I have Created a Workflow of ComfyUI In which how to use AW-Portrait Model Fine tune Model Flux is used
So I’ve been playing with this AWPortrait-FL model lately—it’s a finetuned version of Flux.1-dev, specifically tweaked for portrait work. The key thing here is it was trained on nearly 2,000 high-end fashion photography shots, which means it’s got a sharp eye for composition, skin texture, and those little details that make portraits pop.
I didn’t expect it to handle lighting and facial structure this well, but here we are. It’s like the model absorbed all the best parts of professional fashion photography and learned how to replicate that style. The skin tones come out way more realistic compared to the base Flux model, and the overall framing just feels more intentional.
One thing to note: this is a ShakkerLabs release under DynamicWang’s work, and it follows the same non-commercial license as Flux.1-dev. If you’re curious, you can grab it from their Hugging Face page.
I dropped it into ComfyUI just like any other Flux checkpoint—no special setup needed. The results? Honestly, better than I thought. It’s not magic (you still need decent prompts), but for portraits, it’s a noticeable step up from the base model.
Anyway, if you’re into portrait generation, this one’s worth a look. Just keep the licensing in mind if you’re planning to use it beyond personal experiments.
