Sentest

🚀 SentenceTransformer based on google-bert/bert-base-uncased

This is a sentence-transformers model. It's finetuned from google-bert/bert-base-uncased on the qqp_triplets dataset. It maps sentences and paragraphs to a 768 - dimensional dense vector space. You can use it for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

✨ Features

• Language: Supports English (en).
• Tags: Associated with sentence-transformers, sentence-similarity, feature-extraction, etc.
• Base Model: Built upon google-bert/bert-base-uncased.
• Loss Function: Utilizes TripletLoss.
• Metrics: Evaluated with cosine_accuracy.

📦 Installation

First, you need to install the Sentence Transformers library:

pip install -U sentence-transformers

💻 Usage Examples

Basic Usage

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("palusi/sentest")
# Run inference
sentences = [
    'How can I open my computer if I forget my password?',
    'I forget my PC password what should I do to open it?',
    'I forgot my security code on my Nokia 206 how can I unlock it?',
]
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]

📚 Documentation

Model Details

Model Description

Model Sources

• Documentation: Sentence Transformers Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel 
  (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})
)

Evaluation

Metrics - Triplet

• Dataset: sentest
• Evaluated with: TripletEvaluator

Training Details

Training Dataset - qqp_triplets

• Dataset: qqp_triplets at f475d9c
• Size: 101,762 training samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 6 tokens mean: 13.96 tokens max: 54 tokens	min: 5 tokens mean: 13.99 tokens max: 52 tokens	min: 6 tokens mean: 14.49 tokens max: 73 tokens

• Samples:

  anchor	positive	negative
  Who are Mona Punjabi?	Who are Mona punjabis?	Why are Punjabis so proud of their Punjabi - hood?
  What are some of the best books on/by Bill Gates?	What are the best books of Bill Gates?	Are there any films about Bill Gates?
  Where can I get best pasta in Bangalore?	Where can I get best pasta in Bangalore ?	Where can I get best street food in Bangalore?

• Loss: TripletLoss with these parameters:

  {
      "distance_metric": "TripletDistanceMetric.EUCLIDEAN",
      "triplet_margin": 5
  }
  

Evaluation Dataset - qqp_triplets

• Dataset: qqp_triplets at f475d9c
• Size: 101,762 evaluation samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 6 tokens mean: 13.99 tokens max: 61 tokens	min: 6 tokens mean: 13.76 tokens max: 49 tokens	min: 6 tokens mean: 14.75 tokens max: 78 tokens

• Samples:

  anchor	positive	negative
  How do l study efficiently?	How do you study effectively?	Why can't I study efficiently?
  How do you commit suicide?	What is the easiest way to commite suicide?	What is a way to commit suicide and not damaging your organs so that they can be donated?
  How do you learn to speak a foreign language?	What is the quickest way a person can learn to speak a new language fluently?	What's the easiest foreign language for a native English speaker, living in America, to learn to speak?

• Loss: TripletLoss with these parameters:

  {
      "distance_metric": "TripletDistanceMetric.EUCLIDEAN",
      "triplet_margin": 5
  }
  

Training Hyperparameters

Non - Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• learning_rate: 2e - 05
• weight_decay: 0.01
• num_train_epochs: 1
• warmup_ratio: 0.1
• fp16: True
• load_best_model_at_end: True
• push_to_hub: True
• hub_model_id: palusi/sentest
• batch_sampler: no_duplicates

Training Logs

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.11.11
• Sentence Transformers: 3.4.1
• Transformers: 4.48.2
• PyTorch: 2.5.1+cu124
• Accelerate: 1.3.0
• Datasets: 3.2.0
• Tokenizers: 0.21.0

📄 License

No license information provided in the original document.

📖 Citation

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}

TripletLoss

@misc{hermans2017defense,
    title={In Defense of the Triplet Loss for Person Re-Identification},
    author={Alexander Hermans and Lucas Beyer and Bastian Leibe},
    year={2017},
    eprint={1703.07737},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}