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
@@ -13,7 +15,7 @@ model-index:
       name: Text Classification
       type: text-classification
     dataset:
-      name: glue
+      name: GLUE MNLI
       type: glue
       config: mnli
       split: validation_matched
@@ -21,7 +23,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.40275089149261334
+      value: 0.459519934906428
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -29,10 +31,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # hBERTv1_new_pretrain_w_init_48_ver2_mnli
 
-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 MNLI dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.0687
-- Accuracy: 0.4028
+- Loss: 1.0270
+- Accuracy: 0.4595
 
 ## Model description
 

all_results.json

@@ -0,0 +1,21 @@
+{
+    "epoch": 8.0,
+    "epoch_mm": 8.0,
+    "eval_accuracy": 0.4431991849210392,
+    "eval_accuracy_mm": 0.459519934906428,
+    "eval_loss": 1.0475844144821167,
+    "eval_loss_mm": 1.0269584655761719,
+    "eval_runtime": 50.1044,
+    "eval_runtime_mm": 50.9493,
+    "eval_samples": 9815,
+    "eval_samples_mm": 9832,
+    "eval_samples_per_second": 195.891,
+    "eval_samples_per_second_mm": 192.976,
+    "eval_steps_per_second": 3.074,
+    "eval_steps_per_second_mm": 3.023,
+    "train_loss": 1.0517125135916618,
+    "train_runtime": 49706.4079,
+    "train_samples": 392702,
+    "train_samples_per_second": 118.506,
+    "train_steps_per_second": 1.852
+}

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 8.0,
+    "epoch_mm": 8.0,
+    "eval_accuracy": 0.4431991849210392,
+    "eval_accuracy_mm": 0.459519934906428,
+    "eval_loss": 1.0475844144821167,
+    "eval_loss_mm": 1.0269584655761719,
+    "eval_runtime": 50.1044,
+    "eval_runtime_mm": 50.9493,
+    "eval_samples": 9815,
+    "eval_samples_mm": 9832,
+    "eval_samples_per_second": 195.891,
+    "eval_samples_per_second_mm": 192.976,
+    "eval_steps_per_second": 3.074,
+    "eval_steps_per_second_mm": 3.023
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 8.0,
+    "train_loss": 1.0517125135916618,
+    "train_runtime": 49706.4079,
+    "train_samples": 392702,
+    "train_samples_per_second": 118.506,
+    "train_steps_per_second": 1.852
+}

trainer_state.json

@@ -0,0 +1,148 @@
+{
+  "best_metric": 1.0475844144821167,
+  "best_model_checkpoint": "hBERTv1_new_pretrain_w_init_48_ver2_mnli/checkpoint-18408",
+  "epoch": 8.0,
+  "eval_steps": 500,
+  "global_step": 49088,
+  "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": 1.0624,
+      "step": 6136
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.38400407539480386,
+      "eval_loss": 1.0714952945709229,
+      "eval_runtime": 51.4471,
+      "eval_samples_per_second": 190.778,
+      "eval_steps_per_second": 2.993,
+      "step": 6136
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.466666666666667e-05,
+      "loss": 1.0497,
+      "step": 12272
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.4072338257768721,
+      "eval_loss": 1.054832100868225,
+      "eval_runtime": 51.1874,
+      "eval_samples_per_second": 191.747,
+      "eval_steps_per_second": 3.009,
+      "step": 12272
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 3.2000000000000005e-05,
+      "loss": 1.0421,
+      "step": 18408
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.4431991849210392,
+      "eval_loss": 1.0475844144821167,
+      "eval_runtime": 51.2868,
+      "eval_samples_per_second": 191.375,
+      "eval_steps_per_second": 3.003,
+      "step": 18408
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 2.9333333333333333e-05,
+      "loss": 1.0485,
+      "step": 24544
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.441365257259297,
+      "eval_loss": 1.0546945333480835,
+      "eval_runtime": 51.0485,
+      "eval_samples_per_second": 192.268,
+      "eval_steps_per_second": 3.017,
+      "step": 24544
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 2.6666666666666667e-05,
+      "loss": 1.0473,
+      "step": 30680
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.4553234844625573,
+      "eval_loss": 1.0515663623809814,
+      "eval_runtime": 51.1945,
+      "eval_samples_per_second": 191.72,
+      "eval_steps_per_second": 3.008,
+      "step": 30680
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 2.4e-05,
+      "loss": 1.0498,
+      "step": 36816
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.44268976057055526,
+      "eval_loss": 1.0555789470672607,
+      "eval_runtime": 51.1668,
+      "eval_samples_per_second": 191.824,
+      "eval_steps_per_second": 3.01,
+      "step": 36816
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 2.1333333333333335e-05,
+      "loss": 1.0531,
+      "step": 42952
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.4381049414161997,
+      "eval_loss": 1.0556291341781616,
+      "eval_runtime": 51.1327,
+      "eval_samples_per_second": 191.952,
+      "eval_steps_per_second": 3.012,
+      "step": 42952
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 1.866666666666667e-05,
+      "loss": 1.0609,
+      "step": 49088
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.40275089149261334,
+      "eval_loss": 1.0686805248260498,
+      "eval_runtime": 51.1023,
+      "eval_samples_per_second": 192.066,
+      "eval_steps_per_second": 3.014,
+      "step": 49088
+    },
+    {
+      "epoch": 8.0,
+      "step": 49088,
+      "total_flos": 4.65266900868268e+17,
+      "train_loss": 1.0517125135916618,
+      "train_runtime": 49706.4079,
+      "train_samples_per_second": 118.506,
+      "train_steps_per_second": 1.852
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 92040,
+  "num_train_epochs": 15,
+  "save_steps": 500,
+  "total_flos": 4.65266900868268e+17,
+  "trial_name": null,
+  "trial_params": null
+}