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 MNLI
       type: glue
       config: mnli
       split: validation_matched
@@ -20,7 +22,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.5918492103922568
+      value: 0.5911310008136696
 ---
 
 <!-- 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_new_pretrain_w_init_48_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: 0.9226
-- Accuracy: 0.5918
+- Loss: 0.8954
+- Accuracy: 0.5911
 
 ## Model description
 

all_results.json

@@ -0,0 +1,21 @@
+{
+    "epoch": 18.0,
+    "epoch_mm": 18.0,
+    "eval_accuracy": 0.5917473255221599,
+    "eval_accuracy_mm": 0.5911310008136696,
+    "eval_loss": 0.8925370573997498,
+    "eval_loss_mm": 0.8954058289527893,
+    "eval_runtime": 17.7412,
+    "eval_runtime_mm": 17.7555,
+    "eval_samples": 9815,
+    "eval_samples_mm": 9832,
+    "eval_samples_per_second": 553.233,
+    "eval_samples_per_second_mm": 553.743,
+    "eval_steps_per_second": 4.34,
+    "eval_steps_per_second_mm": 4.337,
+    "train_loss": 0.8583978627810146,
+    "train_runtime": 36993.4768,
+    "train_samples": 392702,
+    "train_samples_per_second": 530.772,
+    "train_steps_per_second": 4.147
+}

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 18.0,
+    "epoch_mm": 18.0,
+    "eval_accuracy": 0.5917473255221599,
+    "eval_accuracy_mm": 0.5911310008136696,
+    "eval_loss": 0.8925370573997498,
+    "eval_loss_mm": 0.8954058289527893,
+    "eval_runtime": 17.7412,
+    "eval_runtime_mm": 17.7555,
+    "eval_samples": 9815,
+    "eval_samples_mm": 9832,
+    "eval_samples_per_second": 553.233,
+    "eval_samples_per_second_mm": 553.743,
+    "eval_steps_per_second": 4.34,
+    "eval_steps_per_second_mm": 4.337
+}

logs/events.out.tfevents.1686096193.serv-9207.2389173.26

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6478350ee311d2ec2233cdec4bef0cc6b7561bf4ecc0f5a9db8cb081ecc3ce6d
+size 698

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 18.0,
+    "train_loss": 0.8583978627810146,
+    "train_runtime": 36993.4768,
+    "train_samples": 392702,
+    "train_samples_per_second": 530.772,
+    "train_steps_per_second": 4.147
+}

trainer_state.json

@@ -0,0 +1,295 @@
+{
+  "best_metric": 0.8925370573997498,
+  "best_model_checkpoint": "hBERTv1_new_pretrain_w_init_48_mnli/checkpoint-39884",
+  "epoch": 18.0,
+  "global_step": 55224,
+  "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": 1.031,
+      "step": 3068
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.4718288334182374,
+      "eval_loss": 0.9883108139038086,
+      "eval_runtime": 18.3267,
+      "eval_samples_per_second": 535.558,
+      "eval_steps_per_second": 4.202,
+      "step": 3068
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.8400000000000005e-05,
+      "loss": 0.9713,
+      "step": 6136
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.5146204788588895,
+      "eval_loss": 0.9751732349395752,
+      "eval_runtime": 18.0531,
+      "eval_samples_per_second": 543.673,
+      "eval_steps_per_second": 4.265,
+      "step": 6136
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 3.76e-05,
+      "loss": 0.9588,
+      "step": 9204
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.5166581762608252,
+      "eval_loss": 0.9701634645462036,
+      "eval_runtime": 18.0165,
+      "eval_samples_per_second": 544.777,
+      "eval_steps_per_second": 4.274,
+      "step": 9204
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 3.680000000000001e-05,
+      "loss": 0.953,
+      "step": 12272
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.5355068772287316,
+      "eval_loss": 0.9589763283729553,
+      "eval_runtime": 17.9451,
+      "eval_samples_per_second": 546.945,
+      "eval_steps_per_second": 4.291,
+      "step": 12272
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 3.6e-05,
+      "loss": 0.9323,
+      "step": 15340
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.5416199694345389,
+      "eval_loss": 0.937126874923706,
+      "eval_runtime": 17.9501,
+      "eval_samples_per_second": 546.794,
+      "eval_steps_per_second": 4.29,
+      "step": 15340
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 3.52e-05,
+      "loss": 0.9068,
+      "step": 18408
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.5590422822210902,
+      "eval_loss": 0.9212560057640076,
+      "eval_runtime": 17.9394,
+      "eval_samples_per_second": 547.119,
+      "eval_steps_per_second": 4.292,
+      "step": 18408
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 3.44e-05,
+      "loss": 0.8891,
+      "step": 21476
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.5607743250127356,
+      "eval_loss": 0.9168298840522766,
+      "eval_runtime": 17.9494,
+      "eval_samples_per_second": 546.814,
+      "eval_steps_per_second": 4.29,
+      "step": 21476
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 3.3600000000000004e-05,
+      "loss": 0.8741,
+      "step": 24544
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.5452878247580234,
+      "eval_loss": 0.947225034236908,
+      "eval_runtime": 17.941,
+      "eval_samples_per_second": 547.072,
+      "eval_steps_per_second": 4.292,
+      "step": 24544
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 3.28e-05,
+      "loss": 0.8589,
+      "step": 27612
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.5793173713703516,
+      "eval_loss": 0.9185450077056885,
+      "eval_runtime": 17.9524,
+      "eval_samples_per_second": 546.723,
+      "eval_steps_per_second": 4.289,
+      "step": 27612
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 3.2000000000000005e-05,
+      "loss": 0.8432,
+      "step": 30680
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.574732552215996,
+      "eval_loss": 0.9133642315864563,
+      "eval_runtime": 17.9284,
+      "eval_samples_per_second": 547.457,
+      "eval_steps_per_second": 4.295,
+      "step": 30680
+    },
+    {
+      "epoch": 11.0,
+      "learning_rate": 3.1200000000000006e-05,
+      "loss": 0.8289,
+      "step": 33748
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": 0.5753438614365767,
+      "eval_loss": 0.9138672947883606,
+      "eval_runtime": 17.9205,
+      "eval_samples_per_second": 547.695,
+      "eval_steps_per_second": 4.297,
+      "step": 33748
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 3.0400000000000004e-05,
+      "loss": 0.8137,
+      "step": 36816
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 0.5766683647478349,
+      "eval_loss": 0.9113235473632812,
+      "eval_runtime": 17.9039,
+      "eval_samples_per_second": 548.205,
+      "eval_steps_per_second": 4.301,
+      "step": 36816
+    },
+    {
+      "epoch": 13.0,
+      "learning_rate": 2.96e-05,
+      "loss": 0.7988,
+      "step": 39884
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": 0.5917473255221599,
+      "eval_loss": 0.8925370573997498,
+      "eval_runtime": 17.9295,
+      "eval_samples_per_second": 547.423,
+      "eval_steps_per_second": 4.295,
+      "step": 39884
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 2.8800000000000002e-05,
+      "loss": 0.7828,
+      "step": 42952
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": 0.5859398879266429,
+      "eval_loss": 0.903650164604187,
+      "eval_runtime": 17.916,
+      "eval_samples_per_second": 547.834,
+      "eval_steps_per_second": 4.298,
+      "step": 42952
+    },
+    {
+      "epoch": 15.0,
+      "learning_rate": 2.8e-05,
+      "loss": 0.7705,
+      "step": 46020
+    },
+    {
+      "epoch": 15.0,
+      "eval_accuracy": 0.5865511971472236,
+      "eval_loss": 0.9128865599632263,
+      "eval_runtime": 17.9075,
+      "eval_samples_per_second": 548.095,
+      "eval_steps_per_second": 4.3,
+      "step": 46020
+    },
+    {
+      "epoch": 16.0,
+      "learning_rate": 2.7200000000000004e-05,
+      "loss": 0.7576,
+      "step": 49088
+    },
+    {
+      "epoch": 16.0,
+      "eval_accuracy": 0.5878757004584819,
+      "eval_loss": 0.923698902130127,
+      "eval_runtime": 17.9131,
+      "eval_samples_per_second": 547.924,
+      "eval_steps_per_second": 4.299,
+      "step": 49088
+    },
+    {
+      "epoch": 17.0,
+      "learning_rate": 2.6400000000000005e-05,
+      "loss": 0.7463,
+      "step": 52156
+    },
+    {
+      "epoch": 17.0,
+      "eval_accuracy": 0.5897096281202241,
+      "eval_loss": 0.9211834669113159,
+      "eval_runtime": 17.9472,
+      "eval_samples_per_second": 546.882,
+      "eval_steps_per_second": 4.29,
+      "step": 52156
+    },
+    {
+      "epoch": 18.0,
+      "learning_rate": 2.5600000000000002e-05,
+      "loss": 0.7341,
+      "step": 55224
+    },
+    {
+      "epoch": 18.0,
+      "eval_accuracy": 0.5918492103922568,
+      "eval_loss": 0.9225720763206482,
+      "eval_runtime": 17.905,
+      "eval_samples_per_second": 548.172,
+      "eval_steps_per_second": 4.3,
+      "step": 55224
+    },
+    {
+      "epoch": 18.0,
+      "step": 55224,
+      "total_flos": 1.0468505269536031e+18,
+      "train_loss": 0.8583978627810146,
+      "train_runtime": 36993.4768,
+      "train_samples_per_second": 530.772,
+      "train_steps_per_second": 4.147
+    }
+  ],
+  "max_steps": 153400,
+  "num_train_epochs": 50,
+  "total_flos": 1.0468505269536031e+18,
+  "trial_name": null,
+  "trial_params": null
+}