You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, this repo still needs a lot of debugging. I just added "test.py" so you can test it yourself. However many parts are not yet tested or still need work.
CUDA/Parallelisation is definitely on my list for the future. However in it's current state, the C++ backbone is not ready to be parallelised easily. Due to the continously changing architecture of the network, the bottleneck is copying this architecture to the GPU. Until there's a solution to this problem, i don't think the use of the GPU will drastically improve performance.
I'm having some good results with it so far.
I enabled self.mutate() in neat.py.
I also plan to drill into attempt to get this to utilize cuda/gpu.
The text was updated successfully, but these errors were encountered: