Skip to content

Latest commit

 

History

History
15 lines (11 loc) · 878 Bytes

README.md

File metadata and controls

15 lines (11 loc) · 878 Bytes

Visual Designer for Deep Neural Networks

The development of neural networks now becomes hassle-free. The neural networks designer provides a user-friendly tool for designing neural networks of various structures, activation functions, loss functions, and more. You create network models, we generate codes for you to train the models. The codes are optimized for GPU computing, compatible with GPUs from different vendors (e.g. AMD and Nvidia). The trained models can be easily embedded into C++, Python, Matlab, and Java programs.

Features

  1. User friendly visual design interface.
  2. Implemented with OpenCL language, so it supports a variety of devices ranging from CPU to GPU of both AMD and Nvidia.

Install

cmake . && make

Project development status

Under development, full features not yet supported.

Developers

Dihong Gong, Siva Prasad and Siliang Xia