![]() You can alter environments while retaining the core features of the place: lokitsar makes many variations of Whiterun from Skyrimĭiffusion Model: Private (but similar to Deliberate)ĬontrolNet Model: control_depth-fp16 (Depth) What is possible with ControlNet? You can copy any pose: ghostintheshell107 uses software Daz3D to create poses, and then applies ControlNet OpenPose on a model called RPG for amazing resultsĬontrolNet Model: control_openpose-fp16 (OpenPose) You can read more about it on the official Github repo and the academic paper. With ControlNet, you can choose exactly which parts to keep from the original image and which to ignore (practical uses right now are controlling poses and compositions, but over time we'll see many more models and use cases). With regular img2img, you had no control over what parts of the original image you wanted to keep and what parts you wanted to ignore. ![]() ![]() ![]() Think of Stable Diffusion's img2img feature on steroids. What exactly is ControlNet and why are Stable Diffusion users so excited about it? ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |