Issues
- 1
Implement padding in pooling layer
#165 opened - 1
Convolutional layer base implementation
#166 opened - 1
Pooling layer base implementation
#161 opened - 1
Pooling layer full implementation
#162 opened - 1
Support 3D input data (2D images with RGB channels)
#172 opened by beomyeol - 1
Convolutional layer full implementation
#171 opened - 1
- 7
Implement n_{fetch} and n_{push}
#92 opened by dongjoon-hyun - 1
Multi-node parameter server
#131 opened by jsjason - 0
Hide ParameterWorker.processReply
#158 opened by jsjason - 0
- 1
- 3
- 0
- 1
Grammatical errors in Matrix's javadoc
#167 opened by beomyeol - 1
- 0
Remove unnecessarily required num_input and num_output fields of LayerConfiguration in protocol buffer configurations
#146 opened by beomyeol - 0
- 0
Clarify how ParameterServerManager should be used
#159 opened by jsjason - 0
- 0
Support neural network models that have layers connected in a directed acyclic graph (DAG)
#153 opened by beomyeol - 0
Add a simple example for single node parameter server
#151 opened by bchocho - 1
Support batch processing of DNN
#141 opened by beomyeol - 4
Incorrect results of matrix multiplication and matrix concatenation of Nd4j library
#142 opened by beomyeol - 0
- 0
- 3
Upgrade to REEF-0.13
#138 opened by dongjoon-hyun - 0
- 1
Setup Travis CI
#134 opened by dongjoon-hyun - 1
- 0
- 0
- 0
- 0
- 0
- 0
- 4
- 0
Save final training/cross validation results to a file instead of simply logging
#114 opened by jsjason - 0
Well-defined configuration input format
#106 opened by jsjason - 2
DNN needs to support multiple dimension inputs.
#108 opened by beomyeol - 0
- 0
Allow parameter initializer to push parameter updates instead of activations and error gradients.
#113 opened by beomyeol - 0
Remove `ActivationAndDerivative` method
#116 opened by beomyeol - 0
Implement kernel_h and kernel_w in PoolingLayer
#107 opened by dongjoon-hyun - 1
Upgrade to use Nd4j (0.4-rc0 or above)
#105 opened by dongjoon-hyun - 4
- 1
- 0
#96 breaks build
#100 opened by jsjason - 1
- 0