Guide: Generating New Images with Amber - A Brief Overview
I apologize as this guide is rather succinct, rest assured, a more comprehensive guide, possibly supplemented with video tutorials, is in the works. For now, here are the essential steps to get you started:
-
Installation: Begin by downloading and installing the latest version of a1111 webui, and make sure you have the most recent ControlNet extension installed as well. Lastly, download the latest Rev Animated model from Civitai.
-
Getting Started: In the ‘original’ folder, you’ll find the initially generated images for each of Amber’s poses. To access the prompt, load the oldest image from the pose you want to copy the prompt from into the ‘PNG Info’ tab on the a1111 webui.
-
Choose Your Approach: From the PNG Info tab, depending on your needs, you can generate new images from text prompts by selecting ‘send to txt2img’, or you can modify specific portions of the image by choosing ‘send to inpaint’.
-
ControlNet Setup: Before you embark on generating new content, enable ControlNet (there’s a checkbox) and switch the mode to “Reference Only”.
-
Reference Image: Load one or more existing images of Amber into Controlnet. The system will use her likeness as a reference to generate new content that maintains Amber’s visual consistency across different poses and expressions. Note that this is not img2img, it’s ControlNet.
Remember, practice makes perfect. The more you engage with these tools, the easier it will be to generate dynamic and engaging content featuring Amber. Keep experimenting and happy creating!
Get Amber Character Sprite Pack - A Runner's Spirit
Amber Character Sprite Pack - A Runner's Spirit
| Status | Released |
| Category | Assets |
| Author | FieraRyan |
| Genre | Visual Novel |
| Tags | 2D, character-art, Characters, Sprites |
More posts
- v1.0 (Initial release)Jun 19, 2023
Comments
Log in with itch.io to leave a comment.
Great work! It brings back memories of the classic game Princess Maker :)
This is AMAZING! I wanted to know in detail how to make this type of images that are repeatable, i.e. not creating another model for each variant pose. Does this guide teach us how to a) make the same character continuous as a base with new modes or movements b) can we extend this to make this PG 13 or Rated R version (minus clothing) ?
Yes and yes, this is basically how I make different poses for the same character (without the need for a LORA). Now one limitation is that its not going to give a consistent look every time! Maybe 1 in 3 images it generates will have the same outfit and character. If an image looks great but has one or two things off (different shorts or random accessories like watches) then I still pick that image and use inpainting to make corrections.
To be honest this is not really a guide, more like very broad directions on what to do. When I have more time I’ll make a detailed guide.
Lastly, if you want to make the same character base with different outfits, simply inpaint the rest of the body while leaving the face intact, and make sure the inpaint area don’t deviate too much from the general body shape. You can also try the Inpaint Anything plugin which makes this easier.
For the rated R version, it will not be released here as this account is SFW only, but you’ll probably spot it when it gets released on a different account.