Digital intelligence company ABBYY has announced a new major update for its cross-platform, open-source machine learning library NeoML. The platform enables developers to build, train and deploy machine learning models, and the new update brings support for the Python programming language, which is the top language for machine learning and AI.
The new framework also involves 5-10x speed improvements and 20+ new ML methods, including 10 network layers and optimization methods. NeoML supports Apple M1 chips, GPU on Linux-based machines, and Intel GPU, all of which means an expansion of addressable use cases and scenarios for the library. It also means developers can use the framework to build AI-powered applications and solutions.
The Popularity of Python
Python is used in various industries for tasks like automation, web development, scripting, web scraping, and data analysis. It is relied on by major companies like Google, Pinterest, Spotiffy, Dropbox, and many others.
Outside of the private sector, academia also uses it to teach students how to program. Python’s versatility is what gives it such a high popularity, and ABBYY’s new development further enables developers and companies to use NeoML to build, train and deploy models for object identification, classification, semantic segmentation, verification and predictive modeling.
With the new speed improvements, NeoML is one of the fastest machine learning frameworks available, offering up to 10 times faster performance for classical algorithms and up to 30% faster neural network training and inference than the previous framework.
When compared to the two top open-source machine learning libraries, NeoML offers 50% faster performance on average. Because of this, the framework is especially useful for customer-facing, cross-platform applications. NeoML’s high cloud efficiency means businesses can use available cloud resources in the best possible way.
Bruce Orcutt is Senior Vice President of Product Marketing at ABBYY.
“Open source is a powerful driver of technological innovation. We aim to support advancements in artificial intelligence by working together with the developer community to further grow and improve our open-source library,” said Orcutt. “NeoML opens new opportunities for developers allowing them to experiment, build and launch ground-breaking initiatives while taking advantage of the framework’s high inference speed, platform independence and support for mobile devices. We invite all developers, data scientists and academia to use and contribute to NeoML on GitHub.”
NeoML can process and analyze data in various different formats, such as text, image, video and more. Models can be applied in the cloud, on-premises, in the browser and on-device, and the library supports C++, Java and Objective C programming languages. It also offers 20+ traditional ML algorithms like classification, regression, and clustering framework.
The framework’s neural network models support more than 100 layer types, and the library is cross-platform, capable of being run on operating systems like Windows, Linux, macOS, iOS and Android, and it is optimized for both CPU and GPU processors.
NeoML is already being used by developers in the US, Canada, Germany, the Netherlands, Brazil, China, India and South Korea. The framework is available on GitHub.