Enregistré dans:
Détails bibliographiques
Auteurs principaux: Zhang, Dun, Zeng, Ziyang, Zhou, Yudong, Lu, Shuyang
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2511.14405
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866911275910430720
author Zhang, Dun
Zeng, Ziyang
Zhou, Yudong
Lu, Shuyang
author_facet Zhang, Dun
Zeng, Ziyang
Zhou, Yudong
Lu, Shuyang
contents This technical report presents the training methodology and evaluation results of the open-source Jasper-Token-Compression-600M model, released in November 2025. Building on previous distillation-based recipes from the English Stella and Jasper models, we successfully extend this approach to a bilingual (English and Chinese) domain, further enhancing model performance through the incorporation of contrastive learning. A key innovation of our model is the introduction of a one-dimensional convolution-based token compression module. We dynamically adjust the compression rate during training, enabling the model to learn more robust and efficient compressed text representations. By combining knowledge distillation with token compression techniques, we achieve significant improvements in both embedding quality and inference efficiency. Our model performs with higher efficiency than a traditional 0.6B model while achieving performance comparable to that of an 8B model. For more information on the model release, visit: https://huggingface.co/infgrad/Jasper-Token-Compression-600M.
format Preprint
id arxiv_https___arxiv_org_abs_2511_14405
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Jasper-Token-Compression-600M Technical Report
Zhang, Dun
Zeng, Ziyang
Zhou, Yudong
Lu, Shuyang
Information Retrieval
This technical report presents the training methodology and evaluation results of the open-source Jasper-Token-Compression-600M model, released in November 2025. Building on previous distillation-based recipes from the English Stella and Jasper models, we successfully extend this approach to a bilingual (English and Chinese) domain, further enhancing model performance through the incorporation of contrastive learning. A key innovation of our model is the introduction of a one-dimensional convolution-based token compression module. We dynamically adjust the compression rate during training, enabling the model to learn more robust and efficient compressed text representations. By combining knowledge distillation with token compression techniques, we achieve significant improvements in both embedding quality and inference efficiency. Our model performs with higher efficiency than a traditional 0.6B model while achieving performance comparable to that of an 8B model. For more information on the model release, visit: https://huggingface.co/infgrad/Jasper-Token-Compression-600M.
title Jasper-Token-Compression-600M Technical Report
topic Information Retrieval
url https://arxiv.org/abs/2511.14405