subject transfer / replacement are pretty neat in Klein (with some minor annoyance)
https://redd.it/1qswthz
@rStableDiffusion

The Bait - LTX2

https://redd.it/1qsukif
@rStableDiffusion

Is Illustrious still the best for anime?

The Lora I like is only available in Illustrious, and is working ok, but are there any other model worth using? Is it hard to train my own lora in these new models? (https://www.google.com/search?scaesv=e53e8bd492977bc3&sxsrf=ANbL-n4p7pqA0mdbBXqgxsiV9NSrTdZYQ:1769951858878&q=chidori+from+full+metal+panic&si=AL3DRZHCeZEsX3-CsvvLZIg5Ht46vC3MF3hh9KlL8p7W7vSo9J8FV1ZAFXHQHNwvagA8nCPN9qmOKvgE47u0G34SHDW4jwaTpRFNJyvEXpII0liu3dM6zlQJr3bynzNigP-0rdjnruxM&sa=X&ved=2ahUKEwjqpMW-sLiSAxX6hf0HHWOdLdMQyNoBKAB6BAghEAA&ictx=1)

https://redd.it/1qsya6n
@rStableDiffusion

Nayelina Z-Anime
https://redd.it/1qt4ygv
@rStableDiffusion

Ref2Font: Generate full font atlases from just two letters (FLUX.2 Klein 9B LoRA)

https://redd.it/1qw83f5
@rStableDiffusion

A few random images from our collection!
https://www.reddit.com/gallery/1qvqjb1

https://redd.it/1qvqk66
@rStableDiffusion

Z-Image workflow to combine two character loras using SAM segmentation

https://redd.it/1qwdl2b
@rStableDiffusion

Z-image lora training news

Many people reported that the lora training sucks for z-image base. Less than 12 hours ago, someone on Bilibili claimed that he/she found the cause - unit 8 used by AdamW8bit optimizer. According to the author, you have to use FP8 optimizer for z-image base. The author pasted some comparisons in his/her post. One can check check https://b23.tv/g7gUFIZ for more info.

https://redd.it/1qw05vn
@rStableDiffusion