Repository for experimenting with Neural-Net related things. For example various architectures, regularization methods, etc. Some of the things will be covered with a blogpost and some are just general improvements (e.g. dropout does not need an entire blogpost)
Yoray :)