Add new SentenceTransformer model

1_Pooling/config.json

@@ -1,7 +1,7 @@
 {
-    "word_embedding_dimension": 768,
-    "pooling_mode_cls_token": true,
-    "pooling_mode_mean_tokens": false,
+    "word_embedding_dimension": 384,
+    "pooling_mode_cls_token": false,
+    "pooling_mode_mean_tokens": true,
     "pooling_mode_max_tokens": false,
     "pooling_mode_mean_sqrt_len_tokens": false,
     "pooling_mode_weightedmean_tokens": false,

README.md

@@ -7,7 +7,7 @@ tags:
 - generated_from_trainer
 - dataset_size:90000
 - loss:MultipleNegativesRankingLoss
-base_model: Alibaba-NLP/gte-modernbert-base
+base_model: thenlper/gte-small
 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 Alibaba-NLP/gte-modernbert-base
+- name: SentenceTransformer based on thenlper/gte-small
   results:
   - task:
       type: information-retrieval
@@ -131,49 +131,49 @@ model-index:
       type: NanoMSMARCO
     metrics:
     - type: cosine_accuracy@1
-      value: 0.44
+      value: 0.38
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.66
+      value: 0.56
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
-      value: 0.72
+      value: 0.68
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.84
+      value: 0.72
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.44
+      value: 0.38
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.22
+      value: 0.18666666666666668
       name: Cosine Precision@3
     - type: cosine_precision@5
-      value: 0.14400000000000002
+      value: 0.136
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.08399999999999999
+      value: 0.07200000000000001
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.44
+      value: 0.38
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.66
+      value: 0.56
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.72
+      value: 0.68
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.84
+      value: 0.72
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.6329590083197414
+      value: 0.5560017600077842
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.5681349206349207
+      value: 0.5023333333333333
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.5748098465861624
+      value: 0.5147876135794199
       name: Cosine Map@100
   - task:
       type: information-retrieval
@@ -183,49 +183,49 @@ model-index:
       type: NanoNQ
     metrics:
     - type: cosine_accuracy@1
-      value: 0.32
+      value: 0.4
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.54
+      value: 0.56
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
       value: 0.64
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.74
+      value: 0.76
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.32
+      value: 0.4
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.18666666666666665
+      value: 0.19333333333333333
       name: Cosine Precision@3
     - type: cosine_precision@5
-      value: 0.132
+      value: 0.14
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.08
+      value: 0.08199999999999999
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.3
+      value: 0.37
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.52
+      value: 0.54
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.6
+      value: 0.63
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.71
+      value: 0.74
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.49769715371181095
+      value: 0.5529706231693197
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.43952380952380954
+      value: 0.505468253968254
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.43617752668168885
+      value: 0.4971957129558554
       name: Cosine Map@100
   - task:
       type: nano-beir
@@ -235,63 +235,63 @@ model-index:
       type: NanoBEIR_mean
     metrics:
     - type: cosine_accuracy@1
-      value: 0.38
+      value: 0.39
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.6000000000000001
+      value: 0.56
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
-      value: 0.6799999999999999
+      value: 0.66
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.79
+      value: 0.74
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.38
+      value: 0.39
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.2033333333333333
+      value: 0.19
       name: Cosine Precision@3
     - type: cosine_precision@5
       value: 0.138
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.08199999999999999
+      value: 0.077
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.37
+      value: 0.375
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.5900000000000001
+      value: 0.55
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.6599999999999999
+      value: 0.655
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.7749999999999999
+      value: 0.73
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.5653280810157761
+      value: 0.554486191588552
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.503829365079365
+      value: 0.5039007936507937
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.5054936866339257
+      value: 0.5059916632676377
       name: Cosine Map@100
 ---
 
-# SentenceTransformer based on Alibaba-NLP/gte-modernbert-base
+# SentenceTransformer based on thenlper/gte-small
 
-This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Alibaba-NLP/gte-modernbert-base](https://huggingface.co/Alibaba-NLP/gte-modernbert-base). It maps sentences & paragraphs to a 768-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 [thenlper/gte-small](https://huggingface.co/thenlper/gte-small). 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:** [Alibaba-NLP/gte-modernbert-base](https://huggingface.co/Alibaba-NLP/gte-modernbert-base) <!-- at revision e7f32e3c00f91d699e8c43b53106206bcc72bb22 -->
+- **Base model:** [thenlper/gte-small](https://huggingface.co/thenlper/gte-small) <!-- at revision 17e1f347d17fe144873b1201da91788898c639cd -->
 - **Maximum Sequence Length:** 128 tokens
-- **Output Dimensionality:** 768 dimensions
+- **Output Dimensionality:** 384 dimensions
 - **Similarity Function:** Cosine Similarity
 <!-- - **Training Dataset:** Unknown -->
 <!-- - **Language:** Unknown -->
@@ -307,8 +307,9 @@ This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [A
 
 ```
 SentenceTransformer(
-  (0): Transformer({'max_seq_length': 128, 'do_lower_case': False, 'architecture': 'ModernBertModel'})
-  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
+  (0): Transformer({'max_seq_length': 128, 'do_lower_case': False, 'architecture': 'BertModel'})
+  (1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
+  (2): Normalize()
 )
 ```
 
@@ -336,14 +337,14 @@ sentences = [
 ]
 embeddings = model.encode(sentences)
 print(embeddings.shape)
-# [3, 768]
+# [3, 384]
 
 # Get the similarity scores for the embeddings
 similarities = model.similarity(embeddings, embeddings)
 print(similarities)
-# tensor([[0.9999, 0.3909, 0.5755],
-#         [0.3909, 0.9999, 0.1633],
-#         [0.5755, 0.1633, 0.9999]])
+# tensor([[1.0000, 0.7010, 0.4682],
+#         [0.7010, 1.0000, 0.2656],
+#         [0.4682, 0.2656, 1.0001]])
 ```
 
 <!--
@@ -379,23 +380,23 @@ You can finetune this model on your own dataset.
 * Datasets: `NanoMSMARCO` and `NanoNQ`
 * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
 
-| Metric              | NanoMSMARCO | NanoNQ     |
-|:--------------------|:------------|:-----------|
-| cosine_accuracy@1   | 0.44        | 0.32       |
-| cosine_accuracy@3   | 0.66        | 0.54       |
-| cosine_accuracy@5   | 0.72        | 0.64       |
-| cosine_accuracy@10  | 0.84        | 0.74       |
-| cosine_precision@1  | 0.44        | 0.32       |
-| cosine_precision@3  | 0.22        | 0.1867     |
-| cosine_precision@5  | 0.144       | 0.132      |
-| cosine_precision@10 | 0.084       | 0.08       |
-| cosine_recall@1     | 0.44        | 0.3        |
-| cosine_recall@3     | 0.66        | 0.52       |
-| cosine_recall@5     | 0.72        | 0.6        |
-| cosine_recall@10    | 0.84        | 0.71       |
-| **cosine_ndcg@10**  | **0.633**   | **0.4977** |
-| cosine_mrr@10       | 0.5681      | 0.4395     |
-| cosine_map@100      | 0.5748      | 0.4362     |
+| Metric              | NanoMSMARCO | NanoNQ    |
+|:--------------------|:------------|:----------|
+| cosine_accuracy@1   | 0.38        | 0.4       |
+| cosine_accuracy@3   | 0.56        | 0.56      |
+| cosine_accuracy@5   | 0.68        | 0.64      |
+| cosine_accuracy@10  | 0.72        | 0.76      |
+| cosine_precision@1  | 0.38        | 0.4       |
+| cosine_precision@3  | 0.1867      | 0.1933    |
+| cosine_precision@5  | 0.136       | 0.14      |
+| cosine_precision@10 | 0.072       | 0.082     |
+| cosine_recall@1     | 0.38        | 0.37      |
+| cosine_recall@3     | 0.56        | 0.54      |
+| cosine_recall@5     | 0.68        | 0.63      |
+| cosine_recall@10    | 0.72        | 0.74      |
+| **cosine_ndcg@10**  | **0.556**   | **0.553** |
+| cosine_mrr@10       | 0.5023      | 0.5055    |
+| cosine_map@100      | 0.5148      | 0.4972    |
 
 #### Nano BEIR
 
@@ -413,21 +414,21 @@ You can finetune this model on your own dataset.
 
 | Metric              | Value      |
 |:--------------------|:-----------|
-| cosine_accuracy@1   | 0.38       |
-| cosine_accuracy@3   | 0.6        |
-| cosine_accuracy@5   | 0.68       |
-| cosine_accuracy@10  | 0.79       |
-| cosine_precision@1  | 0.38       |
-| cosine_precision@3  | 0.2033     |
+| cosine_accuracy@1   | 0.39       |
+| cosine_accuracy@3   | 0.56       |
+| cosine_accuracy@5   | 0.66       |
+| cosine_accuracy@10  | 0.74       |
+| cosine_precision@1  | 0.39       |
+| cosine_precision@3  | 0.19       |
 | cosine_precision@5  | 0.138      |
-| cosine_precision@10 | 0.082      |
-| cosine_recall@1     | 0.37       |
-| cosine_recall@3     | 0.59       |
-| cosine_recall@5     | 0.66       |
-| cosine_recall@10    | 0.775      |
-| **cosine_ndcg@10**  | **0.5653** |
-| cosine_mrr@10       | 0.5038     |
-| cosine_map@100      | 0.5055     |
+| cosine_precision@10 | 0.077      |
+| cosine_recall@1     | 0.375      |
+| cosine_recall@3     | 0.55       |
+| cosine_recall@5     | 0.655      |
+| cosine_recall@10    | 0.73       |
+| **cosine_ndcg@10**  | **0.5545** |
+| cosine_mrr@10       | 0.5039     |
+| cosine_map@100      | 0.506      |
 
 <!--
 ## Bias, Risks and Limitations
@@ -450,10 +451,10 @@ You can finetune this model on your own dataset.
 * Size: 90,000 training samples
 * Columns: <code>anchor</code>, <code>positive</code>, and <code>negative</code>
 * Approximate statistics based on the first 1000 samples:
-  |         | anchor                                                                          | positive                                                                            | negative                                                                            |
-  |:--------|:--------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|
-  | type    | string                                                                          | string                                                                              | string                                                                              |
-  | details | <ul><li>min: 4 tokens</li><li>mean: 9.3 tokens</li><li>max: 27 tokens</li></ul> | <ul><li>min: 20 tokens</li><li>mean: 79.39 tokens</li><li>max: 128 tokens</li></ul> | <ul><li>min: 17 tokens</li><li>mean: 78.77 tokens</li><li>max: 128 tokens</li></ul> |
+  |         | anchor                                                                           | positive                                                                            | negative                                                                            |
+  |:--------|:---------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|
+  | type    | string                                                                           | string                                                                              | string                                                                              |
+  | details | <ul><li>min: 4 tokens</li><li>mean: 9.18 tokens</li><li>max: 43 tokens</li></ul> | <ul><li>min: 19 tokens</li><li>mean: 78.75 tokens</li><li>max: 128 tokens</li></ul> | <ul><li>min: 17 tokens</li><li>mean: 77.97 tokens</li><li>max: 128 tokens</li></ul> |
 * Samples:
   | anchor                                            | positive                                                                                                                                                                                                                                                                                                                                                                          | negative                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
   |:--------------------------------------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
@@ -476,10 +477,10 @@ You can finetune this model on your own dataset.
 * Size: 10,000 evaluation samples
 * Columns: <code>anchor</code>, <code>positive</code>, and <code>negative</code>
 * Approximate statistics based on the first 1000 samples:
-  |         | anchor                                                                           | positive                                                                           | negative                                                                            |
-  |:--------|:---------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|
-  | type    | string                                                                           | string                                                                             | string                                                                              |
-  | details | <ul><li>min: 4 tokens</li><li>mean: 9.33 tokens</li><li>max: 21 tokens</li></ul> | <ul><li>min: 17 tokens</li><li>mean: 79.7 tokens</li><li>max: 128 tokens</li></ul> | <ul><li>min: 18 tokens</li><li>mean: 77.72 tokens</li><li>max: 128 tokens</li></ul> |
+  |         | anchor                                                                           | positive                                                                            | negative                                                                            |
+  |:--------|:---------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|
+  | type    | string                                                                           | string                                                                              | string                                                                              |
+  | details | <ul><li>min: 4 tokens</li><li>mean: 9.18 tokens</li><li>max: 25 tokens</li></ul> | <ul><li>min: 16 tokens</li><li>mean: 78.83 tokens</li><li>max: 128 tokens</li></ul> | <ul><li>min: 18 tokens</li><li>mean: 76.86 tokens</li><li>max: 128 tokens</li></ul> |
 * Samples:
   | anchor                                              | positive                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  | negative                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
   |:----------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
@@ -644,19 +645,19 @@ You can finetune this model on your own dataset.
 ### Training Logs
 | Epoch  | Step | Training Loss | Validation Loss | NanoMSMARCO_cosine_ndcg@10 | NanoNQ_cosine_ndcg@10 | NanoBEIR_mean_cosine_ndcg@10 |
 |:------:|:----:|:-------------:|:---------------:|:--------------------------:|:---------------------:|:----------------------------:|
-| 0      | 0    | -             | 2.7123          | 0.6530                     | 0.6552                | 0.6541                       |
-| 0.3556 | 250  | 1.5463        | 0.9474          | 0.6490                     | 0.5678                | 0.6084                       |
-| 0.7112 | 500  | 0.9209        | 0.8966          | 0.6648                     | 0.5657                | 0.6153                       |
-| 1.0669 | 750  | 0.8879        | 0.8806          | 0.6618                     | 0.5579                | 0.6099                       |
-| 1.4225 | 1000 | 0.8374        | 0.8719          | 0.6586                     | 0.5354                | 0.5970                       |
-| 1.7781 | 1250 | 0.831         | 0.8651          | 0.6657                     | 0.5295                | 0.5976                       |
-| 2.1337 | 1500 | 0.8096        | 0.8626          | 0.6443                     | 0.5006                | 0.5725                       |
-| 2.4893 | 1750 | 0.7797        | 0.8602          | 0.6449                     | 0.5081                | 0.5765                       |
-| 2.8450 | 2000 | 0.7782        | 0.8579          | 0.6376                     | 0.5033                | 0.5704                       |
-| 3.2006 | 2250 | 0.7561        | 0.8611          | 0.6428                     | 0.5007                | 0.5718                       |
-| 3.5562 | 2500 | 0.7494        | 0.8598          | 0.6482                     | 0.4921                | 0.5701                       |
-| 3.9118 | 2750 | 0.7436        | 0.8607          | 0.6427                     | 0.4982                | 0.5704                       |
-| 4.2674 | 3000 | 0.733         | 0.8612          | 0.6330                     | 0.4977                | 0.5653                       |
+| 0      | 0    | -             | 4.1735          | 0.6259                     | 0.6583                | 0.6421                       |
+| 0.3556 | 250  | 3.3399        | 1.1023          | 0.5061                     | 0.5036                | 0.5048                       |
+| 0.7112 | 500  | 1.2152        | 0.9648          | 0.5609                     | 0.5241                | 0.5425                       |
+| 1.0669 | 750  | 1.1157        | 0.9385          | 0.5568                     | 0.5444                | 0.5506                       |
+| 1.4225 | 1000 | 1.0721        | 0.9241          | 0.5523                     | 0.5520                | 0.5521                       |
+| 1.7781 | 1250 | 1.0561        | 0.9159          | 0.5677                     | 0.5454                | 0.5565                       |
+| 2.1337 | 1500 | 1.0411        | 0.9116          | 0.5505                     | 0.5527                | 0.5516                       |
+| 2.4893 | 1750 | 1.0248        | 0.9073          | 0.5433                     | 0.5631                | 0.5532                       |
+| 2.8450 | 2000 | 1.022         | 0.9042          | 0.5424                     | 0.5648                | 0.5536                       |
+| 3.2006 | 2250 | 1.0079        | 0.9015          | 0.5557                     | 0.5626                | 0.5592                       |
+| 3.5562 | 2500 | 1.0092        | 0.8998          | 0.5560                     | 0.5590                | 0.5575                       |
+| 3.9118 | 2750 | 0.9982        | 0.8986          | 0.5560                     | 0.5535                | 0.5548                       |
+| 4.2674 | 3000 | 0.9947        | 0.8986          | 0.5560                     | 0.5530                | 0.5545                       |
 
 
 ### Framework Versions

config_sentence_transformers.json

@@ -1,4 +1,5 @@
 {
+  "model_type": "SentenceTransformer",
   "__version__": {
     "sentence_transformers": "5.2.0",
     "transformers": "4.57.3",
@@ -9,6 +10,5 @@
     "document": ""
   },
   "default_prompt_name": null,
-  "similarity_fn_name": "cosine",
-  "model_type": "SentenceTransformer"
+  "similarity_fn_name": "cosine"
 }

modules.json

@@ -10,5 +10,11 @@
     "name": "1",
     "path": "1_Pooling",
     "type": "sentence_transformers.models.Pooling"
+  },
+  {
+    "idx": 2,
+    "name": "2",
+    "path": "2_Normalize",
+    "type": "sentence_transformers.models.Normalize"
   }
 ]