DeepGlint-AI/UniME-R1

UniME-R1: Training Datasets for Unified Multimodal Retrieval Zelong Sun*, Jun Wang*, Kaicheng Yang, Tiancheng Gu, Ziyong Feng, Zhiwu Lu Glint Lab This repository hosts the training datasets of UniME-R1, an Embedder–Adviser framework that learns Retrieval-Centric Chain-of-Thought (RC-CoT) from retrieval feedback. It covers three stages of data used in the paper: the dual-mode Embedder contrastive data, the Adviser SFT structured-annotation data, and the Adviser GRPO… See the full description on the dataset page: https://huggingface.co/datasets/DeepGlint-AI/UniME-R1.

<iframe src="https://silytics.si/embed/item/3e7e3673-a7bd-4580-a691-10367cb05d62" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.