Amorphant: Difference between revisions

Help

Revision as of 08:17, 1 November 2024

Amorphant is a polymath who loves creation. He wrote this article in the third person.

Amorphant makes virtual photographs of monster girls, trying to ride the line between unsettling and arousing subjects. These include, but are not limited to: slime girls, goblins, kobolds, plant girls, annelids, moth girls and ghouls.

Amorphant has a Patreon with adult content that you should go and visit right now.

While the majority of his work is not "AI generated", it is "AI assisted". His general workflow is:

Use Blender to design a basic 3D scene and character pose, minus textures and details
Export the scene as a mist pass depth map
Import the scene in ControlNet
Create and apply LoRAs as needed for texture and detail
Render the final scene using a diffusion model

Amorphant believes that AI tools should never have been marketed as replacements for artists. He recognizes that they're tools for artists, and are only capable of producing things of value in the hands of artists.

@@ Line 5: / Line 5: @@
 Amorphant has a [https://www.patreon.com/c/MonAmor '''Patreon'''] with adult content that you should go and visit ''[https://www.patreon.com/c/MonAmor right now]''.
-While his work is not "AI generated", it ''is'' "AI assisted". His general [https://sandner.art/how-to-render-blender-3d-models-in-stable-diffusion/ workflow] consists of using [https://www.blender.org/ Blender] to design basic 3D scenes and poses, minus textures and details (mist pass [[depth map]]), then using [https://github.com/lllyasviel/ControlNet ControlNet] to import them into [[Diffusion model|diffusion models]]. Texture and details are applied primarily via [https://medium.com/rendernet/demystifying-loras-what-are-they-and-how-are-they-used-in-stable-diffusion-acc1c8748606 LoRAs] created by him. Final rendering is done by custom diffusion models.
+While the majority of his work is not "AI generated", it ''is'' "AI assisted". His general [https://sandner.art/how-to-render-blender-3d-models-in-stable-diffusion/ workflow] is:
+# Use [https://www.blender.org/ Blender] to design a basic 3D scene and character pose, minus textures and details
+# Export the scene as a mist pass [[depth map]]
+# Import the scene in [https://github.com/lllyasviel/ControlNet ControlNet]
+# Create and apply [https://medium.com/rendernet/demystifying-loras-what-are-they-and-how-are-they-used-in-stable-diffusion-acc1c8748606 LoRAs] as needed for texture and detail
+# Render the final scene using a [[diffusion model]]
 Amorphant believes that AI tools should never have been marketed as replacements for artists. He recognizes that they're tools ''for'' artists, and are only capable of producing things of value in the hands of artists.