README.md

metadata

library_name: sentence-transformers
pipeline_tag: sentence-similarity
tags:
  - sentence-transformers
  - feature-extraction
  - sentence-similarity
  - transformers
  - phobert
  - french
  - sentence-embedding
license: apache-2.0
language:
  - fr
  - en
metrics:
  - pearsonr
  - spearmanr

Model Description:

french-embedding-LongContext is the Embedding Model for French-English language with context length up to 8096 tokens. This model is a specialized text-embedding trained specifically for the french language, which is built upon gte-multilingual and trained using the Multi-Negative Ranking Loss, Matryoshka2dLoss and SimilarityLoss.

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 8192, 'do_lower_case': False}) with Transformer model: BilingualModel 
  (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})
  (2): Normalize()
)

Training and Fine-tuning process

The model underwent a rigorous four-stage training and fine-tuning process, each tailored to enhance its ability to generate precise and contextually relevant sentence embeddings for the french language. Below is an outline of these stages:

Stage 1: Training NLI on dataset XNLI:

• Dataset: XNLI (fr-en)
• Method: Training using Multi-Negative Ranking Loss and Matryoshka2dLoss. This stage focused on improving the model's ability to discern and rank nuanced differences in sentence semantics.

Stage 2: Fine-tuning for Semantic Textual Similarity on STS Benchmark

• Dataset: STS-B (fr-en)
• Method: Fine-tuning specifically for the semantic textual similarity benchmark using Siamese BERT-Networks configured with the 'sentence-transformers' library. This stage honed the model's precision in capturing semantic similarity across various types of french texts.

Usage:

Using this model becomes easy when you have sentence-transformers installed:

pip install -U sentence-transformers

Then you can use the model like this:

from sentence_transformers import SentenceTransformer
sentences = ["Paris est une capitale de la France", "Paris is a capital of France"]



model = SentenceTransformer('dangvantuan/french-embedding-LongContext', trust_remote_code=True)
embeddings = model.encode(sentences)
print(embeddings)

Evaluation

Citation

@article{reimers2019sentence,
   title={Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks},
   author={Nils Reimers, Iryna Gurevych},
   journal={https://arxiv.org/abs/1908.10084},
   year={2019}
}


@article{zhang2024mgte,
  title={mGTE: Generalized Long-Context Text Representation and Reranking Models for Multilingual Text Retrieval},
  author={Zhang, Xin and Zhang, Yanzhao and Long, Dingkun and Xie, Wen and Dai, Ziqi and Tang, Jialong and Lin, Huan and Yang, Baosong and Xie, Pengjun and Huang, Fei and others},
  journal={arXiv preprint arXiv:2407.19669},
  year={2024}
}

@article{li2023towards,
  title={Towards general text embeddings with multi-stage contrastive learning},
  author={Li, Zehan and Zhang, Xin and Zhang, Yanzhao and Long, Dingkun and Xie, Pengjun and Zhang, Meishan},
  journal={arXiv preprint arXiv:2308.03281},
  year={2023}
}

@article{li20242d,
  title={2d matryoshka sentence embeddings},
  author={Li, Xianming and Li, Zongxi and Li, Jing and Xie, Haoran and Li, Qing},
  journal={arXiv preprint arXiv:2402.14776},
  year={2024}
}