Conan-embedding-v1

Performance

Model Average CLS Clustering Reranking Retrieval STS Pair_CLS
gte-Qwen2-7B-instruct 72.05 75.09 66.06 68.92 76.03 65.33 87.48
xiaobu-embedding-v2 72.43 74.67 65.17 72.58 76.5 64.53 91.87
Conan-embedding-v1 72.62 75.03 66.33 72.76 76.67 64.18 91.66

Methods and Training Detials

Please refer to our technical report.

Citation

If you find our models / papers useful in your research, please consider giving ❤️ and citations. Thanks!

@misc{li2024conanembeddinggeneraltextembedding,
  title={Conan-embedding: General Text Embedding with More and Better Negative Samples}, 
  author={Shiyu Li and Yang Tang and Shizhe Chen and Xi Chen},
  year={2024},
  eprint={2408.15710},
  archivePrefix={arXiv},
  primaryClass={cs.CL},
  url={https://arxiv.org/abs/2408.15710}, 
}

About

Created by the Tencent BAC Group. All rights reserved.

License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

Downloads last month
47,500
Safetensors
Model size
326M params
Tensor type
F32
·
Inference API
Unable to determine this model’s pipeline type. Check the docs .

Model tree for TencentBAC/Conan-embedding-v1

Quantizations
2 models

Evaluation results