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:111470
 - loss:MultipleNegativesRankingLoss
-base_model: sentence-transformers/all-MiniLM-L12-v2
+base_model: Alibaba-NLP/gte-modernbert-base
 widget:
 - source_sentence: why are some rocks radioactive
   sentences:
@@ -106,7 +106,7 @@ 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
@@ -116,49 +116,49 @@ model-index:
       type: NanoMSMARCO
     metrics:
     - type: cosine_accuracy@1
-      value: 0.34
+      value: 0.46
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.54
+      value: 0.64
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
-      value: 0.64
+      value: 0.78
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.72
+      value: 0.84
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.34
+      value: 0.46
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.18
+      value: 0.21333333333333332
       name: Cosine Precision@3
     - type: cosine_precision@5
-      value: 0.12800000000000003
+      value: 0.156
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.07200000000000001
+      value: 0.08399999999999999
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.34
+      value: 0.46
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.54
+      value: 0.64
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.64
+      value: 0.78
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.72
+      value: 0.84
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.5240037339745247
+      value: 0.6436072717719088
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.4620555555555555
+      value: 0.5809365079365079
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.47561895719519554
+      value: 0.5875442422678194
       name: Cosine Map@100
   - task:
       type: information-retrieval
@@ -171,46 +171,46 @@ model-index:
       value: 0.42
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.6
+      value: 0.58
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
-      value: 0.6
+      value: 0.66
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.68
+      value: 0.76
       name: Cosine Accuracy@10
     - type: cosine_precision@1
       value: 0.42
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.21333333333333332
+      value: 0.19333333333333333
       name: Cosine Precision@3
     - type: cosine_precision@5
-      value: 0.128
+      value: 0.136
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.07200000000000001
+      value: 0.08399999999999999
       name: Cosine Precision@10
     - type: cosine_recall@1
       value: 0.38
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.58
+      value: 0.54
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.58
+      value: 0.62
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.65
+      value: 0.74
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.5323118123166091
+      value: 0.5512960895859668
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.5109126984126984
+      value: 0.5124126984126984
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.49965941289619986
+      value: 0.4936325526558866
       name: Cosine Map@100
   - task:
       type: nano-beir
@@ -220,63 +220,63 @@ model-index:
       type: NanoBEIR_mean
     metrics:
     - type: cosine_accuracy@1
-      value: 0.38
+      value: 0.44
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.5700000000000001
+      value: 0.61
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
-      value: 0.62
+      value: 0.72
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.7
+      value: 0.8
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.38
+      value: 0.44
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.19666666666666666
+      value: 0.2033333333333333
       name: Cosine Precision@3
     - type: cosine_precision@5
-      value: 0.128
+      value: 0.14600000000000002
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.07200000000000001
+      value: 0.08399999999999999
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.36
+      value: 0.42000000000000004
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.56
+      value: 0.5900000000000001
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.61
+      value: 0.7
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.685
+      value: 0.79
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.5281577731455669
+      value: 0.5974516806789378
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.4864841269841269
+      value: 0.5466746031746031
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.4876391850456977
+      value: 0.540588397461853
       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 -->
@@ -292,9 +292,8 @@ 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})
 )
 ```
 
@@ -322,14 +321,14 @@ 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([[1.0000, 1.0000, 0.9589],
-#         [1.0000, 1.0000, 0.9589],
-#         [0.9589, 0.9589, 1.0001]])
+# tensor([[0.9999, 0.9999, 0.1540],
+#         [0.9999, 0.9999, 0.1540],
+#         [0.1540, 0.1540, 1.0000]])
 ```
 
 <!--
@@ -367,21 +366,21 @@ You can finetune this model on your own dataset.
 
 | Metric              | NanoMSMARCO | NanoNQ     |
 |:--------------------|:------------|:-----------|
-| cosine_accuracy@1   | 0.34        | 0.42       |
-| cosine_accuracy@3   | 0.54        | 0.6        |
-| cosine_accuracy@5   | 0.64        | 0.6        |
-| cosine_accuracy@10  | 0.72        | 0.68       |
-| cosine_precision@1  | 0.34        | 0.42       |
-| cosine_precision@3  | 0.18        | 0.2133     |
-| cosine_precision@5  | 0.128       | 0.128      |
-| cosine_precision@10 | 0.072       | 0.072      |
-| cosine_recall@1     | 0.34        | 0.38       |
-| cosine_recall@3     | 0.54        | 0.58       |
-| cosine_recall@5     | 0.64        | 0.58       |
-| cosine_recall@10    | 0.72        | 0.65       |
-| **cosine_ndcg@10**  | **0.524**   | **0.5323** |
-| cosine_mrr@10       | 0.4621      | 0.5109     |
-| cosine_map@100      | 0.4756      | 0.4997     |
+| cosine_accuracy@1   | 0.46        | 0.42       |
+| cosine_accuracy@3   | 0.64        | 0.58       |
+| cosine_accuracy@5   | 0.78        | 0.66       |
+| cosine_accuracy@10  | 0.84        | 0.76       |
+| cosine_precision@1  | 0.46        | 0.42       |
+| cosine_precision@3  | 0.2133      | 0.1933     |
+| cosine_precision@5  | 0.156       | 0.136      |
+| cosine_precision@10 | 0.084       | 0.084      |
+| cosine_recall@1     | 0.46        | 0.38       |
+| cosine_recall@3     | 0.64        | 0.54       |
+| cosine_recall@5     | 0.78        | 0.62       |
+| cosine_recall@10    | 0.84        | 0.74       |
+| **cosine_ndcg@10**  | **0.6436**  | **0.5513** |
+| cosine_mrr@10       | 0.5809      | 0.5124     |
+| cosine_map@100      | 0.5875      | 0.4936     |
 
 #### Nano BEIR
 
@@ -399,21 +398,21 @@ You can finetune this model on your own dataset.
 
 | Metric              | Value      |
 |:--------------------|:-----------|
-| cosine_accuracy@1   | 0.38       |
-| cosine_accuracy@3   | 0.57       |
-| cosine_accuracy@5   | 0.62       |
-| cosine_accuracy@10  | 0.7        |
-| cosine_precision@1  | 0.38       |
-| cosine_precision@3  | 0.1967     |
-| cosine_precision@5  | 0.128      |
-| cosine_precision@10 | 0.072      |
-| cosine_recall@1     | 0.36       |
-| cosine_recall@3     | 0.56       |
-| cosine_recall@5     | 0.61       |
-| cosine_recall@10    | 0.685      |
-| **cosine_ndcg@10**  | **0.5282** |
-| cosine_mrr@10       | 0.4865     |
-| cosine_map@100      | 0.4876     |
+| cosine_accuracy@1   | 0.44       |
+| cosine_accuracy@3   | 0.61       |
+| cosine_accuracy@5   | 0.72       |
+| cosine_accuracy@10  | 0.8        |
+| cosine_precision@1  | 0.44       |
+| cosine_precision@3  | 0.2033     |
+| cosine_precision@5  | 0.146      |
+| cosine_precision@10 | 0.084      |
+| cosine_recall@1     | 0.42       |
+| cosine_recall@3     | 0.59       |
+| cosine_recall@5     | 0.7        |
+| cosine_recall@10    | 0.79       |
+| **cosine_ndcg@10**  | **0.5975** |
+| cosine_mrr@10       | 0.5467     |
+| cosine_map@100      | 0.5406     |
 
 <!--
 ## Bias, Risks and Limitations
@@ -439,7 +438,7 @@ You can finetune this model on your own dataset.
   |         | anchor                                                                            | positive                                                                           | negative                                                                           |
   |:--------|:----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|
   | type    | string                                                                            | string                                                                             | string                                                                             |
-  | details | <ul><li>min: 4 tokens</li><li>mean: 10.95 tokens</li><li>max: 60 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 67.57 tokens</li><li>max: 128 tokens</li></ul> | <ul><li>min: 7 tokens</li><li>mean: 66.64 tokens</li><li>max: 128 tokens</li></ul> |
+  | details | <ul><li>min: 4 tokens</li><li>mean: 11.17 tokens</li><li>max: 59 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 68.53 tokens</li><li>max: 128 tokens</li></ul> | <ul><li>min: 7 tokens</li><li>mean: 67.56 tokens</li><li>max: 128 tokens</li></ul> |
 * Samples:
   | anchor                                                                | positive                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | negative                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
   |:----------------------------------------------------------------------|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
@@ -465,7 +464,7 @@ You can finetune this model on your own dataset.
   |         | anchor                                                                            | positive                                                                           | negative                                                                           |
   |:--------|:----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|
   | type    | string                                                                            | string                                                                             | string                                                                             |
-  | details | <ul><li>min: 4 tokens</li><li>mean: 11.11 tokens</li><li>max: 66 tokens</li></ul> | <ul><li>min: 7 tokens</li><li>mean: 67.99 tokens</li><li>max: 128 tokens</li></ul> | <ul><li>min: 7 tokens</li><li>mean: 66.08 tokens</li><li>max: 128 tokens</li></ul> |
+  | details | <ul><li>min: 4 tokens</li><li>mean: 11.35 tokens</li><li>max: 64 tokens</li></ul> | <ul><li>min: 7 tokens</li><li>mean: 68.67 tokens</li><li>max: 128 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 67.03 tokens</li><li>max: 128 tokens</li></ul> |
 * Samples:
   | anchor                                                                                                    | positive                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | negative                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
   |:----------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
@@ -487,9 +486,9 @@ 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`: 8e-05
+- `learning_rate`: 4e-05
 - `weight_decay`: 0.005
-- `max_steps`: 1687
+- `max_steps`: 500
 - `warmup_ratio`: 0.1
 - `fp16`: True
 - `dataloader_drop_last`: True
@@ -516,14 +515,14 @@ You can finetune this model on your own dataset.
 - `gradient_accumulation_steps`: 1
 - `eval_accumulation_steps`: None
 - `torch_empty_cache_steps`: None
-- `learning_rate`: 8e-05
+- `learning_rate`: 4e-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`: 1687
+- `max_steps`: 500
 - `lr_scheduler_type`: linear
 - `lr_scheduler_kwargs`: {}
 - `warmup_ratio`: 0.1
@@ -628,16 +627,13 @@ 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    | -             | 1.1142          | 0.5887                     | 0.5786                | 0.5836                       |
-| 0.2874 | 250  | 1.0151        | 0.8262          | 0.5547                     | 0.5323                | 0.5435                       |
-| 0.5747 | 500  | 0.9318        | 0.8010          | 0.5094                     | 0.5342                | 0.5218                       |
-| 0.8621 | 750  | 0.9087        | 0.7803          | 0.5242                     | 0.5461                | 0.5352                       |
-| 1.1494 | 1000 | 0.8448        | 0.7753          | 0.5291                     | 0.5333                | 0.5312                       |
-| 1.4368 | 1250 | 0.798         | 0.7693          | 0.5249                     | 0.5104                | 0.5176                       |
-| 1.7241 | 1500 | 0.7883        | 0.7618          | 0.5240                     | 0.5323                | 0.5282                       |
+| Epoch      | Step    | Training Loss | Validation Loss | NanoMSMARCO_cosine_ndcg@10 | NanoNQ_cosine_ndcg@10 | NanoBEIR_mean_cosine_ndcg@10 |
+|:----------:|:-------:|:-------------:|:---------------:|:--------------------------:|:---------------------:|:----------------------------:|
+| 0          | 0       | -             | 2.5772          | 0.6530                     | 0.6552                | 0.6541                       |
+| 0.2874     | 250     | 0.9403        | 0.7617          | 0.6082                     | 0.5623                | 0.5853                       |
+| **0.5747** | **500** | **0.7488**    | **0.743**       | **0.6436**                 | **0.5513**            | **0.5975**                   |
 
+* The bold row denotes the saved checkpoint.
 
 ### Framework Versions
 - Python: 3.10.18

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"
   }
 ]