Hugging Face acquires development platform XetHub

Recently for mobile devices uvula model SmolLM came to the fore with Hugging Facea teamwork development platform based in Seattle XetHubthat you bought announced. Founded by former Apple researchers, XetHub helps machine learning teams work more efficiently with large data sets and models.

Financial details of the deal have not yet been disclosed. However Hugging Face CEO’su Clem Delangueof As he stated; This is the company’s largest acquisition to date. With this acquisition, the Hugging Face team plans to integrate XetHub’s technology into its own platform. Thus, the company will upgrade the storage backend. In this sense, we should point out that developers can accommodate larger models and datasets than are currently possible with minimal effort. XetHub will close completely.

According to the company’s CTO Julien Chaumond; The XetHub team will be moving to their own better version of LFS as the storage backend for the Hub’s repos. In this way, Hugging Face datasets and models will start their next 5-year growth.

Details of XetHub

XetHub was founded in 2021 by Yucheng Low, Ajit Banerjee and Rajat Arya. The team previously worked on Apple’s internal machine learning infrastructure. The platform enabled businesses to explore, as well as understand and work with large models and data sets.

XetHub offered Git-like version control for repositories up to TBs. With XetHub, teams were able to track changes in their machine learning workflows and work together. Likewise, let’s add that XetHub’s repeatability provides protection benefits.

We can say that the initiative improves storage and transfer processes in this area. For this, Xethub used advanced techniques such as content-defined chunking, deduplication, instant repository assembly, and file streaming. The startup’s customers include Tableau and Gather AI.

Currently, the XetHub platform supports individual files larger than 1TB, while the total storage size of the platform reaches over 100TB. Xethub’s integration into the Hugging Face Hub will allow users to collaborate while working with these models and easily track their progress.