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 STSB
       type: glue
       config: stsb
       split: validation
@@ -20,7 +22,7 @@ model-index:
     metrics:
     - name: Spearmanr
       type: spearmanr
-      value: 0.4507892146083376
+      value: 0.45996385438365645
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -28,12 +30,12 @@ should probably proofread and complete it, then remove this comment. -->
 
 # hBERTv1_new_pretrain_48_emb_com_stsb
 
-This model is a fine-tuned version of [gokuls/bert_12_layer_model_v1_complete_training_new_emb_compress_48](https://huggingface.co/gokuls/bert_12_layer_model_v1_complete_training_new_emb_compress_48) on the glue dataset.
+This model is a fine-tuned version of [gokuls/bert_12_layer_model_v1_complete_training_new_emb_compress_48](https://huggingface.co/gokuls/bert_12_layer_model_v1_complete_training_new_emb_compress_48) on the GLUE STSB dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.9571
-- Pearson: 0.4588
-- Spearmanr: 0.4508
-- Combined Score: 0.4548
+- Loss: 1.9214
+- Pearson: 0.4648
+- Spearmanr: 0.4600
+- Combined Score: 0.4624
 
 ## Model description
 

all_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 27.0,
+    "eval_combined_score": 0.46235708950698207,
+    "eval_loss": 1.9214489459991455,
+    "eval_pearson": 0.46475032463030774,
+    "eval_runtime": 6.7246,
+    "eval_samples": 1500,
+    "eval_samples_per_second": 223.06,
+    "eval_spearmanr": 0.45996385438365645,
+    "eval_steps_per_second": 1.784,
+    "train_loss": 0.6121585881268536,
+    "train_runtime": 2633.1293,
+    "train_samples": 5749,
+    "train_samples_per_second": 109.167,
+    "train_steps_per_second": 0.854
+}

eval_results.json

@@ -0,0 +1,11 @@
+{
+    "epoch": 27.0,
+    "eval_combined_score": 0.46235708950698207,
+    "eval_loss": 1.9214489459991455,
+    "eval_pearson": 0.46475032463030774,
+    "eval_runtime": 6.7246,
+    "eval_samples": 1500,
+    "eval_samples_per_second": 223.06,
+    "eval_spearmanr": 0.45996385438365645,
+    "eval_steps_per_second": 1.784
+}

logs/events.out.tfevents.1686837072.garda.2601621.13

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0b7df4c808478a7923c0b7dc1b3b1109b2f0f997cb4eaf0714d4d31f0883d7e
+size 473

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 27.0,
+    "train_loss": 0.6121585881268536,
+    "train_runtime": 2633.1293,
+    "train_samples": 5749,
+    "train_samples_per_second": 109.167,
+    "train_steps_per_second": 0.854
+}

trainer_state.json

@@ -0,0 +1,484 @@
+{
+  "best_metric": 1.9214489459991455,
+  "best_model_checkpoint": "hBERTv1_new_pretrain_48_emb_com_stsb/checkpoint-990",
+  "epoch": 27.0,
+  "global_step": 1215,
+  "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": 2.5817,
+      "step": 45
+    },
+    {
+      "epoch": 1.0,
+      "eval_combined_score": 0.1961723851082083,
+      "eval_loss": 2.602813720703125,
+      "eval_pearson": 0.2027356377543126,
+      "eval_runtime": 6.748,
+      "eval_samples_per_second": 222.289,
+      "eval_spearmanr": 0.189609132462104,
+      "eval_steps_per_second": 1.778,
+      "step": 45
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.8400000000000005e-05,
+      "loss": 2.1023,
+      "step": 90
+    },
+    {
+      "epoch": 2.0,
+      "eval_combined_score": 0.19863013345614586,
+      "eval_loss": 2.1595816612243652,
+      "eval_pearson": 0.20345395522854448,
+      "eval_runtime": 6.7492,
+      "eval_samples_per_second": 222.249,
+      "eval_spearmanr": 0.19380631168374723,
+      "eval_steps_per_second": 1.778,
+      "step": 90
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 3.76e-05,
+      "loss": 1.9567,
+      "step": 135
+    },
+    {
+      "epoch": 3.0,
+      "eval_combined_score": 0.18933959939433503,
+      "eval_loss": 2.3409199714660645,
+      "eval_pearson": 0.18554383257948298,
+      "eval_runtime": 6.7607,
+      "eval_samples_per_second": 221.871,
+      "eval_spearmanr": 0.19313536620918711,
+      "eval_steps_per_second": 1.775,
+      "step": 135
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 3.680000000000001e-05,
+      "loss": 1.7201,
+      "step": 180
+    },
+    {
+      "epoch": 4.0,
+      "eval_combined_score": 0.2899464590935818,
+      "eval_loss": 2.1790404319763184,
+      "eval_pearson": 0.28650208786193304,
+      "eval_runtime": 6.747,
+      "eval_samples_per_second": 222.32,
+      "eval_spearmanr": 0.29339083032523056,
+      "eval_steps_per_second": 1.779,
+      "step": 180
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 3.6e-05,
+      "loss": 1.5153,
+      "step": 225
+    },
+    {
+      "epoch": 5.0,
+      "eval_combined_score": 0.3366713577400152,
+      "eval_loss": 2.1208479404449463,
+      "eval_pearson": 0.3381105093417233,
+      "eval_runtime": 6.7661,
+      "eval_samples_per_second": 221.694,
+      "eval_spearmanr": 0.3352322061383072,
+      "eval_steps_per_second": 1.774,
+      "step": 225
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 3.52e-05,
+      "loss": 1.2674,
+      "step": 270
+    },
+    {
+      "epoch": 6.0,
+      "eval_combined_score": 0.388996037809745,
+      "eval_loss": 2.122382879257202,
+      "eval_pearson": 0.3881787749232017,
+      "eval_runtime": 6.7466,
+      "eval_samples_per_second": 222.334,
+      "eval_spearmanr": 0.3898133006962884,
+      "eval_steps_per_second": 1.779,
+      "step": 270
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 3.44e-05,
+      "loss": 1.0115,
+      "step": 315
+    },
+    {
+      "epoch": 7.0,
+      "eval_combined_score": 0.42925164874364347,
+      "eval_loss": 2.22530198097229,
+      "eval_pearson": 0.43036479777857173,
+      "eval_runtime": 6.7518,
+      "eval_samples_per_second": 222.163,
+      "eval_spearmanr": 0.4281384997087152,
+      "eval_steps_per_second": 1.777,
+      "step": 315
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 3.3600000000000004e-05,
+      "loss": 0.7449,
+      "step": 360
+    },
+    {
+      "epoch": 8.0,
+      "eval_combined_score": 0.4279396504678673,
+      "eval_loss": 2.323458433151245,
+      "eval_pearson": 0.42358611153077663,
+      "eval_runtime": 6.7608,
+      "eval_samples_per_second": 221.867,
+      "eval_spearmanr": 0.432293189404958,
+      "eval_steps_per_second": 1.775,
+      "step": 360
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 3.28e-05,
+      "loss": 0.66,
+      "step": 405
+    },
+    {
+      "epoch": 9.0,
+      "eval_combined_score": 0.4345560407048856,
+      "eval_loss": 2.361697196960449,
+      "eval_pearson": 0.4340089724216108,
+      "eval_runtime": 6.77,
+      "eval_samples_per_second": 221.566,
+      "eval_spearmanr": 0.4351031089881605,
+      "eval_steps_per_second": 1.773,
+      "step": 405
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 3.2000000000000005e-05,
+      "loss": 0.4678,
+      "step": 450
+    },
+    {
+      "epoch": 10.0,
+      "eval_combined_score": 0.4278924878236148,
+      "eval_loss": 2.0740506649017334,
+      "eval_pearson": 0.42995071418144987,
+      "eval_runtime": 6.7876,
+      "eval_samples_per_second": 220.99,
+      "eval_spearmanr": 0.4258342614657797,
+      "eval_steps_per_second": 1.768,
+      "step": 450
+    },
+    {
+      "epoch": 11.0,
+      "learning_rate": 3.1200000000000006e-05,
+      "loss": 0.4438,
+      "step": 495
+    },
+    {
+      "epoch": 11.0,
+      "eval_combined_score": 0.4289254482191033,
+      "eval_loss": 2.3816161155700684,
+      "eval_pearson": 0.4284999172461374,
+      "eval_runtime": 6.7943,
+      "eval_samples_per_second": 220.772,
+      "eval_spearmanr": 0.4293509791920692,
+      "eval_steps_per_second": 1.766,
+      "step": 495
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 3.0400000000000004e-05,
+      "loss": 0.3192,
+      "step": 540
+    },
+    {
+      "epoch": 12.0,
+      "eval_combined_score": 0.4590614019452154,
+      "eval_loss": 2.1672608852386475,
+      "eval_pearson": 0.45795017178280906,
+      "eval_runtime": 6.8031,
+      "eval_samples_per_second": 220.489,
+      "eval_spearmanr": 0.4601726321076217,
+      "eval_steps_per_second": 1.764,
+      "step": 540
+    },
+    {
+      "epoch": 13.0,
+      "learning_rate": 2.96e-05,
+      "loss": 0.2481,
+      "step": 585
+    },
+    {
+      "epoch": 13.0,
+      "eval_combined_score": 0.43744904640918025,
+      "eval_loss": 2.1544361114501953,
+      "eval_pearson": 0.4392148992975784,
+      "eval_runtime": 6.7836,
+      "eval_samples_per_second": 221.122,
+      "eval_spearmanr": 0.43568319352078216,
+      "eval_steps_per_second": 1.769,
+      "step": 585
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 2.8800000000000002e-05,
+      "loss": 0.2296,
+      "step": 630
+    },
+    {
+      "epoch": 14.0,
+      "eval_combined_score": 0.45925148211887024,
+      "eval_loss": 2.007450580596924,
+      "eval_pearson": 0.46026152385259717,
+      "eval_runtime": 6.8038,
+      "eval_samples_per_second": 220.465,
+      "eval_spearmanr": 0.4582414403851433,
+      "eval_steps_per_second": 1.764,
+      "step": 630
+    },
+    {
+      "epoch": 15.0,
+      "learning_rate": 2.8e-05,
+      "loss": 0.1765,
+      "step": 675
+    },
+    {
+      "epoch": 15.0,
+      "eval_combined_score": 0.4620511023513117,
+      "eval_loss": 2.1394569873809814,
+      "eval_pearson": 0.46235490780864075,
+      "eval_runtime": 6.7896,
+      "eval_samples_per_second": 220.927,
+      "eval_spearmanr": 0.46174729689398264,
+      "eval_steps_per_second": 1.767,
+      "step": 675
+    },
+    {
+      "epoch": 16.0,
+      "learning_rate": 2.7200000000000004e-05,
+      "loss": 0.1533,
+      "step": 720
+    },
+    {
+      "epoch": 16.0,
+      "eval_combined_score": 0.4469345688037569,
+      "eval_loss": 2.2714791297912598,
+      "eval_pearson": 0.45120593215851773,
+      "eval_runtime": 6.8234,
+      "eval_samples_per_second": 219.832,
+      "eval_spearmanr": 0.4426632054489961,
+      "eval_steps_per_second": 1.759,
+      "step": 720
+    },
+    {
+      "epoch": 17.0,
+      "learning_rate": 2.6400000000000005e-05,
+      "loss": 0.1343,
+      "step": 765
+    },
+    {
+      "epoch": 17.0,
+      "eval_combined_score": 0.4428810611082898,
+      "eval_loss": 2.172586679458618,
+      "eval_pearson": 0.44408986762850433,
+      "eval_runtime": 6.787,
+      "eval_samples_per_second": 221.01,
+      "eval_spearmanr": 0.44167225458807524,
+      "eval_steps_per_second": 1.768,
+      "step": 765
+    },
+    {
+      "epoch": 18.0,
+      "learning_rate": 2.5600000000000002e-05,
+      "loss": 0.1373,
+      "step": 810
+    },
+    {
+      "epoch": 18.0,
+      "eval_combined_score": 0.4478234688937688,
+      "eval_loss": 2.02228045463562,
+      "eval_pearson": 0.4532121986881763,
+      "eval_runtime": 6.7875,
+      "eval_samples_per_second": 220.996,
+      "eval_spearmanr": 0.4424347390993613,
+      "eval_steps_per_second": 1.768,
+      "step": 810
+    },
+    {
+      "epoch": 19.0,
+      "learning_rate": 2.4800000000000003e-05,
+      "loss": 0.1277,
+      "step": 855
+    },
+    {
+      "epoch": 19.0,
+      "eval_combined_score": 0.4346959695013698,
+      "eval_loss": 1.9992105960845947,
+      "eval_pearson": 0.4395412779716168,
+      "eval_runtime": 6.7974,
+      "eval_samples_per_second": 220.673,
+      "eval_spearmanr": 0.42985066103112274,
+      "eval_steps_per_second": 1.765,
+      "step": 855
+    },
+    {
+      "epoch": 20.0,
+      "learning_rate": 2.4e-05,
+      "loss": 0.0968,
+      "step": 900
+    },
+    {
+      "epoch": 20.0,
+      "eval_combined_score": 0.46103377995091277,
+      "eval_loss": 2.1078310012817383,
+      "eval_pearson": 0.4620119136399576,
+      "eval_runtime": 6.7915,
+      "eval_samples_per_second": 220.864,
+      "eval_spearmanr": 0.460055646261868,
+      "eval_steps_per_second": 1.767,
+      "step": 900
+    },
+    {
+      "epoch": 21.0,
+      "learning_rate": 2.32e-05,
+      "loss": 0.084,
+      "step": 945
+    },
+    {
+      "epoch": 21.0,
+      "eval_combined_score": 0.460210810698569,
+      "eval_loss": 2.068389415740967,
+      "eval_pearson": 0.46274616538784863,
+      "eval_runtime": 6.7629,
+      "eval_samples_per_second": 221.8,
+      "eval_spearmanr": 0.4576754560092894,
+      "eval_steps_per_second": 1.774,
+      "step": 945
+    },
+    {
+      "epoch": 22.0,
+      "learning_rate": 2.2400000000000002e-05,
+      "loss": 0.0777,
+      "step": 990
+    },
+    {
+      "epoch": 22.0,
+      "eval_combined_score": 0.46235708950698207,
+      "eval_loss": 1.9214489459991455,
+      "eval_pearson": 0.46475032463030774,
+      "eval_runtime": 6.7758,
+      "eval_samples_per_second": 221.376,
+      "eval_spearmanr": 0.45996385438365645,
+      "eval_steps_per_second": 1.771,
+      "step": 990
+    },
+    {
+      "epoch": 23.0,
+      "learning_rate": 2.1600000000000003e-05,
+      "loss": 0.0572,
+      "step": 1035
+    },
+    {
+      "epoch": 23.0,
+      "eval_combined_score": 0.44643779764240843,
+      "eval_loss": 2.0636143684387207,
+      "eval_pearson": 0.45063900222949543,
+      "eval_runtime": 6.7543,
+      "eval_samples_per_second": 222.082,
+      "eval_spearmanr": 0.4422365930553215,
+      "eval_steps_per_second": 1.777,
+      "step": 1035
+    },
+    {
+      "epoch": 24.0,
+      "learning_rate": 2.08e-05,
+      "loss": 0.0615,
+      "step": 1080
+    },
+    {
+      "epoch": 24.0,
+      "eval_combined_score": 0.44382684994837796,
+      "eval_loss": 2.0404300689697266,
+      "eval_pearson": 0.4488704472998438,
+      "eval_runtime": 6.7964,
+      "eval_samples_per_second": 220.705,
+      "eval_spearmanr": 0.43878325259691214,
+      "eval_steps_per_second": 1.766,
+      "step": 1080
+    },
+    {
+      "epoch": 25.0,
+      "learning_rate": 2e-05,
+      "loss": 0.0516,
+      "step": 1125
+    },
+    {
+      "epoch": 25.0,
+      "eval_combined_score": 0.4475345905545639,
+      "eval_loss": 2.059901714324951,
+      "eval_pearson": 0.4515878825695937,
+      "eval_runtime": 6.7915,
+      "eval_samples_per_second": 220.866,
+      "eval_spearmanr": 0.44348129853953405,
+      "eval_steps_per_second": 1.767,
+      "step": 1125
+    },
+    {
+      "epoch": 26.0,
+      "learning_rate": 1.9200000000000003e-05,
+      "loss": 0.0501,
+      "step": 1170
+    },
+    {
+      "epoch": 26.0,
+      "eval_combined_score": 0.4509567766877898,
+      "eval_loss": 2.0359442234039307,
+      "eval_pearson": 0.4530378970256098,
+      "eval_runtime": 6.8254,
+      "eval_samples_per_second": 219.768,
+      "eval_spearmanr": 0.44887565634996973,
+      "eval_steps_per_second": 1.758,
+      "step": 1170
+    },
+    {
+      "epoch": 27.0,
+      "learning_rate": 1.8400000000000003e-05,
+      "loss": 0.0515,
+      "step": 1215
+    },
+    {
+      "epoch": 27.0,
+      "eval_combined_score": 0.45481392235284457,
+      "eval_loss": 1.9571282863616943,
+      "eval_pearson": 0.4588386300973516,
+      "eval_runtime": 6.8054,
+      "eval_samples_per_second": 220.412,
+      "eval_spearmanr": 0.4507892146083376,
+      "eval_steps_per_second": 1.763,
+      "step": 1215
+    },
+    {
+      "epoch": 27.0,
+      "step": 1215,
+      "total_flos": 2.301115836137472e+16,
+      "train_loss": 0.6121585881268536,
+      "train_runtime": 2633.1293,
+      "train_samples_per_second": 109.167,
+      "train_steps_per_second": 0.854
+    }
+  ],
+  "max_steps": 2250,
+  "num_train_epochs": 50,
+  "total_flos": 2.301115836137472e+16,
+  "trial_name": null,
+  "trial_params": null
+}