Marc J. Schmidt

You shouldn't let mysql generate UUIDs due to performance reasons (indices). Better build your own index friendly UUID schema and generate those in PHP. Use `` if you really want...

Indeed, but we haven't this change yet ported to Propel 2. When we port it we should also adjust the documentation then. @Divi, do you want to contribute those changes?

We have already a parser since years that support this, although especially Postgres and SQLite is only with v2 really stable. I've no idea why MySQL is still in v2...

Mh, the thing is: Setting the schema to "public" per default can be wrong, since "public" is not always the default schema in Postgres. We should actually find out in...

@jaugustin, would be cool :sunglasses:

Would love to see this merged. CMake is awesome and would make integrating asio as easy as it gets. :)

I did that and only got 1% improvement. See here the experiment with relu activation for all strided Conv layer: https://trainer.aetros.com/model/marcj/keras:all-conv/job/f1b5348da3f2d7879ebe4390a60d3f5f43d517ac Script diff: https://trainer.aetros.com/api/file-compare/marcj/keras:all-conv/92fcd671c6814c375edd404a65edc66c00ba5aec...marcj/keras:all-conv/f1b5348da3f2d7879ebe4390a60d3f5f43d517ac:allconv.py Here you can see the side-by-side...

Not so cool. When we don't know with what data the net has been trained, we can not compare the network with other ones. Thanks for contacting him/her, we will...

I can confirm, that using the original code (with the fix in https://github.com/MateLabs/All-Conv-Keras/pull/5) and removal of multi_gpu code reveals an accuracy over 81%. My best after 350 epochs using the...

This is because Windows is currently not supported, but we're working on it. In the next big release probably end of the week we release some fixes for windows issue...