(Please do not download the first version anymore)

LZXL2.0 model has undergone approximately half a month of fine-tuning training in two versions, and the latest version 2.0 has been uploaded.

This version looks quite good, but there are still many aspects that need improvement. XL has high configuration requirements, and the fine-tuning training time is quite long. Fortunately, the basic model is excellent, and the final product is still quite good.

The keyword XL has a good effect on natural language in TAG, so everyone can try it.

1boy is a boy, 1man is a man

If you are using it for the first time, you can follow the suggested prompts below.

TAG Example: 1man, analogue, film, washedout, film, colour faded, 400 iso film stock, old-fashioned, mark, (recherche details), (ultra details), 8k resolution, excellent quality, beautiful cinematic lighting, engaging atmosphere,

Negative: (worst quality, low quality, cgi, bad eye, worst eye, illustration, cartoon), deformed, distorted, disfigured, poorly drawn, bad anatomy, wrong anatomy,

Recommended Size: 1024x768, 1280x768

Sampler: DPM++3M/2M Karras

CFG Ratio: 3-7 (BCFG is recommended when creating creative compositions, which may produce unexpected effects. The higher the configuration ratio, the more realistic it may be, depending on the prompts, sampler, and steps)

VAE: built-in