LCM Acceleration - LoRA and CN Control - Second-HD Restoration

Because there are nodes missing in the liblib online workflow in version V1, so some adjustments have been made. Mainly it involves a node to obtain the height and width of an image, which has been replaced by a built-in node in the online workflow for online generation.

For online generation, you need to select large models, LoRA, LCM (LoRA), CN models, etc. It is recommended to add them to the model library for easy selection. I have added notes where selection is needed.

If you download the workflow directly to your local machine and drag it into the interface, these models will still appear as the models used on my end.

The original intention of establishing the workflow was to allow devices with low VRAM to generate high-quality graphics. If you use the online workflow, the step of initially reducing image dimensions can be omitted, or the dimension values can be increased (I have also written a note next to it).

Welcome everyone to generate online! Currently, it's a bit expensive though~