Bring in comfyui

#1
by ryg81 - opened

Hey bring this to ComfyUI please.

Waiting for using it on ComfyUI.

The structure itself seems to be the same as Qwen-Image-Edit.
https://huggingface.co/cocorang/FireRed-Image-Edit-1.0-FP8_And_BF16

Already available for comfyui
https://huggingface.co/Comfy-Org/Qwen-Image-Edit_ComfyUI/blob/main/split_files/diffusion_models/firered_image_edit_1.0_bf16.safetensors

But hopefully there will also be a fp8 version, I am sure they are at it ;-)

VAE is causing the error, is it same as 2509 or 2511 workflow ?

brave_3ZsEqiumHY

VAE is causing the error, is it same as 2509 or 2511 workflow ?

brave_3ZsEqiumHY

got the same error using the vae from this repo. Just use the default qwen_image_vae

its bad model for me ... in all cases QWEN is better
only first impression...
at least for old or small photos dont work ... (input 256pix, output 1024pix) , with prompts like "keep the overall image. Repair the old photograph." (qwen is okay, oc FLUX2.dev is by far the winner)

its bad model for me ... in all cases QWEN is better

Found it to as good if not better in many cases here, and it also gives a different result/aesthetics, that makes it a real nice alternative to have ;-)
And their upcoming image model is for sure an interesting one in that respect

Same seed, same everything:

firered-image-edit-1-0-model-weights-are-released-v0-972l2gmo32kg1

Sign up or log in to comment