File size: 1,664 Bytes
6b3e119 d34dff0 6b3e119 d34dff0 82c03a6 d34dff0 82c03a6 d34dff0 82c03a6 d34dff0 82c03a6 d34dff0 2391bc2 82c03a6 8a0e55e |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 |
---
license: openrail++
language:
- en
tags:
- anime
- image generation
- PonyXL
- Stable Diffusion
- SDXL
---
## MikuFactoryXL
## About
This is my first PonyXL merge, a culmination of somewhat randomly selected models from CivitAI. The results, in my opinion, are really good (though admittedly, I'm biased!). I focused on models that have:
## Model Selection Criteria
* **Sharp Small Details:** Models with a knack for intricate intricacies.
* **Natural Aesthetics:** Minimizing the "AI slop" look, prioritizing clarity etc.
* **Artifact-Freeish Quality:** Selecting models with clean examples and minimal artifacts.
* **Personal Appeal:** A subjective but crucial factor in my choices.
* **Vibrant Colors:** Pretty color other than gray waow.
## Observations
* **Great Quality:** Visually nice and refined.
* **Excellent Prompt Following:** Interprets prompts really accurately compared to others.
* **Generally Strong Anatomy:** Occasional messes up, but mostly accurate.
* **Creative Posing:** Generates creative poses.
* **Overall Creativity:** Seems really creative in a unqiue way.
* **Detailed Faces ("Solv"):** Expressive facial features, less dead pan AI look.
* **Fine Small Details:** Adds intricate details a lot, usually less blurred than others.
* **Thin Line Art:** Nice thin lines compared to some models.
* **Good Conceptual Understanding:** Seems good at understanding concepts.
* **Computer Almost Crapped itself:** Needed 200GB swap file because of ComcfyUI, details can be found in `mega-merge.json` |