Type | |
Stats | 1,777 36,303 26.5k |
Reviews | (339) |
Published | Aug 5, 2024 |
Base Model | |
Hash | AutoV2 EC6546C95C |




The goal is to create a Pony model that has bright, crisp, clean characters with beautiful painted backgrounds. Backgrounds have a slight blur to help the characters stand out.
I try to keep my posts Safe for Work (SFW) but this is a pony model and is capable of pony model things.
What it does:
If you're familiar with Pony models, they are very specific in their prompting and subject focus. I made this model to work the same way, but have more stability and a more refined rendering. It uses the same prompting and LoRAs that other PonyXL models use.
What it doesn't do:
It's not a general model. Like other PonyXL models, it's more designed to work with booru tags and mostly does images with people in them. While I have tried to train more landscape images as you can see in the previews.
What's coming:
Continued training for landscapes and backgrounds.
How to use:
You can experiment with different samplers. I prefer DPM++ 2S a Karras for consistency and DPM++ 3M SDE Exponential to get wild. Prompting for quality should use the score system like other PonyXL models. (score_9, score_8_up, score_7_up) in the positive prompt and (score_3_up, score_4_up, score_5_up) in the negative prompt. Otherwise your images will be very plain.
Alternately, you can use the Pony PDXL embeddings that I created for much easier use.
Discussion
The description says we should use “score_3_up, score_4_up, score_5_up” as negative. Is that really how it works?
I always thought the “up” tags represent a score range between 9 and the specified number, which would mean that good looking styles are going to the negative, so I would rather use “score_3, score_4, score_5”.
But now I’m unsure.