train_runtime": 3241.717,
+      "train_samples_per_second": 1615.548,
+      "train_steps_per_second": 6.324
+    }
+  ],
+  "max_steps": 20500,
+  "num_train_epochs": 50,
+  "total_flos": 1.0584157486002995e+17,
+  "trial_name": null,
+  "trial_params": null
+}