Image Moderation Dataset with Over 15k Images per Class

Over 15,000 images per class for training or evaluating image moderation models. The dataset is packaged in a ZIP file to preserve file structure integrity and contains two classes: SAFE (Safe for Work) and UNSAFE (containing sensitive content). It was created by author cloverxion and last updated on Hugging Face in April 2026.

Use Cases

• Training binary image classifiers based on the described SAFE and UNSAFE categories.
• Evaluating model performance on NSFW (Not Safe For Work) detection tasks.
• Building automated moderation systems for user-generated content platforms.
• Researching computer vision methods for identifying sensitive or violent imagery.

Strengths

• Contains over 15,000 images for each of its two classes, providing a substantial volume of data.
• Explicitly structured for a binary classification task (SAFE vs. UNSAFE), which simplifies model training objectives.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The description metadata is limited; actual data quality and class balance require manual inspection after download.

Provenance

Source

huggingface

Freshness

Last updated 2026-04-12 09:24:37; freshness should be verified.