Description
Please check the instructions for use!
Recommended to pair with the large model
I have created a QQ group for sharing Lora's generated images and illustrations. Everyone is welcome to join: 817686141
The image generated by Lora is a Chinese-style female character with good composition and lighting effects. It is also a simple feeding-style Lora. Adjusting details and adding this Lora can be used. Add More Details This Lora, especially effective at special resolutions 1!!!!!!
After simplifying Lora, it can easily produce good images and can generalize well between realistic and anime-style large models. Based on the style, I have integrated elements , so Lora is foolproof for models; a simple Lora and trigger words can produce excellent results. If you need to change details, you can refer to my example images.
If you don't know how to write tags, it is recommended to modify based on my example images.
Leaning against the dangerous building, feeling the gentle wind, looking far into the spring melancholy, dimly rising on the horizon. In the remaining light of the grass and smoke, no one will understand the thoughts leaning on the balustrade.
Intending to depict a carefree painting while drunk, singing to the wine, enjoying the music without taste. The belt gradually widens, and in the end, there is no regret, becoming haggard for her sake.
Some of my parameters:
Iterations: 20-50 VAE: vae-ft-mse-840000-ema-pruned
Sampling method: DPM++2M Kara DDIM
Remember to enable high-quality restoration!
Upscaling algorithm: 8x_NMKD-Superscale_150000_G or 4x-UltraSharp
High-res iterations: 15 Redraw magnitude: 0.15-0.3
Upscaling factor: 2 Prompt word guidance coefficient: 7-10
To achieve better facial details, enable ADetailer
Recommended front TAGs:
Recommended back tags: (badhandv4:1.2), (worst quality:2), (low quality:2), (normal quality:2), lowres, bad anatomy, bad hands, normal quality, ((monochrome)), ((grayscale)), Freckles, Too many hands and feet,