Add new SentenceTransformer model

1_Pooling/config.json

@@ -1,7 +1,7 @@
 {
-    "word_embedding_dimension": 384,
-    "pooling_mode_cls_token": false,
-    "pooling_mode_mean_tokens": true,
+    "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,

README.md

@@ -7,7 +7,7 @@ tags:
 - generated_from_trainer
 - dataset_size:90000
 - loss:MultipleNegativesRankingLoss
-base_model: sentence-transformers/all-MiniLM-L12-v2
+base_model: Alibaba-NLP/gte-modernbert-base
 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-L12-v2
+- name: SentenceTransformer based on Alibaba-NLP/gte-modernbert-base
   results:
   - task:
       type: information-retrieval
@@ -131,49 +131,49 @@ model-index:
       type: NanoMSMARCO
     metrics:
     - type: cosine_accuracy@1
-      value: 0.36
+      value: 0.44
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.58
+      value: 0.66
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
-      value: 0.64
+      value: 0.72
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.76
+      value: 0.84
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.36
+      value: 0.44
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.19333333333333333
+      value: 0.22
       name: Cosine Precision@3
     - type: cosine_precision@5
-      value: 0.128
+      value: 0.14400000000000002
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.07600000000000001
+      value: 0.08399999999999999
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.36
+      value: 0.44
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.58
+      value: 0.66
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.64
+      value: 0.72
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.76
+      value: 0.84
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.5502773798420649
+      value: 0.6329590083197414
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.4841904761904761
+      value: 0.5681349206349207
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.49554545654198856
+      value: 0.5748098465861624
       name: Cosine Map@100
   - task:
       type: information-retrieval
@@ -183,49 +183,49 @@ model-index:
       type: NanoNQ
     metrics:
     - type: cosine_accuracy@1
-      value: 0.38
+      value: 0.32
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.56
+      value: 0.54
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
-      value: 0.6
+      value: 0.64
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.66
+      value: 0.74
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.38
+      value: 0.32
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.19333333333333333
+      value: 0.18666666666666665
       name: Cosine Precision@3
     - type: cosine_precision@5
-      value: 0.128
+      value: 0.132
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.07
+      value: 0.08
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.37
+      value: 0.3
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.53
+      value: 0.52
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.58
+      value: 0.6
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.63
+      value: 0.71
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.50866692066392
+      value: 0.49769715371181095
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.4758571428571428
+      value: 0.43952380952380954
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.47823183905498623
+      value: 0.43617752668168885
       name: Cosine Map@100
   - task:
       type: nano-beir
@@ -235,63 +235,63 @@ model-index:
       type: NanoBEIR_mean
     metrics:
     - type: cosine_accuracy@1
-      value: 0.37
+      value: 0.38
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.5700000000000001
+      value: 0.6000000000000001
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
-      value: 0.62
+      value: 0.6799999999999999
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.71
+      value: 0.79
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.37
+      value: 0.38
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.19333333333333333
+      value: 0.2033333333333333
       name: Cosine Precision@3
     - type: cosine_precision@5
-      value: 0.128
+      value: 0.138
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.07300000000000001
+      value: 0.08199999999999999
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.365
+      value: 0.37
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.5549999999999999
+      value: 0.5900000000000001
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.61
+      value: 0.6599999999999999
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.6950000000000001
+      value: 0.7749999999999999
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.5294721502529924
+      value: 0.5653280810157761
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.48002380952380946
+      value: 0.503829365079365
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.4868886477984874
+      value: 0.5054936866339257
       name: Cosine Map@100
 ---
 
-# SentenceTransformer based on sentence-transformers/all-MiniLM-L12-v2
+# SentenceTransformer based on Alibaba-NLP/gte-modernbert-base
 
-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.
+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.
 
 ## Model Details
 
 ### Model Description
 - **Model Type:** Sentence Transformer
-- **Base model:** [sentence-transformers/all-MiniLM-L12-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L12-v2) <!-- at revision 936af83a2ecce5fe87a09109ff5cbcefe073173a -->
+- **Base model:** [Alibaba-NLP/gte-modernbert-base](https://huggingface.co/Alibaba-NLP/gte-modernbert-base) <!-- at revision e7f32e3c00f91d699e8c43b53106206bcc72bb22 -->
 - **Maximum Sequence Length:** 128 tokens
-- **Output Dimensionality:** 384 dimensions
+- **Output Dimensionality:** 768 dimensions
 - **Similarity Function:** Cosine Similarity
 <!-- - **Training Dataset:** Unknown -->
 <!-- - **Language:** Unknown -->
@@ -307,9 +307,8 @@ This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [s
 
 ```
 SentenceTransformer(
-  (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()
+  (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})
 )
 ```
 
@@ -337,14 +336,14 @@ sentences = [
 ]
 embeddings = model.encode(sentences)
 print(embeddings.shape)
-# [3, 384]
+# [3, 768]
 
 # Get the similarity scores for the embeddings
 similarities = model.similarity(embeddings, embeddings)
 print(similarities)
-# tensor([[1.0001, 0.5920, 0.3852],
-#         [0.5920, 1.0000, 0.0748],
-#         [0.3852, 0.0748, 1.0001]])
+# tensor([[0.9999, 0.3909, 0.5755],
+#         [0.3909, 0.9999, 0.1633],
+#         [0.5755, 0.1633, 0.9999]])
 ```
 
 <!--
@@ -382,21 +381,21 @@ You can finetune this model on your own dataset.
 
 | Metric              | NanoMSMARCO | NanoNQ     |
 |:--------------------|:------------|:-----------|
-| cosine_accuracy@1   | 0.36        | 0.38       |
-| cosine_accuracy@3   | 0.58        | 0.56       |
-| cosine_accuracy@5   | 0.64        | 0.6        |
-| cosine_accuracy@10  | 0.76        | 0.66       |
-| cosine_precision@1  | 0.36        | 0.38       |
-| cosine_precision@3  | 0.1933      | 0.1933     |
-| cosine_precision@5  | 0.128       | 0.128      |
-| cosine_precision@10 | 0.076       | 0.07       |
-| cosine_recall@1     | 0.36        | 0.37       |
-| cosine_recall@3     | 0.58        | 0.53       |
-| cosine_recall@5     | 0.64        | 0.58       |
-| cosine_recall@10    | 0.76        | 0.63       |
-| **cosine_ndcg@10**  | **0.5503**  | **0.5087** |
-| cosine_mrr@10       | 0.4842      | 0.4759     |
-| cosine_map@100      | 0.4955      | 0.4782     |
+| 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     |
 
 #### Nano BEIR
 
@@ -414,21 +413,21 @@ You can finetune this model on your own dataset.
 
 | Metric              | Value      |
 |:--------------------|:-----------|
-| cosine_accuracy@1   | 0.37       |
-| cosine_accuracy@3   | 0.57       |
-| cosine_accuracy@5   | 0.62       |
-| cosine_accuracy@10  | 0.71       |
-| cosine_precision@1  | 0.37       |
-| cosine_precision@3  | 0.1933     |
-| cosine_precision@5  | 0.128      |
-| cosine_precision@10 | 0.073      |
-| cosine_recall@1     | 0.365      |
-| cosine_recall@3     | 0.555      |
-| cosine_recall@5     | 0.61       |
-| cosine_recall@10    | 0.695      |
-| **cosine_ndcg@10**  | **0.5295** |
-| cosine_mrr@10       | 0.48       |
-| cosine_map@100      | 0.4869     |
+| 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_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     |
 
 <!--
 ## Bias, Risks and Limitations
@@ -451,10 +450,10 @@ 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: 9.18 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> |
+  |         | 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> |
 * Samples:
   | anchor                                            | positive                                                                                                                                                                                                                                                                                                                                                                          | negative                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
   |:--------------------------------------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
@@ -477,10 +476,10 @@ 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: 9.18 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> |
+  |         | 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> |
 * Samples:
   | anchor                                              | positive                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  | negative                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
   |:----------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
@@ -645,19 +644,19 @@ 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    | -             | 1.2073          | 0.5887                     | 0.5786                | 0.5836                       |
-| 0.3556 | 250  | 1.19          | 0.9200          | 0.5466                     | 0.5332                | 0.5399                       |
-| 0.7112 | 500  | 1.0578        | 0.8943          | 0.5396                     | 0.5252                | 0.5324                       |
-| 1.0669 | 750  | 1.0352        | 0.8849          | 0.5497                     | 0.5252                | 0.5375                       |
-| 1.4225 | 1000 | 1.002         | 0.8761          | 0.5484                     | 0.5308                | 0.5396                       |
-| 1.7781 | 1250 | 0.9953        | 0.8732          | 0.5336                     | 0.5213                | 0.5274                       |
-| 2.1337 | 1500 | 0.9828        | 0.8686          | 0.5340                     | 0.5126                | 0.5233                       |
-| 2.4893 | 1750 | 0.965         | 0.8675          | 0.5417                     | 0.5094                | 0.5256                       |
-| 2.8450 | 2000 | 0.9651        | 0.8658          | 0.5467                     | 0.4994                | 0.5230                       |
-| 3.2006 | 2250 | 0.9522        | 0.8650          | 0.5295                     | 0.5097                | 0.5196                       |
-| 3.5562 | 2500 | 0.9521        | 0.8635          | 0.5446                     | 0.5124                | 0.5285                       |
-| 3.9118 | 2750 | 0.9444        | 0.8635          | 0.5529                     | 0.5070                | 0.5299                       |
-| 4.2674 | 3000 | 0.9397        | 0.8632          | 0.5503                     | 0.5087                | 0.5295                       |
+| 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                       |
 
 
 ### Framework Versions

config_sentence_transformers.json

@@ -4,11 +4,11 @@
     "transformers": "4.57.3",
     "pytorch": "2.9.1+cu128"
   },
-  "model_type": "SentenceTransformer",
   "prompts": {
     "query": "",
     "document": ""
   },
   "default_prompt_name": null,
-  "similarity_fn_name": "cosine"
+  "similarity_fn_name": "cosine",
+  "model_type": "SentenceTransformer"
 }

modules.json

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