Hinata is showing why you should not try to render bare feet with SDXL based models.
Added 2025-07-18 17:42:09 +0000 UTCSeriously, I've been wanting to diversify my content for ages, staying on foot-adjacent themes, but mainstream models still suck at feet. Even when boosted with specialized LoRAs, you're lucky if you can get one in ten picture with both feet having the correct number of toes.
That being said, I've acquired a GPU capable of running more powerfull models on my own machine. Now what's left is for me to dive into the arcanes of Open Source UIs to run said models, but it's easier said than done... Wish me luck!
