Google Cloud is partnering with open-source code and model platform Hugging Face for open-source AI software development using Google’s cloud infrastructure.
Developers will be able to use Google’s technical infrastructure, including its in-house AI chips – or tensor processing units (TPUs) – as well as the technical integration required to use the Nvidia H100 GPUs, which is expected to be completed in weeks, to access Hugging Face’s open-source AI software including models and data sets.
Google and Hugging Face have reached a revenue-sharing agreement for the partnership, but terms have not been disclosed.
In February of last year, Hugging Face signed a similar deal with Amazon Web Services. At the time this made AWS the platform’s preferred cloud provider, though it noted it would work with other cloud platforms in the future.
According to the companies, the partnership was motivated by a market need to create AI software that is optimized for specific tasks.
Read more: datacenterdynamics.com