Distributed deep learning with Apache SINGA

Apache SINGA is now an Apache Top Level Project

Maika Möbus
Apache SINGA
© Shutterstock / Andrew Paul Deer

In October, the machine learning library Apache SINGA graduated from the Apache Incubator. Apache SINGA was built with a focus on deep learning and its features make it suitable for a variety of use cases—from healthcare to industrial application. Its maintainers also have some further projects for deep learning in mind.

Apache SINGA, a distributed, scalable machine learning library, has left the Apache Incubator. It all started in 2014, when SINGA was developed at the National University of Singapore in collaboration with Zhejiang University and entered the Apache Incubator the following year.

SEE ALSO: Apache Tika – “Data-driven analytics are at the heart of modern applications”

It is now the first distributed deep learning Apache project to become a Top Level Project (TLP), as Beng Chin Ooi, Distinguished Professor of National University of Singapore, is quoted on the Apache Software Foundation Blog. The initializer of the SINGA project and member of the Apache SINGA Project Management Committee further claims:

It is essential to scale deep learning via distributed computing as the deep learning models are typically large and trained over big datasets, which may take hundreds of days using a single GPU.

Let’s take a closer look at Apache SINGA’s features.

Deep learning with Apache SINGA

Apache SINGA was developed with a focus on deep learning and supports the training of large-scale machine learning models over a cluster of machines. The software stack includes three key components, the first of which is the core component that provides memory management as well as tensor operations. The second main component is the IO with classes for reading and writing data from or to the disk and network. The third is the model component, which offers data structures and algorithms for ML models.

As the Apache Software Foundation states in the blog post, Apache SINGA is currently in use at Carnegie Technologies, Citigroup, National University of Singapore and others. The use cases mentioned in the blog post range from healthcare applications like X-ray diagnostics or cardiovascular disease prediction to industrial application including e-commerce recommendation and smart fabric printing.

Further Apache SINGA plans

Although Apache SINGA has just left the incubator, the SINGA project committers are not sitting still. They are currently working on SINGA-lite, intended for deep learning on edge devices, as well as SINGA-easy for making AI accessible to domain experts without a deep AI background.

SEE ALSO: Big data database Apache Rya becomes a Top Level Project

The Apache Software Foundation Blog offers more information on Apache SINGA’s new status as Top Level Project and future plans. Apache SINGA is available in version 2.0.0 under the Apache 2.0 license.

Maika Möbus
Maika Möbus has been an editor for Software & Support Media since January 2019. She studied Sociology at Goethe University Frankfurt and Johannes Gutenberg University Mainz.

Inline Feedbacks
View all comments