Baidu puts open source deep learning into smartphones

Computer vision, deep learning, and the camera in your phone

A year after it open sourced its PaddlePaddle deep learning suite, Baidu has dropped another piece of AI tech into the public domain – a project to put AI on smartphones.

Mobile Deep Learning (MDL) landed at GitHub under the MIT license a day ago, along with the exhortation “Be all eagerness to see it”.

MDL is a convolution-based neural network designed to fit on a mobile device. Baidu said it is suitable for applications such as recognising objects in an image using a smartphone's camera.

The neural network's calculations are offloaded to a phone's GPU, the company says at its repo, with high speed and low complexity. While it runs on either iOS or Android, the project's documentation leans more towards Apple than Google, with GPU support still an Android to-do, along with TensorFlow model support.

Baidu says the code fits in around 4 MB with no third-party library dependencies (beyond the neural networking models). The developers recommend the company's PaddlePaddle model for use with MDL, but it can also use Caffe.

The screen shots below show MDL can recognise a smartphone in less than half a second.

Baidu AI identifies phone in pic

Another MDL demo identified the pattern on a china teacup, and used that as in a search to find matching products for sale.

Baidu AI matching patterns on china

MDL supports MobileNet, SqueezeNet and GoogLeNet. ?


Biting the hand that feeds IT ? 1998–2017

<acronym id="haujiCA"></acronym>
<tr id="haujiCA"><optgroup id="haujiCA"></optgroup></tr>
<acronym id="haujiCA"><small id="haujiCA"></small></acronym>
<acronym id="haujiCA"><small id="haujiCA"></small></acronym>
<acronym id="haujiCA"><optgroup id="haujiCA"></optgroup></acronym>
<tr id="haujiCA"><optgroup id="haujiCA"></optgroup></tr>
<acronym id="haujiCA"><optgroup id="haujiCA"></optgroup></acronym>
  • 397531913 2018-01-24
  • 792411912 2018-01-24
  • 235953911 2018-01-24
  • 51841910 2018-01-24
  • 439944909 2018-01-24
  • 593438908 2018-01-24
  • 378341907 2018-01-24
  • 947792906 2018-01-24
  • 188953905 2018-01-24
  • 6206904 2018-01-23
  • 722120903 2018-01-23
  • 3141902 2018-01-23
  • 352694901 2018-01-23
  • 219550900 2018-01-23
  • 49328899 2018-01-23
  • 839992898 2018-01-23
  • 887926897 2018-01-23
  • 468948896 2018-01-23
  • 491400895 2018-01-23
  • 643446894 2018-01-23