Vers2.0 I removed the Sam3 node and connected the "krea2 edit model" node from the first sampler. In the second sampler, the face improves even more. 3. sampler for details.
Hi everyone! I'm sharing my custom ComfyUI workflow, built on the Krea2 architecture, which combines identity editing and advanced painting/masking processes. With Image to Image, you can slightly alter an artwork you're inspired by to fit it into your desired identity. You can turn a pose that inspires you from an influencer into your own pose. It's an experimental process designed to create visually stunning compositions by blending realistic photographic textures or your preferred styles. I tried to avoid identity confusion on the face by masking it with SAM3.
Workflow Highlights:
Base & Model: Krea2 Jib Mix https://civarchive.com/models/2799984/jib-mix-krea-2
Text Encoder: I prefer the heretic one https://huggingface.co/DreamFast/Qwen3-VL-4b-Heretic-ComfyUI/tree/main
Masking & Inpaint: SAM3 (
sam3.1_multiplex_fp16.safetensors) Native NodeLLM Prompter: Krea's own text encoder Qwen3 VL
Optimization: Speed-optimized pipeline with Sage Attention support.
☕ Buy Me a Coffee https://buymeacoffee.com/ugurdoyduk
Happy generating, and I'd love to see your results in the gallery!
Description
FAQ
Comments (8)
I apologize, but I'm a bit confused. I replicated your exact setup - same model, same SAM3, same settings - and everything runs without errors. However, the output bears no resemblance to the original face. It creates an entirely different character with only a distant similarity to the source - nothing like your demo images. Perhaps there are some subtle tweaks for achieving likeness that I've missed?
İs the lora active?
@ugurdoyduk341 Yes, the LoRA krea2_identity_edit_v1_2.safetensors is active. But I'm using jibMixKrea2_v30Poblano.safetensors instead of your Krea2 jibMix.safetensors. Could that be the issue?
@Cali_Gulla I was using the same model as you and yeah the faces don't look anything similar.
If the faces are not similar at all, I don't know if you are experiencing a technical problem. But if you think it is less similar, set the ground pix setting to 1024 and the ref boost setting to 8. Put a collage of 3-4 different photos of your model in the "face" section.
@ugurdoyduk341Thank you. I will try to do that.
@Cali_Gulla @nikolaibloom805 I made a new version, you can try it.
@ugurdoyduk341 hello chief, i also can't get it working properly, even with the v2, the results are not looking like the face image at all :/ everything is as it should be, i don't know what to do.
also i don't know if it's intended like that, but the image comparer is not showing my base reference image, but 2 generated images, that have no likeness to my face image.