Reproducibility issue

[tlikhomanenko]

All of neural nets (pybrain, neurolab, theanets) have problems with reproducibility:

• pybrain Reproducibility pybrain/pybrain#170
• neurolab Add random_state zueve/neurolab#26
• theanets Add random_state to Dataset lmjohns3/downhill#3

Meanwhile there are some dirty hacks with changing global seed, which is awful. (And they don't work for pyBrain, so things there are totally irreproducible)

[arogozhnikov]

there is one more problem, which is common for all NNs - they aren't supporting weights, but in theanets this will be probably resolved.

[tlikhomanenko]

In theanets reproducibility issue is fixed in 0.7.3 by adding rng during Downhill.Dataset creating https://github.com/lmjohns3/theanets/blob/v0.7/theanets/graph.py#L327 and this fix lmjohns3/theanets@94662f5.

This is shown in lmjohns3/theanets#102.