Sign inGet started
Dataset · Hugging Face

whybe-choi/EDIRIT2IRetrieval

EDIRIT2IRetrieval An MTEB dataset Massive Text Embedding Benchmark Retrieve edited images from a reference image and textual modification instruction using the streamlined EDIR benchmark. Task category Any2AnyRetrieval (image+text-to-image) Domains Scene, Constructed Reference Rethinking Composed Image Retrieval Evaluation: A Fine-Grained Benchmark from Image Editing Source datasets: whybe-choi/EDIRIT2IRetrieval How to evaluate on this task You… See the full description on the dataset page: https://huggingface.co/datasets/whybe-choi/EDIRIT2IRetrieval.

Author / organizationwhybe-choi
License (as declared)mit
Downloads129 (as reported by Hugging Face)
Likes0
Created19 August 2026
Last updated at source19 August 2026
mtebimagetext
Embed this entry
<iframe src="https://silytics.si/embed/item/9c7fc162-236e-419d-a5cb-ae7cada8ffb4" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

Popularity, last 2 days

129 now (-11 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.

More from whybe-choi

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 Information may be incomplete or out of date.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.