Business only (hides war / politics / culture / sports)
Saturday, Sep 19, 2026
1 Developer Releases Full arXiv Archive of 3.1 Million Papers as 16 TB Hugging Face DatasetPVT:HGFC 🤖 AI Sep 19, 5:20 PM EDT 5/5
1
Developer Releases Full arXiv Archive of 3.1 Million Papers as 16 TB Hugging Face DatasetPVT:HGFC
🤖 AI Sep 19, 5:20 PM EDT 5/5
A developer uploaded a massive repository of scientific research documents to a public AI hub. The collection, hosted on Hugging Face, contains 3,148,796 papers in LaTeX, PDF, PostScript, and HTML formats, covering every version of the documents found on the arXiv website.
The total size of the upload is 16 TB, providing a centralized resource for researchers to perform large scale analysis or train machine learning models. The data was published on September 19, 2026, creating a consolidated version of the scholarly site for the artificial intelligence community.
You're reading an older version of the story.
User Releases First Complete 16 TB arXiv Corpus of 3.1 Million Papers on Hugging Face 5 tweets • 5 sources