FastestDemon

a year ago

Hey, thats awesome! Can I suggest some things? 1. Will you be adding new models? I personally would really like to see rouwei 0.8 (either eps or vpred version, preferably both but I understand that it may be difficult to maintain both) 2. In the default negative prompt for anime models (hassuka and animagine) you are using "censor" which isn't a proper danbooru tag, it should be "censored" 3. When listing models can you tell us which version it is? for example whether it is animagine 3.1 or 4? Same with other models. Thanks for the awesome website!

Joul

a year ago

Thanks a lot for your feedback FastestDemon! Yes, we're planning to add new models soon, both image and video, as we work on expanding our GPU capacity. I’ll check the Rouwei 0.8 models and review the license. I’ve fixed the negative prompt on Hassuka and Animagine, thanks for spotting that. Good point about the model versions. I already store them in the database, but they’re not shown on the site yet. I’ve added that to the to-do list.

Joul

a year ago

Hello FastestDemon, I have added model version on model selector.

Twinwaffle

9 months ago

Hi, the GPT model errors out for me every single time.

Joul

8 months ago

Hi, GPT model use an external API out of gen-image and it is very picky on nsfw content or IP content. I will add soon a check on prompt before sending it to the API