CodeReclaimers / neat-python

Python implementation of the NEAT neuroevolution algorithm
BSD 3-Clause "New" or "Revised" License
1.4k stars 490 forks source link

XOR example, evolve-spiking.py is not working. #38

Open evolvingfridge opened 7 years ago

evolvingfridge commented 7 years ago

Testing master branch, almost all examples are working on python 2.7.12, so far only issue with evole-spiking. First error I got was: <TypeError: eval_fitness() takes exactly 1 argument (2 given)> . Tried to hack it :) by adding argument to eval_func and got error: <AttributeError: 'tuple' object has no attribute 'node_genes'> I am trying to fix it, but so far all seems very painfully complicated, any tips on how to fix it or eta for a fix ?

CodeReclaimers commented 7 years ago

Yep, I still need to update the Izhikevitch genome implementation to work with the current framework. :) I'm hoping I can get to that over the weekend.

Thanks for testing with 2.7! I've been working in 3.5 for development, so it's good to know things aren't too bad in 2.x.

abrahamrhoffman commented 7 years ago

Using 2.7.12 myself. Let me know if I can help somehow. I'd be happy to aid in the overhaul effort.

CodeReclaimers commented 7 years ago

Just an update: the spiking example now runs, but doesn't seem to really ever make any progress towards a solution. I'm guessing there's probably something wrong with the neuron or network implementation.

evolvingfridge commented 7 years ago

Thank you, it will make much easier to figure out whats not working, if have time next will will dedicate some time for spiking, I find spiking are most interesting neural model.