Dataset · Hugging Face

IndexTeam/Sandglass-Bench

Name: Sandglass-Bench
Creator: IndexTeam
License: mit
Keywords: task_categories:translation, language:zh, language:en, language:ja, language:es, language:ar, license:mit, size_categories:1K<n<10K, format:json, modality:text, library:datasets, library:pandas, library:polars, library:mlcroissant, arxiv:2601.10187

Sandglass: Syllable-Controlled Translation Benchmark Sandglass is a benchmark for evaluating syllable-controlled translation quality and controllability. It tests whether models can translate Chinese video subtitles into target languages while adhering to precise syllable count constraints. Overview 3,600 test cases: 300 real video subtitles × 4 languages × 3 length targets Languages: English (en), Japanese (ja), Spanish (es), Arabic (ar) Length targets: short… See the full description on the dataset page: https://huggingface.co/datasets/IndexTeam/Sandglass-Bench.

View on Hugging FaceReport an issue

Author / organization	IndexTeam
License (as declared)	mit
Downloads	202 (as reported by Hugging Face)
Likes	1
Created	3 October 2026
Last updated at source	3 October 2026

benchmarkevaluationsyllable-controlconstrained-translation

Embed this entry

<iframe src="https://silytics.si/embed/item/68aca4df-e3be-4deb-921e-516c3d635dde" 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

202 now. Counts as reported by the source, snapshotted daily by Silytics.

More from IndexTeam

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 Information may be incomplete or out of date.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.