Sign inGet started
Dataset · Hugging Face

Shu1L0n9/CleanSTL-10

Dataset Card for STL-10 Cleaned (Deduplicated Training Set) Paper | Code Dataset Description This dataset is a modified version of the STL-10 dataset. The primary modification involves deduplicating the training set by removing any images that are exact byte-for-byte matches (based on SHA256 hash) with images present in the original STL-10 test set. The dataset comprises this cleaned training set and the original, unmodified STL-10 test set. The goal is to provide… See the full description on the dataset page: https://huggingface.co/datasets/Shu1L0n9/CleanSTL-10.

Author / organizationShu1L0n9
License (as declared)apache-2.0
Downloads236 (as reported by Hugging Face)
Likes2
Created5 June 2025
Last updated at source23 July 2025
semi-supervised-learningdeduplicatedstl-10
Embed this entry
<iframe src="https://silytics.si/embed/item/d4817931-94f2-4070-9636-a0f83e6d8963" 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

236 now (+14 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.

Recent changes

  • 16 Oct 2026 Description: Dataset Card for STL-10 Cleaned (Deduplicated Training Set) Paper | Code Dataset Description This dataset is a modified version of the STL-10 dataset. The primary modification involves deduplicating the training set by removing any images t → Dataset Card for STL-10 Cleaned (Deduplicated Training Set) Paper | Code Dataset Description This dataset is a modified version of the STL-10 dataset. The primary modification involves deduplicating the training set by removing any images t

Changes detected by Silytics between refreshes of the source.

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 Information may be incomplete or out of date.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.