v0.6.2

• some errors corrected
• added output classes for speeding up
• structure slightly changed
• net changed from doubles to floats
• LDA extension (experimental)

v0.6.0

• changed structure of class LSTM
• numpy arrays changed to memoryviews (a little bit faster)

v0.5.0

Basic release for first experiments