Requires SDXL_vae.

Completely virtual portrait, the material is 2.5D cartoon image, recommended to open ADetailer to get better face effect.The source of image data used for training Checkpoint is 2.5D images generated by Blonde Sasaki SDXL Lora.

https://civitai.com/models/605133/blonde-sasaki-sdxlpony-lora

The v0.5 version is slightly better than v0.1 in terms of screen sharpness.

Thanks to the author lizardon1024 for the test data, you can visit his page. https://civitai.com/articles/5300

描述:

训练词语: