I finetuned Qwen3-1.7B to imitate original Z-Image text encoder. 21% less VRAM
https://redd.it/1t71hvm
@rStableDiffusion
https://redd.it/1t71hvm
@rStableDiffusion
Reddit
From the StableDiffusion community on Reddit: I finetuned Qwen3-1.7B to imitate original Z-Image text encoder. 21% less VRAM
Explore this post and more from the StableDiffusion community