V3 Model Overview
The V1 and V2 versions were my initial attempts to achieve what I believe is now fully realized in V3.
Keywords for Generation
You can generate the model with either just the dress or with brown pantyhose using the following keywords:
- Only Dress: s3x6_t1gh_bl4ck_dr3ss, black dress
- With Pantyhose: s3x6_t1gh_bl4ck_dr3ss, black dress, pantyhose, brown pantyhose
## Recommended Sizes
The model was primarily trained on two sizes:
- 896x1152
- 832x1216
It's best to start with these sizes, but feel free to experiment with others.
Weight Settings
For the best results, I recommend using a weight of 0.8, as it has provided the best outcomes in my tests. However, you can try values between 0.6 and 1 for potentially good results as well.
Feedback and Suggestions
If you like the model, feel free to leave a comment! Let me know if you find a particular checkpoint that works exceptionally well or if you have suggestions for improvements.
If you enjoyed this model, please share some buzz! I’d love to know if you'd like to see an improved version or a different model altogether. Your feedback will help guide future updates.
Description
Increased number of images in the dataset and tried to get more diversity to make it more versatile. I'm more satisfied with this model than my previous versions and it tooks much more works to achieve it.
This version require some trigger words to works fine.
For generation, I suggest starting from prefectPony with DPM++ 2M or Euler a and 20 to 30 steps.
For the weight of the model, I like 0.8 but between 0.6 and 1 you can get it working.
I'm waiting for comments about things that I can try to improve or suggestion for other checkpoint and settings, so that everyone can makes the best use of it.
FAQ
Looks like we don't have an active mirror for this one right now.
This archive is kept alive by community contributions, so there are still some gaps, especially for newly removed content.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.