QwenImage_2512_fp8

This model is an enhanced optimization model of Qwen-Image2512, which can be used for image generation on the client side and quick image generation in ComfyUI workflows.

Due to the horizontal line bug that has not been resolved in online client-side image generation, it is recommended to download it to your local computer for deployment and then generate images.

Recommended prompt words: in summer, light wind, Asian girl, sunlight on the face, long hair, light eyes, Luffy's straw hat covers half the sunlight, highlighting the messy beauty of the hair, lively, moist lips slightly open, film photo, bright sunlight, white light spots, large grain texture, Kodak film texture, blurry, overexposed, blue tone, mid-shot shooting,

Parameter configuration

Cfg range: 1-4

Sampler selection: euler a  DPM++ 2M euler ancestral

Scheduler selection: beta57, beta, etc.

Recommended step settings: 10-40 (adjust according to your own situation)

Recommended workflow:

Qwen_lora General Text-to-Image Photography Visual Workflow