UAE Falcons AI's NSFW Classification Model Enters Top Ranks on Hugging Face with 50.8 Million Downloads in 28 Days

The NSFW (Not Safe for Work) image classification model developed by Falcons AI, based in Ras Al Khaimah, UAE, has garnered attention in the global open-source AI ecosystem by recording 50.8 million downloads on Hugging Face over the past 28 days.

Download Metrics and Comparison

• The model ranked 6th among more than 1,400 models registered on Hugging Face.
• It recorded 12 to 13 times higher total downloads than DeepSeek models during the same period, suggesting a surge in global demand for content moderation and safety verification tools.

Technical Features

• Based on a Fine-Tuned Vision Transformer (ViT) architecture, it applies the Transformer structure, primarily used for natural language processing, to image classification.
• Trained on a curated dataset of 80,000 images, using a learning rate of 5e-5 and a batch size of 16.
• Released under the Apache 2.0 license, allowing free use by corporate and individual developers.

Implications

• With the increase in user-generated content (UGC) driven by generative AI, demand for automated content moderation in corporate governance and media platforms is growing.
• Falcons AI's success demonstrates the expanding influence of the UAE in AI research and the open-source ecosystem.