|
Framework |
Details |
 |
Developed by Google's research division in
Machine Intelligence. Interfaces: C++, Python |
| |
|
 |
Python
library integrated with NumPy; transparent use of GPU.
Developed at the University of Montreal. Interfaces: Python
Theano: a Python
framework for fast computation of mathematical expressions |
| |
|
 |
Developed by the community for scientific processing. Efficient
paralellization in both CPU and GPU. Interfaces: C, C++, Lua
Videotutorials |
| |
|
 |
Developed
by the Berkeley Vision and Learning Center with community support.
Interfaces: C, C++, Python, MATLAB.
Getting
started with the Caffe framework |
| |
|
 |
Extended version of Caffe. Interfaces: C++, Python
Caffe2 learning resources |
| |
|
 |
Written in Python and runs over TensorFlow or Theano, aiming at fast
prototyping, runs in both, CPU or GPU. Interfaces: Python |
| |
|
 |
Developed
by Microsoft Research, also known as the Microsoft Cognitive Toolkit. Interfaces: Python, C++, C#
Download code here |
| |
|
 |
Deep learning for Java
can import models written in Keras, TensorFlow, Caffe, Torch and Theano. |
| |
|
 |
Toolbox for MATLAB to implement CNN models |
| |
|
 |
Library for the construction of deep learning models using Theano.
Download code here |
