The neural network models have been optimized. Users of 1.019 report a 10-15% increase in processing speed compared to earlier 1.0x versions, especially on Apple Silicon (M1/M2/M3) and modern NVIDIA RTX cards.
Early adopters on 4K and 5K monitors experienced tiny, unreadable UI buttons. Version 1.019 introduces proper high-DPI scaling, making the standalone app interfaces usable on high-resolution displays.