π“Ÿ host1.bitchan.one

𓀐 @YGCGrnZN

⏱ 3d seenβ›“ 0πŸ‘ 0β§– 1

filter [-]

languages:onother

it comes taste, preference and tolerance at this point. some people are easily pleased and settle for barebones effort. Not so much for me anymore since krea2 and also exploring api models when i get annoyed at the current dry state of local. I don't mindlessly mass download loras all kinds anymore. If a lora baker can't even bothered to caption the dataset and share the prompt and triggers, why should i download the lora and fight tooth and nail to get decent results? stop taking it personal.⏱ 3d↩ 1
three weeks ago a chinese lab released a model trained on NSFW, gore, (which has never been done before), without synthetic images, bbox support, and Comfy just released an 8mp lora for it. anons are training and posting loras for it and discussing what it can generate, and you say local is dry right now? are you living under a rock?⏱ 3d↩ 2