Apple Releases Turi ML Software as Open Source
Apple last week released Turi Create, an open source package that it says will make it easy for mobile app developers to infuse machine learning into their products with just a few lines of code.
“You don’t have to be a machine learning expert to add recommendations, object detection, image classification, image similarity, or activity classification to your app,” the company says in the GitHub description for Turi Create. “Focus on tasks instead of algorithms.”
From a desktop computer running macOS, Linux, or Windows, Turi Create allows users to apply several machine learning algorithms, including classifiers (like nearest neighbor, SVM, random forests); regression (logistic regression, boosted decision trees); graph analytics (PageCount, K-Core decomposition, triangle count); clustering (K-Means, DBSCAN); and topic models.
The software automates the application of the algorithms to a variety of input data, including text, images, audio, video, and sensor data. Users can work with large data sets with a single machine, Apple says. What’s more, they can visualize streaming data with the software, which requires Python version 2.7 (with version 3.5 support coming soon).
Once a model has been developed, developers can export them to CoreML, the machine learning library that Apple released earlier this year for mobile app developers targeting iOS and other Apple platforms, including macOS, watchOS, and tvOS.
No GPUs are required for model training or
inference, although they can be used to accelerate processing, particularly for deep learning approaches supported by the software. The software uses CUDA 8.0 and cuDNN 5 libraries to power the MXNet deep learning algorithm for image-based workloads.
Turi Create handles various aspects of the machine learning process. For example, if a developer is building a recommender, Turi Create will automatically choose which model is most appropriate for the developer’s data set. If using image classifier or image similarity functions, it brings a pre-trained convolutional neural network (CNN) to bear on the problem.
Turi Create is based on software Apple obtained with its 2016 acquisition of Turi, the Seattle, Washington-based machine learning software company that computer science professor Carlos Guestrin founded as GraphLab back in 2013. Apple reportedly paid around $200 million for Turi and its machine learning software and expertise back in 2016.
The Turi influence on Apple’s software first surfaced this June with Core ML, a new machine learning library that enables software developers to incorporate Apple’s machine learning technology into iOS apps.
— Carlos Guestrin (@guestrin) December 8, 2017
Core ML was originally designed to do just several tasks, such as detecting human faces, landmarks, and text. Apple relied on Core ML to power the new face detection function in the new iPhones unveiled in September. The high-end iPhone X, in particular, features the high-powered A11 Bionic chip, which includes a six-core CPU glued to a three-core GPU.
Apple’s release of Turi Create coincided with NIPS 2017, the 31st annual conference on Neural Information Processing Systems, which took place in Long Beach, California last week.