UAE AI model tops 50 million monthly downloads
Falcons AI's NSFW classifier ranks among top global open-source models
#UAE #opensource - An AI model developed by Ras Al Khaimah-based AI company Falcons AI has surpassed 50.8 million downloads in the past 28 days on Hugging Face, a leading global open-source AI platform. The Fine-Tuned Vision Transformer (ViT) model, designed to classify NSFW (Not Safe for Work) images, is currently the sixth most downloaded AI model out of over 1,400 models hosted on the platform. In comparison, Deepseek, one of the most talked-about AI models recently, recorded 3.7 million downloads in the same period. Falcons AI's success highlights the growing influence of UAE-based AI research in the global open-source community.
SO WHAT? - The rapid increase in downloads of Falcons AI’s open-source model underscores both community trust in the developer and the increasing global demand for robust AI-powered content moderation tools. Content moderation for online platforms and corporate environments is a big and growing use case which applies to a variety of domains from corporate governance, to media publishing and broadcasting, to regulatory usage. The new wave of Generative AI image models is only adding to the proliferation of user-generated content across digital platforms. Falcons AI has hit a sweet spot with its NSFW model, clearly appealing to a broad global audience.
Here are some key details about Falcons AI and the NSFW AI model:
A NSFW (not safe for work) classification model developed by UAE-based Falcons AI has been downloaded over 50.8 million times in the past 28 days (as of 14-Feb-25), making it one of the world’s most downloaded open-source AI models.
Headquartered in Ras Al Khaimah, Falcons AI provides AI-powered solutions for businesses and governments, focusing on automation, efficiency, and decision-making support. It also has an advanced open-source model programme.
Ranked sixth (out of more than 1,400 models) on Hugging Face in number of downloads over the past month, the NSFW model is based on the transformer architecture commonly used in natural language processing but adapted for image classification.
By way of comparison, the Falcons AI model has registered 12-13 times more downloads via the Hugging Face open-source community than all Deepseek AI models combine over the same period, despite the Chinese model’s massive global publicity.
The Falcons AI Fine-Tuned Vision Transformer (ViT) for NSFW images was trained on 80,000 curated images to distinguish between normal and explicit content, using a learning rate of 5e-5 and a batch size of 16 for optimal performance.
The NSFW Image Classification Model is available for download and use under an Apache 2.0 open-source licence.
ZOOM OUT - The UAE’s strategic investment in its AI ecosystem is paying off, with development from R&D labs, commercial AI ventures, corporates and independent experts starting to make an impact on the global stage. In addition to well publicised model releases from big UAE R&D labs and shining stars like the Falcons AI NSFW model, the UAE ecosystem is also earning a reputation for growing developer resources, communities and global collaboration.
LINKS
NSFW Image Classification Model (Hugging Face)
Falcons AI (Hugging Face)
Read more about open-source models from the UAE:
MBZUAI releases Arabic 'inclusive' multimodal AI model (Middle East AI News)
TII launches powerful SLMs under 13B parameters (Middle East AI News)
MBZUAI open-sources NANDA LLM (Middle East AI News)
TII launches Falcon's first SSLM (Middle East AI News)
TII debuts multimodal Falcon 2 Series (Middle East AI News)
Powerful open-source K2-65B LLM costs 35% less to train (Middle East AI News)