Add new SentenceTransformer model

README.md

@@ -5,38 +5,110 @@ tags:
 - feature-extraction
 - dense
 - generated_from_trainer
-- dataset_size:100000
+- dataset_size:713743
 - loss:MultipleNegativesRankingLoss
 base_model: prajjwal1/bert-small
 widget:
-- source_sentence: How do I calculate IQ?
+- source_sentence: 'Abraham Lincoln: Why is the Gettysburg Address so memorable?'
   sentences:
-  - What is the easiest way to know my IQ?
-  - How do I calculate not IQ ?
-  - What are some creative and innovative business ideas with less investment in India?
-- source_sentence: How can I learn martial arts in my home?
+  - 'Abraham Lincoln: Why is the Gettysburg Address so memorable?'
+  - What does the Gettysburg Address really mean?
+  - What is eatalo.com?
+- source_sentence: Has the influence of Ancient Carthage in science, math, and society
+    been underestimated?
   sentences:
-  - How can I learn martial arts by myself?
-  - What are the advantages and disadvantages of investing in gold?
-  - Can people see that I have looked at their pictures on instagram if I am not following
-    them?
-- source_sentence: When Enterprise picks you up do you have to take them back?
+  - How does one earn money online without an investment from home?
+  - Has the influence of Ancient Carthage in science, math, and society been underestimated?
+  - Has the influence of the Ancient Etruscans in science and math been underestimated?
+- source_sentence: Is there any app that shares charging to others like share it how
+    we transfer files?
   sentences:
-  - Are there any software Training institute in Tuticorin?
-  - When Enterprise picks you up do you have to take them back?
-  - When Enterprise picks you up do them have to take youback?
-- source_sentence: What are some non-capital goods?
+  - How do you think of Chinese claims that the present Private Arbitration is illegal,
+    its verdict violates the UNCLOS and is illegal?
+  - Is there any app that shares charging to others like share it how we transfer
+    files?
+  - Are there any platforms that provides end-to-end encryption for file transfer/
+    sharing?
+- source_sentence: Why AAP’s MLA Dinesh Mohaniya has been arrested?
   sentences:
-  - What are capital goods?
-  - How is the value of [math]\pi[/math] calculated?
-  - What are some non-capital goods?
-- source_sentence: What is the QuickBooks technical support phone number in New York?
+  - What are your views on the latest sex scandal by AAP MLA Sandeep Kumar?
+  - What is a dc current? What are some examples?
+  - Why AAP’s MLA Dinesh Mohaniya has been arrested?
+- source_sentence: What is the difference between economic growth and economic development?
   sentences:
-  - What caused the Great Depression?
-  - Can I apply for PR in Canada?
-  - Which is the best QuickBooks Hosting Support Number in New York?
+  - How cold can the Gobi Desert get, and how do its average temperatures compare
+    to the ones in the Simpson Desert?
+  - the difference between economic growth and economic development is What?
+  - What is the difference between economic growth and economic development?
 pipeline_tag: sentence-similarity
 library_name: sentence-transformers
+metrics:
+- cosine_accuracy@1
+- cosine_accuracy@3
+- cosine_accuracy@5
+- cosine_precision@1
+- cosine_precision@3
+- cosine_precision@5
+- cosine_recall@1
+- cosine_recall@3
+- cosine_recall@5
+- cosine_ndcg@10
+- cosine_mrr@1
+- cosine_mrr@5
+- cosine_mrr@10
+- cosine_map@100
+model-index:
+- name: SentenceTransformer based on prajjwal1/bert-small
+  results:
+  - task:
+      type: information-retrieval
+      name: Information Retrieval
+    dataset:
+      name: val
+      type: val
+    metrics:
+    - type: cosine_accuracy@1
+      value: 0.7467
+      name: Cosine Accuracy@1
+    - type: cosine_accuracy@3
+      value: 0.81875
+      name: Cosine Accuracy@3
+    - type: cosine_accuracy@5
+      value: 0.842275
+      name: Cosine Accuracy@5
+    - type: cosine_precision@1
+      value: 0.7467
+      name: Cosine Precision@1
+    - type: cosine_precision@3
+      value: 0.27291666666666664
+      name: Cosine Precision@3
+    - type: cosine_precision@5
+      value: 0.16845500000000002
+      name: Cosine Precision@5
+    - type: cosine_recall@1
+      value: 0.7467
+      name: Cosine Recall@1
+    - type: cosine_recall@3
+      value: 0.81875
+      name: Cosine Recall@3
+    - type: cosine_recall@5
+      value: 0.842275
+      name: Cosine Recall@5
+    - type: cosine_ndcg@10
+      value: 0.8088581445720447
+      name: Cosine Ndcg@10
+    - type: cosine_mrr@1
+      value: 0.7467
+      name: Cosine Mrr@1
+    - type: cosine_mrr@5
+      value: 0.784354583333328
+      name: Cosine Mrr@5
+    - type: cosine_mrr@10
+      value: 0.7884659325396792
+      name: Cosine Mrr@10
+    - type: cosine_map@100
+      value: 0.7917670616349511
+      name: Cosine Map@100
 ---
 
 # SentenceTransformer based on prajjwal1/bert-small
@@ -85,12 +157,12 @@ Then you can load this model and run inference.
 from sentence_transformers import SentenceTransformer
 
 # Download from the 🤗 Hub
-model = SentenceTransformer("sentence_transformers_model_id")
+model = SentenceTransformer("redis/model-b-structured")
 # Run inference
 sentences = [
-    'What is the QuickBooks technical support phone number in New York?',
-    'Which is the best QuickBooks Hosting Support Number in New York?',
-    'Can I apply for PR in Canada?',
+    'What is the difference between economic growth and economic development?',
+    'What is the difference between economic growth and economic development?',
+    'the difference between economic growth and economic development is What?',
 ]
 embeddings = model.encode(sentences)
 print(embeddings.shape)
@@ -99,9 +171,9 @@ print(embeddings.shape)
 # Get the similarity scores for the embeddings
 similarities = model.similarity(embeddings, embeddings)
 print(similarities)
-# tensor([[1.0000, 0.8563, 0.0594],
-#         [0.8563, 1.0000, 0.1245],
-#         [0.0594, 0.1245, 1.0000]])
+# tensor([[1.0000, 1.0000, 0.9993],
+#         [1.0000, 1.0000, 0.9993],
+#         [0.9993, 0.9993, 1.0000]])
 ```
 
 <!--
@@ -128,6 +200,32 @@ You can finetune this model on your own dataset.
 *List how the model may foreseeably be misused and address what users ought not to do with the model.*
 -->
 
+## Evaluation
+
+### Metrics
+
+#### Information Retrieval
+
+* Dataset: `val`
+* Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
+
+| Metric             | Value      |
+|:-------------------|:-----------|
+| cosine_accuracy@1  | 0.7467     |
+| cosine_accuracy@3  | 0.8187     |
+| cosine_accuracy@5  | 0.8423     |
+| cosine_precision@1 | 0.7467     |
+| cosine_precision@3 | 0.2729     |
+| cosine_precision@5 | 0.1685     |
+| cosine_recall@1    | 0.7467     |
+| cosine_recall@3    | 0.8187     |
+| cosine_recall@5    | 0.8423     |
+| **cosine_ndcg@10** | **0.8089** |
+| cosine_mrr@1       | 0.7467     |
+| cosine_mrr@5       | 0.7844     |
+| cosine_mrr@10      | 0.7885     |
+| cosine_map@100     | 0.7918     |
+
 <!--
 ## Bias, Risks and Limitations
 
@@ -146,23 +244,49 @@ You can finetune this model on your own dataset.
 
 #### Unnamed Dataset
 
-* Size: 100,000 training samples
-* Columns: <code>sentence_0</code>, <code>sentence_1</code>, and <code>sentence_2</code>
+* Size: 713,743 training samples
+* Columns: <code>anchor</code>, <code>positive</code>, and <code>negative</code>
 * Approximate statistics based on the first 1000 samples:
-  |         | sentence_0                                                                        | sentence_1                                                                        | sentence_2                                                                        |
+  |         | anchor                                                                            | positive                                                                          | negative                                                                          |
   |:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|
   | type    | string                                                                            | string                                                                            | string                                                                            |
-  | details | <ul><li>min: 6 tokens</li><li>mean: 15.79 tokens</li><li>max: 66 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 15.68 tokens</li><li>max: 66 tokens</li></ul> | <ul><li>min: 7 tokens</li><li>mean: 16.37 tokens</li><li>max: 67 tokens</li></ul> |
+  | details | <ul><li>min: 6 tokens</li><li>mean: 16.07 tokens</li><li>max: 53 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 16.03 tokens</li><li>max: 53 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 16.81 tokens</li><li>max: 58 tokens</li></ul> |
 * Samples:
-  | sentence_0                                                       | sentence_1                                                       | sentence_2                                                                        |
-  |:-----------------------------------------------------------------|:-----------------------------------------------------------------|:----------------------------------------------------------------------------------|
-  | <code>Is masturbating bad for boys?</code>                       | <code>Is masturbating bad for boys?</code>                       | <code>How harmful or unhealthy is masturbation?</code>                            |
-  | <code>Does a train engine move in reverse?</code>                | <code>Does a train engine move in reverse?</code>                | <code>Time moves forward, not in reverse. Doesn't that make time a vector?</code> |
-  | <code>What is the most badass thing anyone has ever done?</code> | <code>What is the most badass thing anyone has ever done?</code> | <code>anyone is the most badass thing Whathas ever done?</code>                   |
+  | anchor                                                                         | positive                                                                              | negative                                                                                                          |
+  |:-------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------------------------------------|
+  | <code>Which one is better Linux OS? Ubuntu or Mint?</code>                     | <code>Why do you use Linux Mint?</code>                                               | <code>Which one is not better Linux OS ? Ubuntu or Mint ?</code>                                                  |
+  | <code>What is flow?</code>                                                     | <code>What is flow?</code>                                                            | <code>What are flow lines?</code>                                                                                 |
+  | <code>How is Trump planning to get Mexico to pay for his supposed wall?</code> | <code>How is it possible for Donald Trump to force Mexico to pay for the wall?</code> | <code>Why do we connect the positive terminal before the negative terminal to ground in a vehicle battery?</code> |
 * Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
   ```json
   {
-      "scale": 20.0,
+      "scale": 1.0,
+      "similarity_fct": "cos_sim",
+      "gather_across_devices": false
+  }
+  ```
+
+### Evaluation Dataset
+
+#### Unnamed Dataset
+
+* Size: 40,000 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: 6 tokens</li><li>mean: 15.52 tokens</li><li>max: 74 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 15.51 tokens</li><li>max: 74 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 16.79 tokens</li><li>max: 69 tokens</li></ul> |
+* Samples:
+  | anchor                                                                                                                                                       | positive                                                                                                                                                     | negative                                                                                                                                                            |
+  |:-------------------------------------------------------------------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------|
+  | <code>Why are all my questions on Quora marked needing improvement?</code>                                                                                   | <code>Why are all my questions immediately being marked as needing improvement?</code>                                                                       | <code>For a post-graduate student in IIT, is it allowed to take an external scholarship as a top-up to his/her MHRD assistantship?</code>                           |
+  | <code>Can blue butter fly needle with vaccum tube be reused? Is it HIV risk? . Heard the needle is too small to be reused . Had blood draw at clinic?</code> | <code>Can blue butter fly needle with vaccum tube be reused? Is it HIV risk? . Heard the needle is too small to be reused . Had blood draw at clinic?</code> | <code>Can blue butter fly needle with vaccum tube be reused not ? Is it HIV risk ? . Heard the needle is too small to be reused . Had blood draw at clinic ?</code> |
+  | <code>Why do people still believe the world is flat?</code>                                                                                                  | <code>Why are there still people who believe the world is flat?</code>                                                                                       | <code>I'm not able to buy Udemy course .it is not accepting mine and my friends debit card.my card can be used for Flipkart .how to purchase now?</code>            |
+* Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
+  ```json
+  {
+      "scale": 1.0,
       "similarity_fct": "cos_sim",
       "gather_across_devices": false
   }
@@ -171,36 +295,49 @@ You can finetune this model on your own dataset.
 ### Training Hyperparameters
 #### Non-Default Hyperparameters
 
-- `per_device_train_batch_size`: 64
-- `per_device_eval_batch_size`: 64
+- `eval_strategy`: steps
+- `per_device_train_batch_size`: 1024
+- `per_device_eval_batch_size`: 1024
+- `learning_rate`: 2e-05
+- `weight_decay`: 0.0001
+- `max_steps`: 2000
+- `warmup_ratio`: 0.1
 - `fp16`: True
-- `multi_dataset_batch_sampler`: round_robin
+- `dataloader_drop_last`: True
+- `dataloader_num_workers`: 1
+- `dataloader_prefetch_factor`: 1
+- `load_best_model_at_end`: True
+- `optim`: adamw_torch
+- `ddp_find_unused_parameters`: False
+- `push_to_hub`: True
+- `hub_model_id`: redis/model-b-structured
+- `eval_on_start`: True
 
 #### All Hyperparameters
 <details><summary>Click to expand</summary>
 
 - `overwrite_output_dir`: False
 - `do_predict`: False
-- `eval_strategy`: no
+- `eval_strategy`: steps
 - `prediction_loss_only`: True
-- `per_device_train_batch_size`: 64
-- `per_device_eval_batch_size`: 64
+- `per_device_train_batch_size`: 1024
+- `per_device_eval_batch_size`: 1024
 - `per_gpu_train_batch_size`: None
 - `per_gpu_eval_batch_size`: None
 - `gradient_accumulation_steps`: 1
 - `eval_accumulation_steps`: None
 - `torch_empty_cache_steps`: None
-- `learning_rate`: 5e-05
-- `weight_decay`: 0.0
+- `learning_rate`: 2e-05
+- `weight_decay`: 0.0001
 - `adam_beta1`: 0.9
 - `adam_beta2`: 0.999
 - `adam_epsilon`: 1e-08
-- `max_grad_norm`: 1
-- `num_train_epochs`: 3
-- `max_steps`: -1
+- `max_grad_norm`: 1.0
+- `num_train_epochs`: 3.0
+- `max_steps`: 2000
 - `lr_scheduler_type`: linear
 - `lr_scheduler_kwargs`: {}
-- `warmup_ratio`: 0.0
+- `warmup_ratio`: 0.1
 - `warmup_steps`: 0
 - `log_level`: passive
 - `log_level_replica`: warning
@@ -228,14 +365,14 @@ You can finetune this model on your own dataset.
 - `tpu_num_cores`: None
 - `tpu_metrics_debug`: False
 - `debug`: []
-- `dataloader_drop_last`: False
-- `dataloader_num_workers`: 0
-- `dataloader_prefetch_factor`: None
+- `dataloader_drop_last`: True
+- `dataloader_num_workers`: 1
+- `dataloader_prefetch_factor`: 1
 - `past_index`: -1
 - `disable_tqdm`: False
 - `remove_unused_columns`: True
 - `label_names`: None
-- `load_best_model_at_end`: False
+- `load_best_model_at_end`: True
 - `ignore_data_skip`: False
 - `fsdp`: []
 - `fsdp_min_num_params`: 0
@@ -245,23 +382,23 @@ You can finetune this model on your own dataset.
 - `parallelism_config`: None
 - `deepspeed`: None
 - `label_smoothing_factor`: 0.0
-- `optim`: adamw_torch_fused
+- `optim`: adamw_torch
 - `optim_args`: None
 - `adafactor`: False
 - `group_by_length`: False
 - `length_column_name`: length
 - `project`: huggingface
 - `trackio_space_id`: trackio
-- `ddp_find_unused_parameters`: None
+- `ddp_find_unused_parameters`: False
 - `ddp_bucket_cap_mb`: None
 - `ddp_broadcast_buffers`: False
 - `dataloader_pin_memory`: True
 - `dataloader_persistent_workers`: False
 - `skip_memory_metrics`: True
 - `use_legacy_prediction_loop`: False
-- `push_to_hub`: False
+- `push_to_hub`: True
 - `resume_from_checkpoint`: None
-- `hub_model_id`: None
+- `hub_model_id`: redis/model-b-structured
 - `hub_strategy`: every_save
 - `hub_private_repo`: None
 - `hub_always_push`: False
@@ -288,32 +425,45 @@ You can finetune this model on your own dataset.
 - `neftune_noise_alpha`: None
 - `optim_target_modules`: None
 - `batch_eval_metrics`: False
-- `eval_on_start`: False
+- `eval_on_start`: True
 - `use_liger_kernel`: False
 - `liger_kernel_config`: None
 - `eval_use_gather_object`: False
 - `average_tokens_across_devices`: True
 - `prompts`: None
 - `batch_sampler`: batch_sampler
-- `multi_dataset_batch_sampler`: round_robin
+- `multi_dataset_batch_sampler`: proportional
 - `router_mapping`: {}
 - `learning_rate_mapping`: {}
 
 </details>
 
 ### Training Logs
-| Epoch  | Step | Training Loss |
-|:------:|:----:|:-------------:|
-| 0.3199 | 500  | 0.4294        |
-| 0.6398 | 1000 | 0.1268        |
-| 0.9597 | 1500 | 0.1           |
-| 1.2796 | 2000 | 0.0792        |
-| 1.5995 | 2500 | 0.0706        |
-| 1.9194 | 3000 | 0.0687        |
-| 2.2393 | 3500 | 0.0584        |
-| 2.5592 | 4000 | 0.057         |
-| 2.8791 | 4500 | 0.0581        |
-
+| Epoch      | Step     | Training Loss | Validation Loss | val_cosine_ndcg@10 |
+|:----------:|:--------:|:-------------:|:---------------:|:------------------:|
+| 0          | 0        | -             | 7.3381          | 0.7794             |
+| 0.1435     | 100      | 7.1693        | 6.7524          | 0.8134             |
+| 0.2869     | 200      | 6.8567        | 6.6976          | 0.7954             |
+| 0.4304     | 300      | 6.7958        | 6.6911          | 0.7901             |
+| 0.5739     | 400      | 6.7758        | 6.6861          | 0.7883             |
+| 0.7174     | 500      | 6.7656        | 6.6830          | 0.7886             |
+| 0.8608     | 600      | 6.7582        | 6.6809          | 0.7894             |
+| 1.0043     | 700      | 6.7527        | 6.6786          | 0.7903             |
+| 1.1478     | 800      | 6.7479        | 6.6767          | 0.7915             |
+| 1.2912     | 900      | 6.7445        | 6.6760          | 0.7925             |
+| 1.4347     | 1000     | 6.7415        | 6.6744          | 0.7940             |
+| 1.5782     | 1100     | 6.7394        | 6.6739          | 0.7954             |
+| 1.7217     | 1200     | 6.7372        | 6.6731          | 0.7966             |
+| 1.8651     | 1300     | 6.735         | 6.6727          | 0.7984             |
+| 2.0086     | 1400     | 6.7338        | 6.6722          | 0.8004             |
+| 2.1521     | 1500     | 6.7325        | 6.6716          | 0.8026             |
+| 2.2956     | 1600     | 6.7313        | 6.6713          | 0.8046             |
+| 2.4390     | 1700     | 6.7304        | 6.6710          | 0.8067             |
+| 2.5825     | 1800     | 6.7299        | 6.6706          | 0.8080             |
+| 2.7260     | 1900     | 6.729         | 6.6704          | 0.8085             |
+| **2.8694** | **2000** | **6.7289**    | **6.6703**      | **0.8089**         |
+
+* The bold row denotes the saved checkpoint.
 
 ### Framework Versions
 - Python: 3.10.18