Startup Hugging Face, which runs a program to host open-source artificial intelligence projects, made an announcement on Thursday that it acquired $235 million in a funding round.
The fourth stage of seed funding saw participation from tech company’s leading players including Google LLC, Amazon, IBM corp., along with Qualcomm Inc., Intel Corp and Nvidia Corp.
This investment is going to make Hugging Face value $4.5 billion, which is an increase since its past funding round.
Hugging Face handles software of a similar name which developers utilize to share open-source AI models. Its interface is just like GitHub. The Hugging Face page for an AI model is sectioned into three major parts: One consists of a brief description of this model, the second provides entry to its code and third consists of a discussion forum.
Developers count on Hugging Face for sharing AI models as well as inter-related technical assets. According to the start-up, its platform can host several training datasets which organizations can use to make brand new neural networks. They also provide access to Spaces, a sample application which demonstrates the practical use for AI models on the platform. For the most part Hugging Face hosts over 120,000 AI models along with 20,000 datasets.
The organization earns money by giving a paid version of the software called Enterprise Hub. The paid version lets organizations make an inside Hugging Face environment for the tech team. Developers will be able to use this environment for allocating AI models and linked assets with employees, along with access to files from public variation.
Other than Enterprise Hub, the organization also offers two more commercialized products which focus on making AI model development easy.
The first product is AutoTrain, customers can upload a training dataset on AutoTrain and it will find the most relevant open-source AI model for its project. After choosing a neural network, this service trains on the user-provided information.
AI models made with AutoTrain which are uploaded by developers to the Hugging Face Service are called Inference Endpoints. It will provide a systematic cloud environment to host neural networks. This service will spare users the hard work needed in assembling AI software infrastructure.
Clement Delangue, CEO at Hugging Face stated that the company will be using this funding for “more open-source AI and platform building.” They will increase their 170-employees to cultivate the effort.