---
language:
- en
library_name: sentence-transformers
tags:
- sentence-transformers
- sentence-similarity
- feature-extraction
- generated_from_trainer
- dataset_size:314315
- loss:AdaptiveLayerLoss
- loss:MultipleNegativesRankingLoss
base_model: microsoft/deberta-v3-small
datasets:
- stanfordnlp/snli
- sentence-transformers/stsb
metrics:
- pearson_cosine
- spearman_cosine
- pearson_manhattan
- spearman_manhattan
- pearson_euclidean
- spearman_euclidean
- pearson_dot
- spearman_dot
- pearson_max
- spearman_max
- cosine_accuracy
- cosine_accuracy_threshold
- cosine_f1
- cosine_f1_threshold
- cosine_precision
- cosine_recall
- cosine_ap
- dot_accuracy
- dot_accuracy_threshold
- dot_f1
- dot_f1_threshold
- dot_precision
- dot_recall
- dot_ap
- manhattan_accuracy
- manhattan_accuracy_threshold
- manhattan_f1
- manhattan_f1_threshold
- manhattan_precision
- manhattan_recall
- manhattan_ap
- euclidean_accuracy
- euclidean_accuracy_threshold
- euclidean_f1
- euclidean_f1_threshold
- euclidean_precision
- euclidean_recall
- euclidean_ap
- max_accuracy
- max_accuracy_threshold
- max_f1
- max_f1_threshold
- max_precision
- max_recall
- max_ap
widget:
- source_sentence: The pitcher is pitching the ball in a game of baseball.
sentences:
- the lady digs into the ground
- A group of people are sitting at tables.
- The pitcher throws the ball.
- source_sentence: People are conversing at a dining table under a canopy.
sentences:
- A canine is using his legs.
- The people are creative.
- People at a party are seated for dinner on the lawn.
- source_sentence: Two teenage girls conversing next to lockers.
sentences:
- Girls talking about their problems next to lockers.
- A group of people play in the ocean.
- The man is testing the bike.
- source_sentence: A young boy in a hoodie climbs a red slide sitting on a red and
green checkered background.
sentences:
- People are buying food from a street vendor.
- A boy is playing.
- A dog outside digging.
- source_sentence: A professional swimmer spits water out after surfacing while grabbing
the hand of someone helping him back to land.
sentences:
- A group of people wait in a line.
- A tourist has his picture taken on Easter Island.
- The swimmer almost drowned after being sucked under a fast current.
pipeline_tag: sentence-similarity
model-index:
- name: SentenceTransformer based on microsoft/deberta-v3-small
results:
- task:
type: semantic-similarity
name: Semantic Similarity
dataset:
name: Unknown
type: unknown
metrics:
- type: pearson_cosine
value: 0.7641416788909702
name: Pearson Cosine
- type: spearman_cosine
value: 0.763668633314844
name: Spearman Cosine
- type: pearson_manhattan
value: 0.7808845626705342
name: Pearson Manhattan
- type: spearman_manhattan
value: 0.783960481366303
name: Spearman Manhattan
- type: pearson_euclidean
value: 0.7714319160162553
name: Pearson Euclidean
- type: spearman_euclidean
value: 0.7750607015673249
name: Spearman Euclidean
- type: pearson_dot
value: 0.587659176024498
name: Pearson Dot
- type: spearman_dot
value: 0.6010467058509925
name: Spearman Dot
- type: pearson_max
value: 0.7808845626705342
name: Pearson Max
- type: spearman_max
value: 0.783960481366303
name: Spearman Max
- task:
type: binary-classification
name: Binary Classification
dataset:
name: Unknown
type: unknown
metrics:
- type: cosine_accuracy
value: 0.6773826673743271
name: Cosine Accuracy
- type: cosine_accuracy_threshold
value: 0.5830236673355103
name: Cosine Accuracy Threshold
- type: cosine_f1
value: 0.7209834880077135
name: Cosine F1
- type: cosine_f1_threshold
value: 0.5085207223892212
name: Cosine F1 Threshold
- type: cosine_precision
value: 0.6137273007079102
name: Cosine Precision
- type: cosine_recall
value: 0.873667299547247
name: Cosine Recall
- type: cosine_ap
value: 0.7219177301725319
name: Cosine Ap
- type: dot_accuracy
value: 0.6389415421942528
name: Dot Accuracy
- type: dot_accuracy_threshold
value: 45.1016845703125
name: Dot Accuracy Threshold
- type: dot_f1
value: 0.7090406632451638
name: Dot F1
- type: dot_f1_threshold
value: 32.459449768066406
name: Dot F1 Threshold
- type: dot_precision
value: 0.5775450202131569
name: Dot Precision
- type: dot_recall
value: 0.9180663064115671
name: Dot Recall
- type: dot_ap
value: 0.6795197111227502
name: Dot Ap
- type: manhattan_accuracy
value: 0.6625217984684206
name: Manhattan Accuracy
- type: manhattan_accuracy_threshold
value: 158.29489135742188
name: Manhattan Accuracy Threshold
- type: manhattan_f1
value: 0.7041269465332466
name: Manhattan F1
- type: manhattan_f1_threshold
value: 178.5047607421875
name: Manhattan F1 Threshold
- type: manhattan_precision
value: 0.5921131248755228
name: Manhattan Precision
- type: manhattan_recall
value: 0.8684095224185775
name: Manhattan Recall
- type: manhattan_ap
value: 0.7054112942825768
name: Manhattan Ap
- type: euclidean_accuracy
value: 0.6578967321252559
name: Euclidean Accuracy
- type: euclidean_accuracy_threshold
value: 7.951424598693848
name: Euclidean Accuracy Threshold
- type: euclidean_f1
value: 0.7015471831817645
name: Euclidean F1
- type: euclidean_f1_threshold
value: 9.045232772827148
name: Euclidean F1 Threshold
- type: euclidean_precision
value: 0.5888767720828789
name: Euclidean Precision
- type: euclidean_recall
value: 0.8675332262304659
name: Euclidean Recall
- type: euclidean_ap
value: 0.7024193897121154
name: Euclidean Ap
- type: max_accuracy
value: 0.6773826673743271
name: Max Accuracy
- type: max_accuracy_threshold
value: 158.29489135742188
name: Max Accuracy Threshold
- type: max_f1
value: 0.7209834880077135
name: Max F1
- type: max_f1_threshold
value: 178.5047607421875
name: Max F1 Threshold
- type: max_precision
value: 0.6137273007079102
name: Max Precision
- type: max_recall
value: 0.9180663064115671
name: Max Recall
- type: max_ap
value: 0.7219177301725319
name: Max Ap
---
# SentenceTransformer based on microsoft/deberta-v3-small
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [microsoft/deberta-v3-small](https://huggingface.co./microsoft/deberta-v3-small) on the [stanfordnlp/snli](https://huggingface.co./datasets/stanfordnlp/snli) dataset. 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:** [microsoft/deberta-v3-small](https://huggingface.co./microsoft/deberta-v3-small)
- **Maximum Sequence Length:** 512 tokens
- **Output Dimensionality:** 768 tokens
- **Similarity Function:** Cosine Similarity
- **Training Dataset:**
- [stanfordnlp/snli](https://huggingface.co./datasets/stanfordnlp/snli)
- **Language:** en
### Model Sources
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co./models?library=sentence-transformers)
### Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: DebertaV2Model
(1): Pooling({'word_embedding_dimension': 768, '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})
)
```
## Usage
### Direct Usage (Sentence Transformers)
First install the Sentence Transformers library:
```bash
pip install -U sentence-transformers
```
Then you can load this model and run inference.
```python
from sentence_transformers import SentenceTransformer
# Download from the 🤗 Hub
model = SentenceTransformer("bobox/DeBERTaV3-small-SentenceTransformer-AdaptiveLayerAll")
# Run inference
sentences = [
'A professional swimmer spits water out after surfacing while grabbing the hand of someone helping him back to land.',
'The swimmer almost drowned after being sucked under a fast current.',
'A group of people wait in a line.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
```
## Evaluation
### Metrics
#### Semantic Similarity
* Evaluated with [EmbeddingSimilarityEvaluator
](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| pearson_cosine | 0.7641 |
| **spearman_cosine** | **0.7637** |
| pearson_manhattan | 0.7809 |
| spearman_manhattan | 0.784 |
| pearson_euclidean | 0.7714 |
| spearman_euclidean | 0.7751 |
| pearson_dot | 0.5877 |
| spearman_dot | 0.601 |
| pearson_max | 0.7809 |
| spearman_max | 0.784 |
#### Binary Classification
* Evaluated with [BinaryClassificationEvaluator
](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.BinaryClassificationEvaluator)
| Metric | Value |
|:-----------------------------|:-----------|
| cosine_accuracy | 0.6774 |
| cosine_accuracy_threshold | 0.583 |
| cosine_f1 | 0.721 |
| cosine_f1_threshold | 0.5085 |
| cosine_precision | 0.6137 |
| cosine_recall | 0.8737 |
| cosine_ap | 0.7219 |
| dot_accuracy | 0.6389 |
| dot_accuracy_threshold | 45.1017 |
| dot_f1 | 0.709 |
| dot_f1_threshold | 32.4594 |
| dot_precision | 0.5775 |
| dot_recall | 0.9181 |
| dot_ap | 0.6795 |
| manhattan_accuracy | 0.6625 |
| manhattan_accuracy_threshold | 158.2949 |
| manhattan_f1 | 0.7041 |
| manhattan_f1_threshold | 178.5048 |
| manhattan_precision | 0.5921 |
| manhattan_recall | 0.8684 |
| manhattan_ap | 0.7054 |
| euclidean_accuracy | 0.6579 |
| euclidean_accuracy_threshold | 7.9514 |
| euclidean_f1 | 0.7015 |
| euclidean_f1_threshold | 9.0452 |
| euclidean_precision | 0.5889 |
| euclidean_recall | 0.8675 |
| euclidean_ap | 0.7024 |
| max_accuracy | 0.6774 |
| max_accuracy_threshold | 158.2949 |
| max_f1 | 0.721 |
| max_f1_threshold | 178.5048 |
| max_precision | 0.6137 |
| max_recall | 0.9181 |
| **max_ap** | **0.7219** |
## Training Details
### Training Dataset
#### stanfordnlp/snli
* Dataset: [stanfordnlp/snli](https://huggingface.co./datasets/stanfordnlp/snli) at [cdb5c3d](https://huggingface.co./datasets/stanfordnlp/snli/tree/cdb5c3d5eed6ead6e5a341c8e56e669bb666725b)
* Size: 314,315 training samples
* Columns: sentence1
, sentence2
, and label
* Approximate statistics based on the first 1000 samples:
| | sentence1 | sentence2 | label |
|:--------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:-----------------------------|
| type | string | string | int |
| details |
A person on a horse jumps over a broken down airplane.
| A person is outdoors, on a horse.
| 0
|
| Children smiling and waving at camera
| There are children present
| 0
|
| A boy is jumping on skateboard in the middle of a red bridge.
| The boy does a skateboarding trick.
| 0
|
* Loss: [AdaptiveLayerLoss
](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#adaptivelayerloss) with these parameters:
```json
{
"loss": "MultipleNegativesRankingLoss",
"n_layers_per_step": -1,
"last_layer_weight": 1,
"prior_layers_weight": 1,
"kl_div_weight": 1.2,
"kl_temperature": 1.2
}
```
### Evaluation Dataset
#### sentence-transformers/stsb
* Dataset: [sentence-transformers/stsb](https://huggingface.co./datasets/sentence-transformers/stsb) at [ab7a5ac](https://huggingface.co./datasets/sentence-transformers/stsb/tree/ab7a5ac0e35aa22088bdcf23e7fd99b220e53308)
* Size: 1,500 evaluation samples
* Columns: sentence1
, sentence2
, and score
* Approximate statistics based on the first 1000 samples:
| | sentence1 | sentence2 | score |
|:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------|
| type | string | string | float |
| details | A man with a hard hat is dancing.
| A man wearing a hard hat is dancing.
| 1.0
|
| A young child is riding a horse.
| A child is riding a horse.
| 0.95
|
| A man is feeding a mouse to a snake.
| The man is feeding a mouse to the snake.
| 1.0
|
* Loss: [AdaptiveLayerLoss
](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#adaptivelayerloss) with these parameters:
```json
{
"loss": "MultipleNegativesRankingLoss",
"n_layers_per_step": -1,
"last_layer_weight": 1,
"prior_layers_weight": 1,
"kl_div_weight": 1.2,
"kl_temperature": 1.2
}
```
### Training Hyperparameters
#### Non-Default Hyperparameters
- `eval_strategy`: steps
- `per_device_train_batch_size`: 32
- `per_device_eval_batch_size`: 16
- `learning_rate`: 5e-06
- `weight_decay`: 1e-07
- `warmup_ratio`: 0.33
- `save_safetensors`: False
- `fp16`: True
- `push_to_hub`: True
- `hub_model_id`: bobox/DeBERTaV3-small-SentenceTransformer-AdaptiveLayerAlln
- `hub_strategy`: checkpoint
- `batch_sampler`: no_duplicates
#### All Hyperparameters