Add new SentenceTransformer model

README.md

@@ -7,7 +7,7 @@ tags:
 - generated_from_trainer
 - dataset_size:90000
 - loss:MultipleNegativesRankingLoss
-base_model: sentence-transformers/all-MiniLM-L6-v2
+base_model: sentence-transformers/all-MiniLM-L12-v2
 widget:
 - source_sentence: what is the maximum i can contribute to a traditional ira
   sentences:
@@ -121,7 +121,7 @@ metrics:
 - cosine_mrr@10
 - cosine_map@100
 model-index:
-- name: SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2
+- name: SentenceTransformer based on sentence-transformers/all-MiniLM-L12-v2
   results:
   - task:
       type: information-retrieval
@@ -131,10 +131,10 @@ model-index:
       type: NanoMSMARCO
     metrics:
     - type: cosine_accuracy@1
-      value: 0.3
+      value: 0.32
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.62
+      value: 0.52
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
       value: 0.62
@@ -143,10 +143,10 @@ model-index:
       value: 0.72
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.3
+      value: 0.32
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.20666666666666667
+      value: 0.1733333333333333
       name: Cosine Precision@3
     - type: cosine_precision@5
       value: 0.124
@@ -155,10 +155,10 @@ model-index:
       value: 0.07200000000000001
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.3
+      value: 0.32
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.62
+      value: 0.52
       name: Cosine Recall@3
     - type: cosine_recall@5
       value: 0.62
@@ -167,13 +167,13 @@ model-index:
       value: 0.72
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.5180879550984706
+      value: 0.5182449787606596
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.4529126984126984
+      value: 0.45404761904761903
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.4665743772173455
+      value: 0.4681213273999474
       name: Cosine Map@100
   - task:
       type: information-retrieval
@@ -183,49 +183,49 @@ model-index:
       type: NanoNQ
     metrics:
     - type: cosine_accuracy@1
-      value: 0.36
+      value: 0.38
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.5
+      value: 0.52
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
-      value: 0.54
+      value: 0.58
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.6
+      value: 0.7
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.36
+      value: 0.38
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.1733333333333333
+      value: 0.18666666666666665
       name: Cosine Precision@3
     - type: cosine_precision@5
-      value: 0.11600000000000002
+      value: 0.128
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.066
+      value: 0.076
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.34
+      value: 0.36
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.47
+      value: 0.5
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.52
+      value: 0.57
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.59
+      value: 0.67
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.4736769259177555
+      value: 0.5134978713592498
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.44483333333333336
+      value: 0.47138888888888886
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.4475757336608197
+      value: 0.4692659575514759
       name: Cosine Map@100
   - task:
       type: nano-beir
@@ -235,61 +235,61 @@ model-index:
       type: NanoBEIR_mean
     metrics:
     - type: cosine_accuracy@1
-      value: 0.32999999999999996
+      value: 0.35
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.56
+      value: 0.52
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
-      value: 0.5800000000000001
+      value: 0.6
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.6599999999999999
+      value: 0.71
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.32999999999999996
+      value: 0.35
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.19
+      value: 0.18
       name: Cosine Precision@3
     - type: cosine_precision@5
-      value: 0.12000000000000001
+      value: 0.126
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.069
+      value: 0.07400000000000001
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.32
+      value: 0.33999999999999997
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.5449999999999999
+      value: 0.51
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.5700000000000001
+      value: 0.595
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.655
+      value: 0.6950000000000001
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.49588244050811303
+      value: 0.5158714250599548
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.44887301587301587
+      value: 0.46271825396825395
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.4570750554390826
+      value: 0.46869364247571166
       name: Cosine Map@100
 ---
 
-# SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2
+# SentenceTransformer based on sentence-transformers/all-MiniLM-L12-v2
 
-This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [sentence-transformers/all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2). It maps sentences & paragraphs to a 384-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
+This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [sentence-transformers/all-MiniLM-L12-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L12-v2). It maps sentences & paragraphs to a 384-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
 
 ## Model Details
 
 ### Model Description
 - **Model Type:** Sentence Transformer
-- **Base model:** [sentence-transformers/all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2) <!-- at revision c9745ed1d9f207416be6d2e6f8de32d1f16199bf -->
+- **Base model:** [sentence-transformers/all-MiniLM-L12-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L12-v2) <!-- at revision 936af83a2ecce5fe87a09109ff5cbcefe073173a -->
 - **Maximum Sequence Length:** 128 tokens
 - **Output Dimensionality:** 384 dimensions
 - **Similarity Function:** Cosine Similarity
@@ -342,9 +342,9 @@ print(embeddings.shape)
 # Get the similarity scores for the embeddings
 similarities = model.similarity(embeddings, embeddings)
 print(similarities)
-# tensor([[1.0000, 0.5217, 0.5263],
-#         [0.5217, 0.9999, 0.2880],
-#         [0.5263, 0.2880, 1.0000]])
+# tensor([[0.9999, 0.5493, 0.3900],
+#         [0.5493, 1.0000, 0.1239],
+#         [0.3900, 0.1239, 1.0001]])
 ```
 
 <!--
@@ -382,21 +382,21 @@ You can finetune this model on your own dataset.
 
 | Metric              | NanoMSMARCO | NanoNQ     |
 |:--------------------|:------------|:-----------|
-| cosine_accuracy@1   | 0.3         | 0.36       |
-| cosine_accuracy@3   | 0.62        | 0.5        |
-| cosine_accuracy@5   | 0.62        | 0.54       |
-| cosine_accuracy@10  | 0.72        | 0.6        |
-| cosine_precision@1  | 0.3         | 0.36       |
-| cosine_precision@3  | 0.2067      | 0.1733     |
-| cosine_precision@5  | 0.124       | 0.116      |
-| cosine_precision@10 | 0.072       | 0.066      |
-| cosine_recall@1     | 0.3         | 0.34       |
-| cosine_recall@3     | 0.62        | 0.47       |
-| cosine_recall@5     | 0.62        | 0.52       |
-| cosine_recall@10    | 0.72        | 0.59       |
-| **cosine_ndcg@10**  | **0.5181**  | **0.4737** |
-| cosine_mrr@10       | 0.4529      | 0.4448     |
-| cosine_map@100      | 0.4666      | 0.4476     |
+| cosine_accuracy@1   | 0.32        | 0.38       |
+| cosine_accuracy@3   | 0.52        | 0.52       |
+| cosine_accuracy@5   | 0.62        | 0.58       |
+| cosine_accuracy@10  | 0.72        | 0.7        |
+| cosine_precision@1  | 0.32        | 0.38       |
+| cosine_precision@3  | 0.1733      | 0.1867     |
+| cosine_precision@5  | 0.124       | 0.128      |
+| cosine_precision@10 | 0.072       | 0.076      |
+| cosine_recall@1     | 0.32        | 0.36       |
+| cosine_recall@3     | 0.52        | 0.5        |
+| cosine_recall@5     | 0.62        | 0.57       |
+| cosine_recall@10    | 0.72        | 0.67       |
+| **cosine_ndcg@10**  | **0.5182**  | **0.5135** |
+| cosine_mrr@10       | 0.454       | 0.4714     |
+| cosine_map@100      | 0.4681      | 0.4693     |
 
 #### Nano BEIR
 
@@ -414,21 +414,21 @@ You can finetune this model on your own dataset.
 
 | Metric              | Value      |
 |:--------------------|:-----------|
-| cosine_accuracy@1   | 0.33       |
-| cosine_accuracy@3   | 0.56       |
-| cosine_accuracy@5   | 0.58       |
-| cosine_accuracy@10  | 0.66       |
-| cosine_precision@1  | 0.33       |
-| cosine_precision@3  | 0.19       |
-| cosine_precision@5  | 0.12       |
-| cosine_precision@10 | 0.069      |
-| cosine_recall@1     | 0.32       |
-| cosine_recall@3     | 0.545      |
-| cosine_recall@5     | 0.57       |
-| cosine_recall@10    | 0.655      |
-| **cosine_ndcg@10**  | **0.4959** |
-| cosine_mrr@10       | 0.4489     |
-| cosine_map@100      | 0.4571     |
+| cosine_accuracy@1   | 0.35       |
+| cosine_accuracy@3   | 0.52       |
+| cosine_accuracy@5   | 0.6        |
+| cosine_accuracy@10  | 0.71       |
+| cosine_precision@1  | 0.35       |
+| cosine_precision@3  | 0.18       |
+| cosine_precision@5  | 0.126      |
+| cosine_precision@10 | 0.074      |
+| cosine_recall@1     | 0.34       |
+| cosine_recall@3     | 0.51       |
+| cosine_recall@5     | 0.595      |
+| cosine_recall@10    | 0.695      |
+| **cosine_ndcg@10**  | **0.5159** |
+| cosine_mrr@10       | 0.4627     |
+| cosine_map@100      | 0.4687     |
 
 <!--
 ## Bias, Risks and Limitations
@@ -502,9 +502,9 @@ You can finetune this model on your own dataset.
 - `eval_strategy`: steps
 - `per_device_train_batch_size`: 128
 - `per_device_eval_batch_size`: 128
-- `learning_rate`: 0.0001
+- `learning_rate`: 8e-05
 - `weight_decay`: 0.005
-- `max_steps`: 562
+- `max_steps`: 500
 - `warmup_ratio`: 0.1
 - `fp16`: True
 - `dataloader_drop_last`: True
@@ -531,14 +531,14 @@ You can finetune this model on your own dataset.
 - `gradient_accumulation_steps`: 1
 - `eval_accumulation_steps`: None
 - `torch_empty_cache_steps`: None
-- `learning_rate`: 0.0001
+- `learning_rate`: 8e-05
 - `weight_decay`: 0.005
 - `adam_beta1`: 0.9
 - `adam_beta2`: 0.999
 - `adam_epsilon`: 1e-08
 - `max_grad_norm`: 1.0
 - `num_train_epochs`: 3.0
-- `max_steps`: 562
+- `max_steps`: 500
 - `lr_scheduler_type`: linear
 - `lr_scheduler_kwargs`: {}
 - `warmup_ratio`: 0.1
@@ -643,12 +643,13 @@ You can finetune this model on your own dataset.
 </details>
 
 ### Training Logs
-| Epoch  | Step | Training Loss | Validation Loss | NanoMSMARCO_cosine_ndcg@10 | NanoNQ_cosine_ndcg@10 | NanoBEIR_mean_cosine_ndcg@10 |
-|:------:|:----:|:-------------:|:---------------:|:--------------------------:|:---------------------:|:----------------------------:|
-| 0      | 0    | -             | 1.2542          | 0.5540                     | 0.5931                | 0.5735                       |
-| 0.3556 | 250  | 1.138         | 0.9359          | 0.5039                     | 0.4586                | 0.4813                       |
-| 0.7112 | 500  | 1.0665        | 0.9102          | 0.5181                     | 0.4737                | 0.4959                       |
+| Epoch      | Step    | Training Loss | Validation Loss | NanoMSMARCO_cosine_ndcg@10 | NanoNQ_cosine_ndcg@10 | NanoBEIR_mean_cosine_ndcg@10 |
+|:----------:|:-------:|:-------------:|:---------------:|:--------------------------:|:---------------------:|:----------------------------:|
+| 0          | 0       | -             | 1.2073          | 0.5887                     | 0.5786                | 0.5836                       |
+| 0.3556     | 250     | 1.0774        | 0.8959          | 0.5097                     | 0.5021                | 0.5059                       |
+| **0.7112** | **500** | **1.0113**    | **0.8709**      | **0.5182**                 | **0.5135**            | **0.5159**                   |
 
+* The bold row denotes the saved checkpoint.
 
 ### Framework Versions
 - Python: 3.10.18