Data Science Investment Counter —Funding raised by UK data science companies in 2018.
£ 5.640 Million

Facebook releases new, open-source version of its PyTorch AI framework

Facebook has launched a developer preview of its long-awaited PyTorch artificial intelligence (AI) software framework, the company announced on Tuesday.

The Python-based tool, which helps accelerate the deployment of AI-based applications, provides developers with the power to seamlessly move from the research phase to production using a single framework.

PyTorch 1.0 integrates PyTorch’s original research-oriented aspects with the production-focused capabilities of Caffe2, a popular deep learning framework, and ONNX (Open Neural Network Exchange), an open format to represent deep learning models.

“Today, we’re pleased to announce that engineers on Google’s TPU team are actively collaborating with core PyTorch developers to connect PyTorch to Cloud TPUs,” Google Cloud director of product management Rajen Sheth wrote in a blog post.

“The long-term goal is to enable everyone to enjoy the simplicity and flexibility of PyTorch while benefiting from the performance, scalability, and cost-efficiency of Cloud TPUs.”

PyTorch is already used in many of Facebook’s products. The best-known example is Facebook Translate, using AI on neural networks to perform billions of translations a day.

Since Facebook open-sourced PyTorch, the project has gained many supporters. Tech giants like Amazon, Google, and Microsoft have picked up the new version of PyTorch, as well as technology providers ARM, Intel, IBM, NVIDIA, and Qualcomm.

“Data scientists and machine learning engineers have a wide variety of open source tools to choose from today when it comes to developing intelligent systems,” Information Services Group principal analyst Blair Hanley Frank commented for Fortune.

“This announcement is a critical step to help ensure more people have access to the best hardware and software capabilities to create AI models”, he said.

“Expanding framework support can help cloud providers like AWS, Google and Microsoft drive additional usage of their platforms. That means it makes sense for them to support as broad a set of development tools as possible, to try and attract as many customers as they can.”


Co-working space and blog dedicated to all things data science.

Subscribe to our newsletter