[
https://issues.apache.org/jira/browse/SINGA-65?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14730739#comment-14730739
]
ASF subversion and git services commented on SINGA-65:
------------------------------------------------------
Commit 98b499527012ab6b61fd854b39225b35d7370716 in incubator-singa's branch
refs/heads/tutorial from wangwei
[ https://git-wip-us.apache.org/repos/asf?p=incubator-singa.git;h=98b4995 ]
SINGA-65 Add an example of implementing user's own layers
The folder examples/mlp includes files for a user-defined layer, i.e.,
HiddenLayer class. The main.cc register this class. The myproto.proto defines
the configuration of this layer and extends it to the base LayerProto.
Makefile.example contains instructions for compiling and linking
user code.
job.conf specifies a simple model with only one hidden layer.
deep.conf configures a deep model with 6 hidden layers, which is slower
per iteration, but converges faster than the simple model.
> Add an example of writing user-defined layers
> ---------------------------------------------
>
> Key: SINGA-65
> URL: https://issues.apache.org/jira/browse/SINGA-65
> Project: Singa
> Issue Type: New Feature
> Reporter: wangwei
>
> SINGA's code has been modularized to support user extension.
> This ticket creates a full example of writing training using user defined
> layers.
> This example has the same neural net as the deep MLP example for digits
> recognition over the MNIST dataset. Instead of using the built-in
> InnerProductLayer and STanhLayer, we create a HiddenLayer as user-defined
> layer which applies inner-product and scaled Tanh transformations together.
> Makefile.example is provided to compile and link user code.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
