It is trained on the same images as the previous LoRa, probably wont update the LoRa

Does basically the same thing only with more detail it works better at 512x1024 compared to the LoRa