2511 FP8 Please
We need the FP8 2511 model in half-precision
- 1
It's here with lighting fused inside (if you use fp8 and lightx2v lora separately, you will have lines artifactacts) https://huggingface.co/lightx2v/Qwen-Image-Edit-2511-Lightning/blob/main/qwen_image_edit_2511_fp8_e4m3fn_scaled_lightning_comfyui.safetensors
It's here with lighting fused inside (if you use fp8 and lightx2v lora separately, you will have lines artifactacts) https://huggingface.co/lightx2v/Qwen-Image-Edit-2511-Lightning/blob/main/qwen_image_edit_2511_fp8_e4m3fn_scaled_lightning_comfyui.safetensors
what evidence edit2511lightning is better than the fp8model+lightning lora?
It produces artifacts, vertical and diagonal white lines all over the image
The same was with Qwen Image and QIE 2508 2509. They are especially noticeable when you use loras
The same was with Qwen Image and QIE 2508 2509. They are especially noticeable when you use loras
you right here i encountered that.
so now i have to use the already fused with lightning lora full model from lightx2v ?