AI development platform, Hugging Face, has collaborated with third-party cloud services including SambaNova, launching a feature called Inference Providers. Aimed at streamlining the process for developers on their platform, it allows AI models to run on a preferred infrastructure.
Fellow partners collaborating on this venture are Fal, Replicate, and Together AI. Hugging Face asserts that these partners have integrated their data centers into Hugging Face’s platform, creating seamless access for running various models. As a result, developers can effortlessly launch a model on SambaNova’s servers via their Hugging Face project page.
Although it provided an internal solution for operating AI models, Hugging Face revealed via a recent blog that its focus has pivoted towards collaboration, storage, and model distribution.
As serverless providers began to prosper, Hugging Face felt the need to facilitate straightforward access to their services. Consequently, they joined forces with next-generation serverless inference providers, primed for model-centric, serverless inference.
These serverless inferences empower developers to deploy and scale AI models without having to manage any hardware. Service providers such as SambaNova spontaneously allocate the necessary computing resources based on use.
Regarding the cost, developers utilizing third-party cloud providers on Hugging Face’s platform will pay the standard provider API rates. Depending on future potential revenue-sharing agreements with providers, this standard may change, says Hugging Face. Hugging Face Pro subscribers also receive an additional $2 monthly credit.
Starting as a chatbot startup in 2016, Hugging Face has grown into a global AI model hosting and development platform, successfully raising approximately $400 million from investors such as Salesforce, Google, Amazon, and NVIDIA.
Original source: Read the full article on TechCrunch
