Add new SentenceTransformer model

README.md

@@ -7,7 +7,7 @@ tags:
 - generated_from_trainer
 - dataset_size:111470
 - loss:MultipleNegativesRankingLoss
-base_model: thenlper/gte-small
+base_model: sentence-transformers/all-MiniLM-L6-v2
 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 thenlper/gte-small
+- name: SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2
   results:
   - task:
       type: information-retrieval
@@ -116,49 +116,49 @@ model-index:
       type: NanoMSMARCO
     metrics:
     - type: cosine_accuracy@1
-      value: 0.4
+      value: 0.32
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
       value: 0.5
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
-      value: 0.58
+      value: 0.56
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.68
+      value: 0.7
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.4
+      value: 0.32
       name: Cosine Precision@1
     - type: cosine_precision@3
       value: 0.16666666666666669
       name: Cosine Precision@3
     - type: cosine_precision@5
-      value: 0.11600000000000002
+      value: 0.11200000000000002
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.068
+      value: 0.07
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.4
+      value: 0.32
       name: Cosine Recall@1
     - type: cosine_recall@3
       value: 0.5
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.58
+      value: 0.56
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.68
+      value: 0.7
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.523462400095123
+      value: 0.4962486706422321
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.4752380952380952
+      value: 0.43346031746031743
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.48914996910623887
+      value: 0.44415856354878636
       name: Cosine Map@100
   - task:
       type: information-retrieval
@@ -168,49 +168,49 @@ model-index:
       type: NanoNQ
     metrics:
     - type: cosine_accuracy@1
-      value: 0.48
+      value: 0.16
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.56
+      value: 0.26
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
-      value: 0.64
+      value: 0.32
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.74
+      value: 0.46
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.48
+      value: 0.16
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.18666666666666665
+      value: 0.08666666666666666
       name: Cosine Precision@3
     - type: cosine_precision@5
-      value: 0.132
+      value: 0.068
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.078
+      value: 0.04800000000000001
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.45
+      value: 0.15
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.52
+      value: 0.23
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.61
+      value: 0.3
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.7
+      value: 0.43
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.5662755634296394
+      value: 0.27247558178705156
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.5436031746031746
+      value: 0.23207936507936502
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.5289614062254305
+      value: 0.234397839045408
       name: Cosine Map@100
   - task:
       type: nano-beir
@@ -220,61 +220,61 @@ model-index:
       type: NanoBEIR_mean
     metrics:
     - type: cosine_accuracy@1
-      value: 0.44
+      value: 0.24
       name: Cosine Accuracy@1
     - type: cosine_accuracy@3
-      value: 0.53
+      value: 0.38
       name: Cosine Accuracy@3
     - type: cosine_accuracy@5
-      value: 0.61
+      value: 0.44000000000000006
       name: Cosine Accuracy@5
     - type: cosine_accuracy@10
-      value: 0.71
+      value: 0.58
       name: Cosine Accuracy@10
     - type: cosine_precision@1
-      value: 0.44
+      value: 0.24
       name: Cosine Precision@1
     - type: cosine_precision@3
-      value: 0.17666666666666667
+      value: 0.12666666666666668
       name: Cosine Precision@3
     - type: cosine_precision@5
-      value: 0.12400000000000001
+      value: 0.09000000000000001
       name: Cosine Precision@5
     - type: cosine_precision@10
-      value: 0.07300000000000001
+      value: 0.05900000000000001
       name: Cosine Precision@10
     - type: cosine_recall@1
-      value: 0.42500000000000004
+      value: 0.235
       name: Cosine Recall@1
     - type: cosine_recall@3
-      value: 0.51
+      value: 0.365
       name: Cosine Recall@3
     - type: cosine_recall@5
-      value: 0.595
+      value: 0.43000000000000005
       name: Cosine Recall@5
     - type: cosine_recall@10
-      value: 0.69
+      value: 0.565
       name: Cosine Recall@10
     - type: cosine_ndcg@10
-      value: 0.5448689817623812
+      value: 0.38436212621464183
       name: Cosine Ndcg@10
     - type: cosine_mrr@10
-      value: 0.5094206349206349
+      value: 0.3327698412698412
       name: Cosine Mrr@10
     - type: cosine_map@100
-      value: 0.5090556876658348
+      value: 0.3392782012970972
       name: Cosine Map@100
 ---
 
-# SentenceTransformer based on thenlper/gte-small
+# SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2
 
-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.
+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.
 
 ## Model Details
 
 ### Model Description
 - **Model Type:** Sentence Transformer
-- **Base model:** [thenlper/gte-small](https://huggingface.co/thenlper/gte-small) <!-- at revision 17e1f347d17fe144873b1201da91788898c639cd -->
+- **Base model:** [sentence-transformers/all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2) <!-- at revision c9745ed1d9f207416be6d2e6f8de32d1f16199bf -->
 - **Maximum Sequence Length:** 128 tokens
 - **Output Dimensionality:** 384 dimensions
 - **Similarity Function:** Cosine Similarity
@@ -327,9 +327,9 @@ print(embeddings.shape)
 # Get the similarity scores for the embeddings
 similarities = model.similarity(embeddings, embeddings)
 print(similarities)
-# tensor([[1.0001, 1.0001, 0.9810],
-#         [1.0001, 1.0001, 0.9810],
-#         [0.9810, 0.9810, 1.0000]])
+# tensor([[1.0000, 1.0000, 0.9955],
+#         [1.0000, 1.0000, 0.9955],
+#         [0.9955, 0.9955, 1.0000]])
 ```
 
 <!--
@@ -367,21 +367,21 @@ You can finetune this model on your own dataset.
 
 | Metric              | NanoMSMARCO | NanoNQ     |
 |:--------------------|:------------|:-----------|
-| cosine_accuracy@1   | 0.4         | 0.48       |
-| cosine_accuracy@3   | 0.5         | 0.56       |
-| cosine_accuracy@5   | 0.58        | 0.64       |
-| cosine_accuracy@10  | 0.68        | 0.74       |
-| cosine_precision@1  | 0.4         | 0.48       |
-| cosine_precision@3  | 0.1667      | 0.1867     |
-| cosine_precision@5  | 0.116       | 0.132      |
-| cosine_precision@10 | 0.068       | 0.078      |
-| cosine_recall@1     | 0.4         | 0.45       |
-| cosine_recall@3     | 0.5         | 0.52       |
-| cosine_recall@5     | 0.58        | 0.61       |
-| cosine_recall@10    | 0.68        | 0.7        |
-| **cosine_ndcg@10**  | **0.5235**  | **0.5663** |
-| cosine_mrr@10       | 0.4752      | 0.5436     |
-| cosine_map@100      | 0.4891      | 0.529      |
+| cosine_accuracy@1   | 0.32        | 0.16       |
+| cosine_accuracy@3   | 0.5         | 0.26       |
+| cosine_accuracy@5   | 0.56        | 0.32       |
+| cosine_accuracy@10  | 0.7         | 0.46       |
+| cosine_precision@1  | 0.32        | 0.16       |
+| cosine_precision@3  | 0.1667      | 0.0867     |
+| cosine_precision@5  | 0.112       | 0.068      |
+| cosine_precision@10 | 0.07        | 0.048      |
+| cosine_recall@1     | 0.32        | 0.15       |
+| cosine_recall@3     | 0.5         | 0.23       |
+| cosine_recall@5     | 0.56        | 0.3        |
+| cosine_recall@10    | 0.7         | 0.43       |
+| **cosine_ndcg@10**  | **0.4962**  | **0.2725** |
+| cosine_mrr@10       | 0.4335      | 0.2321     |
+| cosine_map@100      | 0.4442      | 0.2344     |
 
 #### Nano BEIR
 
@@ -399,21 +399,21 @@ You can finetune this model on your own dataset.
 
 | Metric              | Value      |
 |:--------------------|:-----------|
-| cosine_accuracy@1   | 0.44       |
-| cosine_accuracy@3   | 0.53       |
-| cosine_accuracy@5   | 0.61       |
-| cosine_accuracy@10  | 0.71       |
-| cosine_precision@1  | 0.44       |
-| cosine_precision@3  | 0.1767     |
-| cosine_precision@5  | 0.124      |
-| cosine_precision@10 | 0.073      |
-| cosine_recall@1     | 0.425      |
-| cosine_recall@3     | 0.51       |
-| cosine_recall@5     | 0.595      |
-| cosine_recall@10    | 0.69       |
-| **cosine_ndcg@10**  | **0.5449** |
-| cosine_mrr@10       | 0.5094     |
-| cosine_map@100      | 0.5091     |
+| cosine_accuracy@1   | 0.24       |
+| cosine_accuracy@3   | 0.38       |
+| cosine_accuracy@5   | 0.44       |
+| cosine_accuracy@10  | 0.58       |
+| cosine_precision@1  | 0.24       |
+| cosine_precision@3  | 0.1267     |
+| cosine_precision@5  | 0.09       |
+| cosine_precision@10 | 0.059      |
+| cosine_recall@1     | 0.235      |
+| cosine_recall@3     | 0.365      |
+| cosine_recall@5     | 0.43       |
+| cosine_recall@10    | 0.565      |
+| **cosine_ndcg@10**  | **0.3844** |
+| cosine_mrr@10       | 0.3328     |
+| cosine_map@100      | 0.3393     |
 
 <!--
 ## Bias, Risks and Limitations
@@ -449,7 +449,7 @@ You can finetune this model on your own dataset.
 * Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
   ```json
   {
-      "scale": 7.0,
+      "scale": 3.0,
       "similarity_fct": "cos_sim",
       "gather_across_devices": false
   }
@@ -475,7 +475,7 @@ You can finetune this model on your own dataset.
 * Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
   ```json
   {
-      "scale": 7.0,
+      "scale": 3.0,
       "similarity_fct": "cos_sim",
       "gather_across_devices": false
   }
@@ -487,9 +487,9 @@ 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
-- `weight_decay`: 0.005
-- `max_steps`: 3375
+- `learning_rate`: 0.0001
+- `weight_decay`: 0.001
+- `max_steps`: 5062
 - `warmup_ratio`: 0.1
 - `fp16`: True
 - `dataloader_drop_last`: True
@@ -516,14 +516,14 @@ 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
-- `weight_decay`: 0.005
+- `learning_rate`: 0.0001
+- `weight_decay`: 0.001
 - `adam_beta1`: 0.9
 - `adam_beta2`: 0.999
 - `adam_epsilon`: 1e-08
 - `max_grad_norm`: 1.0
 - `num_train_epochs`: 3.0
-- `max_steps`: 3375
+- `max_steps`: 5062
 - `lr_scheduler_type`: linear
 - `lr_scheduler_kwargs`: {}
 - `warmup_ratio`: 0.1
@@ -630,20 +630,27 @@ 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    | -             | 4.0678          | 0.6259                     | 0.6583                | 0.6421                       |
-| 0.2874 | 250  | 1.9612        | 0.8621          | 0.5278                     | 0.5499                | 0.5389                       |
-| 0.5747 | 500  | 0.9657        | 0.8137          | 0.5173                     | 0.5494                | 0.5333                       |
-| 0.8621 | 750  | 0.9272        | 0.7913          | 0.5136                     | 0.5526                | 0.5331                       |
-| 1.1494 | 1000 | 0.8618        | 0.7881          | 0.5281                     | 0.5308                | 0.5295                       |
-| 1.4368 | 1250 | 0.8225        | 0.7816          | 0.5286                     | 0.5265                | 0.5276                       |
-| 1.7241 | 1500 | 0.8128        | 0.7757          | 0.5442                     | 0.5449                | 0.5445                       |
-| 2.0115 | 1750 | 0.8077        | 0.7718          | 0.5665                     | 0.5175                | 0.5420                       |
-| 2.2989 | 2000 | 0.734         | 0.7798          | 0.5271                     | 0.5350                | 0.5311                       |
-| 2.5862 | 2250 | 0.7325        | 0.7758          | 0.5371                     | 0.5666                | 0.5518                       |
-| 2.8736 | 2500 | 0.7245        | 0.7776          | 0.5173                     | 0.5766                | 0.5470                       |
-| 3.1609 | 2750 | 0.7067        | 0.7787          | 0.5209                     | 0.5548                | 0.5379                       |
-| 3.4483 | 3000 | 0.6855        | 0.7807          | 0.5180                     | 0.5611                | 0.5395                       |
-| 3.7356 | 3250 | 0.6831        | 0.7813          | 0.5235                     | 0.5663                | 0.5449                       |
+| 0      | 0    | -             | 3.3212          | 0.5540                     | 0.5931                | 0.5735                       |
+| 0.2874 | 250  | 3.2509        | 3.0429          | 0.4590                     | 0.4189                | 0.4389                       |
+| 0.5747 | 500  | 3.1458        | 3.0222          | 0.4855                     | 0.3752                | 0.4303                       |
+| 0.8621 | 750  | 3.1119        | 3.0053          | 0.4708                     | 0.3715                | 0.4211                       |
+| 1.1494 | 1000 | 3.0646        | 2.9901          | 0.4632                     | 0.3600                | 0.4116                       |
+| 1.4368 | 1250 | 3.0381        | 2.9852          | 0.5014                     | 0.3426                | 0.4220                       |
+| 1.7241 | 1500 | 3.0301        | 2.9781          | 0.4967                     | 0.3029                | 0.3998                       |
+| 2.0115 | 1750 | 3.0238        | 2.9768          | 0.4706                     | 0.2717                | 0.3712                       |
+| 2.2989 | 2000 | 2.9739        | 2.9735          | 0.4828                     | 0.2734                | 0.3781                       |
+| 2.5862 | 2250 | 2.9709        | 2.9696          | 0.4896                     | 0.2257                | 0.3576                       |
+| 2.8736 | 2500 | 2.9652        | 2.9693          | 0.4816                     | 0.2553                | 0.3684                       |
+| 3.1609 | 2750 | 2.9475        | 2.9720          | 0.4815                     | 0.2618                | 0.3717                       |
+| 3.4483 | 3000 | 2.9313        | 2.9715          | 0.5048                     | 0.2831                | 0.3939                       |
+| 3.7356 | 3250 | 2.9309        | 2.9705          | 0.4606                     | 0.2879                | 0.3743                       |
+| 4.0230 | 3500 | 2.9264        | 2.9712          | 0.5049                     | 0.2774                | 0.3911                       |
+| 4.3103 | 3750 | 2.9056        | 2.9722          | 0.4758                     | 0.2532                | 0.3645                       |
+| 4.5977 | 4000 | 2.9056        | 2.9708          | 0.5004                     | 0.2724                | 0.3864                       |
+| 4.8851 | 4250 | 2.9038        | 2.9705          | 0.5066                     | 0.2675                | 0.3870                       |
+| 5.1724 | 4500 | 2.8932        | 2.9729          | 0.4890                     | 0.2627                | 0.3759                       |
+| 5.4598 | 4750 | 2.8884        | 2.9710          | 0.5016                     | 0.2822                | 0.3919                       |
+| 5.7471 | 5000 | 2.8876        | 2.9712          | 0.4962                     | 0.2725                | 0.3844                       |
 
 
 ### Framework Versions

config_sentence_transformers.json

@@ -1,10 +1,10 @@
 {
-  "model_type": "SentenceTransformer",
   "__version__": {
     "sentence_transformers": "5.2.0",
     "transformers": "4.57.3",
     "pytorch": "2.9.1+cu128"
   },
+  "model_type": "SentenceTransformer",
   "prompts": {
     "query": "",
     "document": ""