This repository has been archived by the owner on Jul 1, 2024. It is now read-only.
forked from keras-team/keras
-
Notifications
You must be signed in to change notification settings - Fork 65
Bug in K.equal operator #146
Labels
Comments
Root Cause: we should use mx.sym.broadcast_equal() when x and y have different shape. Will fix together in pr #145 Reproducible code in mxnet:
Error Msg:
|
same applies to other element wise operators
|
This was referenced Jul 27, 2018
closing as fixed in #145 |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
The current K.equal(x, y) operator can not handle the situation when x and y have different shape with MXNet backend.
Reproducible code:
with TensorFlow backend output is:
with MXNet backend, error msg:
however, pure mxnet ndarray and symbol works fine:
produce
The text was updated successfully, but these errors were encountered: