Recently nvidia released a tensor extension for the webui
It would be cool if Andrew made a tutorial on how to make RT engines and convert loras. Personally using an SD1.5 model i managed to get on a batch size of 4 images on a 3090RTX images of 512×768 and using the LCM lora in only just 3 seconds! It would be cool if Andrew made a detailed tutorial on how to use those various settings when creating the engines and explain more in-depth thanks.