[
https://issues.apache.org/jira/browse/SINGA-218?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15413851#comment-15413851
]
ASF subversion and git services commented on SINGA-218:
-------------------------------------------------------
Commit dfc422e5b6229de5f598ee3f0226f1a0d082eb16 in incubator-singa's branch
refs/heads/dev from [~wangwei.cs]
[ https://git-wip-us.apache.org/repos/asf?p=incubator-singa.git;h=dfc422e ]
SINGA-218 Implementation for RNN CUDNN version
Add an example using the char-rnn model.
The trained model (with 2 stacks of lstm) over linux kernel source code
could generate source code with some meaning full patterns, e.g.,
indention, comments, variable definition, assignments.
> Implementation for RNN CUDNN version
> ------------------------------------
>
> Key: SINGA-218
> URL: https://issues.apache.org/jira/browse/SINGA-218
> Project: Singa
> Issue Type: New Feature
> Reporter: ZHAOJING
>
> (1) cudnn rnn implementation (cudnn_rnn,h, cudnn_rnn.cc, rnn.cc, rnn.h,
> test_cudnn_rnn.cc).
> (2) The weight shape now are manually calculated instead of using API
> provided by CUDNN.
> (3) Test for RNN_cudnn_Tanh (unidirectional, 1 hidden layer).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
