SD1.5 V1
Lora for stable diffusion 1.5
Use with a weight of 1
Compatible with many models
That's all, enjoy it and share your generations!!
PonyXl V1
Use it whit a weight of a 1 to 1.2
Do not use source_anime, source_cartoon tags,
The best performing samplers are: Euler a and DPM++2M kARRAS.
20 to 30 steps
Remember to use the triggerword for best results
Use any SDXL compatiblde resolution
Have a fun and share your results!
Description
V1
Use it whit a weight of a 1 to 1.2
Do not use source_anime, source_cartoon tags,
The best performing samplers are: Euler a and DPM++2M kARRAS.
20 to 30 steps
Remember to use the triggerword for best results
Use any SDXL compatiblde resolution
Have a fun and share your results!
FAQ
Comments (6)
Its so nice to see small loras!
Thanks!!, I found a way to get a smaller size, but still work the same as a 200MB one!!!
@WhiteZ please make a guide, are you making loras online or localy?
@Rating_Agent online using this: https://colab.research.google.com/github/hollowstrawberry/kohya-colab/blob/main/Lora_Trainer_XL.ipynb
@WhiteZ Thanks alot! Can you please also share what setting do you use in this collab
@Rating_Agent well it's basically how well the notebook is configured, but I use prodigy and the network dim and alpha at 4
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.