Cuda?

[aarmn80]

Maybe wrong place to ask but is there any neataptic or neat-python version with cuda support?
is it useful in neat process (I mean like tensorflow neat has lots of parallel process or not)?

[ghost]

Why do you want to invent the wheel ? Tensorflow with Keras gives you supercomputer power easily. You would use projects like this maily to validate that your concept works on different frameworks with the same result.

[aarmn80]

@Pummelchen
You mean there is no use for a powerful neat and all of neat AIs are easily portable to things like pytorch keras and tensorflow?
Because based on what I know they are layer based but neat is more flexible in spec connections
(Im noob in AI, please don't blame me if I'm saying sth idiot😁)

[ghost]

In the end you decide which framework to use, but in commercial use, its Tensorflow/Keras in combination with Multi-GPU Tensorcore Nvidia cards. Frameworks like these here are used for validation and testing, to ensure your prediction model has no bugs related to a specific framework.

[aarmn80]

Ok. Thanks ☺️

[ghost]

You can learn more here:
https://machinelearningmastery.com/blog/