search

dmlc / cxxnet

move forward to https://github.com/dmlc/mxnet
Other
1.03k stars 414 forks source link

Is it easy to implement a RNN? #46

Open byzhang opened 9 years ago

byzhang commented 9 years ago

Could you please add a simple example if possible?

tqchen commented 9 years ago

RNN was not yet implemented in cxxnet. However, it could be implemented with mshadow easily with matrix operation.

A notable thing about V2 is that we enables natural layer weight sharing by distinguishing layer and connections, this property could be helpful when building up an RNN.

byzhang commented 9 years ago

Do you have an estimation on when will the V2 be public?

tqchen commented 9 years ago

We are on final stages in testing things to make sure things are stable before pulling into master. You can already use the code in V2-refactor branch. On Saturday, March 7, 2015, byzhang notifications@github.com wrote:

Do you have an estimation on when will the V2 be public?

— Reply to this email directly or view it on GitHub https://github.com/antinucleon/cxxnet/issues/46#issuecomment-77713834.

Sincerely,

Tianqi Chen Computer Science & Engineering, University of Washington

byzhang commented 9 years ago

Do you mind to add a RNN example in V2-refactor branch? Thanks!

huashiyiqike commented 9 years ago

I have a RNN Library much like CXXNET, which is also built on top of Mshadow and also implemented LSTM and RTRBM layers. It is at https://github.com/huashiyiqike/NETLAB.