5 Commits
Author SHA1 Message Date
Mike Clark aae13f0126 NaN's with multiple pooling layers with MNIST clouds :(
Tried building a model with these ops, but with multiple layers of pooling I get NaN value after a while. I think there is a problem with an unstable gradient in the this layer.
2018-10-16 08:14:51 +08:00
PatWie e2fdad61d9 fix layer, when input is not a tensor 2018-10-04 14:08:40 +02:00
PatWie 8be1b9ccf9 add pooling, transpose convolution to layers.py and update example.py 2018-10-02 13:44:32 +02:00
PatWie ab8a28c922 edit readme and add project page 2018-09-24 14:20:36 +02:00
Patrick Wieschollek 5a9cd25aa9 cherry pick files for public release 2018-09-23 12:44:08 +02:00