User Activity
-
Posted
a comment
on
discussion General Discussion
on
CURRENNT
You missed the peephole weights (one from the cell state to each gate unit, i.e....
-
Posted
a comment
on
discussion General Discussion
on
CURRENNT
You can just add dummy targets (e.g., zeros) of the right size (one integer per frame...
-
Posted
a comment
on
discussion General Discussion
on
CURRENNT
I'm using it with CUDA 6.5. You need to apply a small patch in order to get best...
-
Posted
a comment
on
discussion General Discussion
on
CURRENNT
Hi Alex, The outputs of the bidirectional layer are the concatenation of the activations...
-
Posted
a comment
on
discussion General Discussion
on
CURRENNT
The input layer has no weights, thus the learning rate has no effect. Conceptually,...
-
Posted
a comment
on
discussion General Discussion
on
CURRENNT
I cannot reproduce this bug, sorry. If I insert "learningRate": 0 as a member into...
-
Posted
a comment
on
discussion General Discussion
on
CURRENNT
The targetStrings variable is not required, and in fact it has no effect in CURRENNT....
-
Posted
a comment
on
discussion General Discussion
on
CURRENNT
A BLSTM layer consists of 2 LSTM layers of the same size, i.e., a BLSTM of size N...
Personal Data
- Username:
- weninger
- Joined:
- 2013-07-08 19:03:38
Projects
This is a list of open source software projects that Felix Weninger is associated with:
- CURRENNT CUDA-enabled machine learning library for recurrent neural networks Last Updated:
