Announcing Clandestine XL 1.0
Yesterday we launched yet another opinionated and stylized
#AnimagineXL
V3 fine-tuned model. The model is available to use and download on the Yodayo AI Model Hub platform and in a separate repository in our Hugging Face organization. The art style
Comparing Stable Diffusion 3 Large, Ultra and SD3M.
> {style*} best quality, trending on art station, looking at viewer, 1girl has a panel which is written [AICU], V-sign , in Shibuya crossroad
Today I hit 77 images per seconds at 512x512 with my pipeline, stable-fast, and sd-turbo. I've been doing realtime sd since hours after LCM came out. Coined the term RTSD a month ago.
100 fps here we come!
Actual gen time is <13ms but the display slows it. Will optimize more.