Hugging Face has announced a partnership with Microsoft to democratise machine learning through open source collaboration and make the Hugging Face machine learning platform accessible to Microsoft Azure customers. The company has introduced Hugging Face Endpoints on Azure, a new service to turn Hugging Face models into scalable production solutions.

Machine Learning is helping industries all around the world run more efficiently. For example, financial companies use text extraction to clean insights from company fillings, healthcare companies use image recognition to detect diseases, e-commerce companies leverage recommender systems to surface relevant products, and more. However, many companies have strict performance, security, compliance, and privacy requirements that require hosting models on controlled internal infrastructure and massive engineering efforts. This makes deploying and scaling production-grade ML models incredibly time-consuming and difficult, with most machine learning projects today never making it into production.

Hugging Face is working with Microsoft to solve the growing interest in a simple off-the-shelf solution that provides companies with the freedom and control to deploy any ML model on internal infrastructure. 

Hugging Face Endpoints on Azure is a simple, scalable, and secure solution to deploy Hugging Face models on Azure infrastructure powered by Azure Machine Learning Services within minutes and without a single line of code. Hugging Face Endpoints are available in public beta in all Azure Regions where Azure Machine Learning Services are available.

“The mission of Hugging Face is to democratise good machine learning. We are striving to help every developer and organisation build high-quality, ML-powered applications that have a positive impact on society and businesses. With Hugging Face Endpoints, we have made it simpler than ever to deploy state-of-the-art models, and we can’t wait to see what Azure customers will build with them,” said Clément Delangue, co-founder and CEO of Hugging Face.

“Hugging Face Endpoints take care of the most pressing issues when it comes to model deployment. With just a few clicks or a few lines of Microsoft Azure SDK code, you just have to select a model and a task type, and you can start predicting in minutes. For the first time, I don’t have to worry about infrastructure, or about scaling up and down. This brings a huge opportunity both for testing and production, and it becomes very easy to iterate quickly,” said Mabu Manaileng, principal AI engineer at Standard Bank.