Dain-App 0.46 (Patreon)

This version have the default options "Half-Precision" and "Benchmark" turned on.

I was able to fix some problems with the memory and now the application can get some really benefits from Tensor Core. Just keep "Half-Precision" and "Benchmark" turned on to use the mixed-precision model.

Using Half-Precision + Benchmark + "Cartoon or anime: Media with little or no depth" my computer could do a batch = 6 on a 578X326 media file at a pretty good speed.