I understand the frustration. I'm just getting into ANNs myself and the options are overwhelming. Trying to keep up with the state of the art is definitely a ch

I understand the frustration. I'm just getting into ANNs myself and the options are overwhelming. Trying to keep up with the state of the art is definitely a challenge. However, if you have a model that gives you the accuracy you need for whatever problem you're trying to solve, why not just stick with it? One of the nice things about using Neural nets these days is there's a lot of great software and infrastructure available for training and serving models.