Skip to content

Latest commit

 

History

History
29 lines (22 loc) · 527 Bytes

README.md

File metadata and controls

29 lines (22 loc) · 527 Bytes

ConvNetCL4J

OpenCL Accelerated Deep Learning for Java using Aparapi

This is still WIP.

Developement has stopped due to better alternatives such as:

  • Google Colab
  • Kaggle Kernels

Currently done:
Convolutional Layer
Transpose Convolution Layer
ReLU Layer, LeakyReLU
SGD
SGD-Momentum
MSE Loss
Charbonnier Loss

High priority:
Add gradient checks and unit tests

Near future plans:
Other Rectifiers (ELU, etc)
Max Pooling Layer
Fully Connected Layer
Dilated Convolution Layer
Adam