Hugging Face's logo

Sampath1987
/
EnergyEmbed-v1

Sentence Similarity
sentence-transformers
Safetensors
new
feature-extraction
dense
Generated from Trainer
dataset_size:53913
loss:MultipleNegativesRankingLoss
custom_code
Eval Results
text-embeddings-inference
Model card Files
xet
Community
Sampath1987 commited on
Commit
6ef3486
·
verified ·
1 Parent(s): 28c7271

fine-tuned model-v1 5 epochs

Browse files
Files changed (2) hide show
  1. README.md +36 -24
  2. model.safetensors +1 -1
README.md CHANGED
@@ -495,7 +495,7 @@ model-index:
495
  type: ai-job-validation
496
  metrics:
497
  - type: cosine_accuracy
498
- value: 0.7850282788276672
499
  name: Cosine Accuracy
500
  ---
501
 
@@ -561,9 +561,9 @@ print(embeddings.shape)
561
  # Get the similarity scores for the embeddings
562
  similarities = model.similarity(embeddings, embeddings)
563
  print(similarities)
564
- # tensor([[1.0000, 0.4457, 0.3235],
565
- # [0.4457, 1.0000, 0.3388],
566
- # [0.3235, 0.3388, 1.0000]])
567
  ```
568
 
569
  <!--
@@ -599,9 +599,9 @@ You can finetune this model on your own dataset.
599
  * Dataset: `ai-job-validation`
600
  * Evaluated with [<code>TripletEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.TripletEvaluator)
601
 
602
- | Metric | Value |
603
- |:--------------------|:----------|
604
- | **cosine_accuracy** | **0.785** |
605
 
606
  <!--
607
  ## Bias, Risks and Limitations
@@ -678,6 +678,7 @@ You can finetune this model on your own dataset.
678
  - `per_device_train_batch_size`: 16
679
  - `per_device_eval_batch_size`: 16
680
  - `learning_rate`: 2e-05
 
681
  - `warmup_ratio`: 0.1
682
 
683
  #### All Hyperparameters
@@ -700,7 +701,7 @@ You can finetune this model on your own dataset.
700
  - `adam_beta2`: 0.999
701
  - `adam_epsilon`: 1e-08
702
  - `max_grad_norm`: 1.0
703
- - `num_train_epochs`: 3
704
  - `max_steps`: -1
705
  - `lr_scheduler_type`: linear
706
  - `lr_scheduler_kwargs`: {}
@@ -806,22 +807,33 @@ You can finetune this model on your own dataset.
806
  ### Training Logs
807
  | Epoch | Step | Training Loss | Validation Loss | ai-job-validation_cosine_accuracy |
808
  |:------:|:-----:|:-------------:|:---------------:|:---------------------------------:|
809
- | 0.1795 | 1000 | - | 1.1634 | 0.6597 |
810
- | 0.3590 | 2000 | - | 1.0971 | 0.6821 |
811
- | 0.5385 | 3000 | - | 1.0596 | 0.7050 |
812
- | 0.7180 | 4000 | - | 1.0336 | 0.7193 |
813
- | 0.8975 | 5000 | 1.2066 | 1.0073 | 0.7312 |
814
- | 1.0770 | 6000 | - | 1.0060 | 0.7331 |
815
- | 1.2565 | 7000 | - | 0.9794 | 0.7465 |
816
- | 1.4360 | 8000 | - | 0.9657 | 0.7580 |
817
- | 1.6155 | 9000 | - | 0.9498 | 0.7593 |
818
- | 1.7950 | 10000 | 0.935 | 0.9387 | 0.7678 |
819
- | 1.9745 | 11000 | - | 0.9293 | 0.7623 |
820
- | 2.1540 | 12000 | - | 0.9313 | 0.7769 |
821
- | 2.3335 | 13000 | - | 0.9245 | 0.7794 |
822
- | 2.5130 | 14000 | - | 0.9190 | 0.7787 |
823
- | 2.6925 | 15000 | 0.7607 | 0.9139 | 0.7782 |
824
- | 2.8720 | 16000 | - | 0.9094 | 0.7850 |
 
 
 
 
 
 
 
 
 
 
 
825
 
826
 
827
  ### Framework Versions
 
495
  type: ai-job-validation
496
  metrics:
497
  - type: cosine_accuracy
498
+ value: 0.8222780823707581
499
  name: Cosine Accuracy
500
  ---
501
 
 
561
  # Get the similarity scores for the embeddings
562
  similarities = model.similarity(embeddings, embeddings)
563
  print(similarities)
564
+ # tensor([[1.0000, 0.3074, 0.1837],
565
+ # [0.3074, 1.0000, 0.1640],
566
+ # [0.1837, 0.1640, 1.0000]])
567
  ```
568
 
569
  <!--
 
599
  * Dataset: `ai-job-validation`
600
  * Evaluated with [<code>TripletEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.TripletEvaluator)
601
 
602
+ | Metric | Value |
603
+ |:--------------------|:-----------|
604
+ | **cosine_accuracy** | **0.8223** |
605
 
606
  <!--
607
  ## Bias, Risks and Limitations
 
678
  - `per_device_train_batch_size`: 16
679
  - `per_device_eval_batch_size`: 16
680
  - `learning_rate`: 2e-05
681
+ - `num_train_epochs`: 5
682
  - `warmup_ratio`: 0.1
683
 
684
  #### All Hyperparameters
 
701
  - `adam_beta2`: 0.999
702
  - `adam_epsilon`: 1e-08
703
  - `max_grad_norm`: 1.0
704
+ - `num_train_epochs`: 5
705
  - `max_steps`: -1
706
  - `lr_scheduler_type`: linear
707
  - `lr_scheduler_kwargs`: {}
 
807
  ### Training Logs
808
  | Epoch | Step | Training Loss | Validation Loss | ai-job-validation_cosine_accuracy |
809
  |:------:|:-----:|:-------------:|:---------------:|:---------------------------------:|
810
+ | 0.1795 | 1000 | - | 1.1970 | 0.6482 |
811
+ | 0.3590 | 2000 | - | 1.1165 | 0.6762 |
812
+ | 0.5385 | 3000 | - | 1.0740 | 0.6986 |
813
+ | 0.7180 | 4000 | - | 1.0460 | 0.7152 |
814
+ | 0.8975 | 5000 | 1.2294 | 1.0200 | 0.7252 |
815
+ | 1.0770 | 6000 | - | 1.0162 | 0.7259 |
816
+ | 1.2565 | 7000 | - | 0.9827 | 0.7445 |
817
+ | 1.4360 | 8000 | - | 0.9690 | 0.7592 |
818
+ | 1.6155 | 9000 | - | 0.9499 | 0.7590 |
819
+ | 1.7950 | 10000 | 0.9515 | 0.9396 | 0.7673 |
820
+ | 1.9745 | 11000 | - | 0.9297 | 0.7617 |
821
+ | 2.1540 | 12000 | - | 0.9290 | 0.7770 |
822
+ | 2.3335 | 13000 | - | 0.9128 | 0.7862 |
823
+ | 2.5130 | 14000 | - | 0.9076 | 0.7846 |
824
+ | 2.6925 | 15000 | 0.744 | 0.8964 | 0.7815 |
825
+ | 2.8720 | 16000 | - | 0.8777 | 0.7990 |
826
+ | 3.0515 | 17000 | - | 0.8798 | 0.7966 |
827
+ | 3.2310 | 18000 | - | 0.8713 | 0.8026 |
828
+ | 3.4105 | 19000 | - | 0.8658 | 0.8062 |
829
+ | 3.5900 | 20000 | 0.5671 | 0.8513 | 0.8055 |
830
+ | 3.7695 | 21000 | - | 0.8387 | 0.8143 |
831
+ | 3.9490 | 22000 | - | 0.8295 | 0.8144 |
832
+ | 4.1285 | 23000 | - | 0.8327 | 0.8192 |
833
+ | 4.3080 | 24000 | - | 0.8332 | 0.8189 |
834
+ | 4.4875 | 25000 | 0.4463 | 0.8267 | 0.8192 |
835
+ | 4.6670 | 26000 | - | 0.8236 | 0.8208 |
836
+ | 4.8465 | 27000 | - | 0.8205 | 0.8223 |
837
 
838
 
839
  ### Framework Versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:10e213517c8ea188d9928352e4f840c7dc99f2a01e0ca935229048b0d46df0e8
3
  size 1221487872
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a63ea997a7559fcd0e90ca98efd4f32a45eeff8627ac95cbfa641380020706d5
3
  size 1221487872