Baidu develops open source deep learning framework for smartphones

Chinese Internet giant Baidu aims to get bigger in the world of artificial intelligence (AI) by launching its open source mobile deep learning framework. Called Mobile

Deep Learning (MDL), this convolution-based neural network uses the image sensor on Android and iOS mobile devices to recognise objects.

MDL’s neural network has low complexity and high speed. Interestingly, the new model is also designed to use the GPU on iOS devices to enable calculations through the network. This takes Apple devices a step ahead of the Android models.

A couple of demonstrations put up by Baidu on the GitHub repository show its different features. The first demonstration reveals that MDL is capable of recognising an object in less than a second. The second demo shows the project’s ability to identify an object from an image and find suitable matches from the Web.

While MDL does support developments made on UC Berkeley’s deep learning framework Caffe, Baidu recommends that developers use its PaddlePaddle platform that was launched last year. Users can convert their existing PaddlePaddle model to an MDL project using simple Python skills. Baidu has also given support for projects such as MobileNet, SqueezeNet and GoogLeNet.

MDL is available on GitHub under the MIT licence.