Google Colab to Run: https://drive.google.com/file/d/1nJhEqgfKNhbd9VVZ-0LR2UechftB4z3i/view?usp=sharing
https://stable-diffusion-art.com/controlnet/
Run down of features:
https://civitai.com/models/9251/controlnet-pre-trained-models
How to enable multicontrolnet: (If Running Locally click restart UI after changing the settings and you should be able to see the same results shown in the video)
https://www.youtube.com/watch?v=22VGlk3pR_I
Since I am still new to testing this feature, it would be best to refer to the link above
Different image used as base image,
Image generated using prompt used for testing:
Result:
Original image used as base image,
Image generated using different prompt:
masterpiece, best quality, 1girl, beautiful detailed eyes, perfect face, beautiful detailed face, looking at viewer, amazing sharp focus, ultra detailed, (black hair:1.4), large breasts, bodysuit, wetsuit, scuba diver, green eyes, long hair, high ponytail, sleepy, sweat, breathing, soft skin, outdoors, poolside, afterglow
Negative prompt: (worst quality, low quality:1.4), (monochrome:1.3), (NSFW:1.4), 3d, text, frame, jpeg artifacts, grids, watermark, logo, username, text, flowers, particles, (missing fingers:1.3), bad hands,
Steps: 20, Sampler: DDIM, CFG scale: 8, Seed: 3934657044, Size: 448x576, Model hash: 6806d4c0f8, Model: anything3pruned, Denoising strength: 0.7, Clip skip: 2, ControlNet Enabled: True, ControlNet Module: depth, ControlNet Model: control_depth-fp16 [400750f6], ControlNet Weight: 1, ControlNet Guidance Start: 0, ControlNet Guidance End: 1, Hires upscale: 1.85, Hires steps: 20, Hires upscaler: Latent (nearest-exact)