History

0.1.4 (????-??-??)

• Either a TFLearn optimizer object or a string key can be passed to an algorithm
• Fix replay memory bug where size would only reach max size - 1
• Improve discount_rewards efficiency with scipy.signal.lfilter
• TODO: improve session management so user doesn’t have to create session
• TODO: implement policy duplicate function, where the duplicated ops are in the same graph with a different scope

0.1.2/3 (2016-12-17)

• Improving meta data and fixing __init__ scripts to load subpackages properly

0.1.0 (2016-12-16)

• First release on PyPI.