search

pybrain / pybrain

BSD 3-Clause "New" or "Revised" License
2.86k stars 792 forks source link

Reproducibility #170

Open tlikhomanenko opened 9 years ago

tlikhomanenko commented 9 years ago

Hi!

I found that pybrain training isn't reproducible, see http://nbviewer.ipython.org/gist/tata-antares/6e67100f6b67ee5bd225. The same training with even fixed random seed (numpy.random.seed) gives different results. Why this happened?

Also sometimes trained net gives different predictions for the same data if call predict twice.

Thanks in advance!

philipobrien commented 8 years ago

+1

franciscogmm commented 7 years ago

still an issue. :/