Write an article about
Flux is a great model, don’t get me wrong but there is one thing that gives flux a bad name which is the plastic skin and the flux chin. The fix had been to train and use various different LoRAs to enhance the skin but in my experience many of them work well when you generate the first image within 1024×1024 dimension. After that when you start to increase the size and upscale the image lot of the flux characteristics come back to haunt you.
Here is what I mean with the plastic skin.
But this week I learnt about a new model that introduces realistic skin details and textures that finally remove the glossy and oily skin that we were used to get out of Flux Dev/Schnell. The model is called SRPO by Tencent and is available on Hugging Face, it is massive though weighing in at 47GB.
Of course at happens in the AI space these days, a few minutes later we have distilled versions available as well:
The original and distilled version all are designed to run on ComfyUI and there are workflows (JSON) available in the repos along with the models.
I have to say that I went with the FP8 version by Wikeeyang and tested on my RTX4080 with 16GB VRAM.
If you use my RunPod notebook I have already updated the configurations for the Model Manager with the Original and FP8/16 versions on it. GGUF version is quite small so that’s more local use only so I don’t include it in the configurations of Model Manager.
Results with this new model you can compare below. Some of the images generated are with the same seed and prompt so the comparison is easily possible.
If you’d like to support our site please consider buying us a Ko-fi, grab a product or subscribe. Need a faster GPU, get access to fastest GPUs for less than $1 per hour with RunPod.io
.Organize the content with appropriate headings and subheadings ( h2, h3, h4, h5, h6). Include conclusion section and FAQs section with Proper questions and answers at the end. do not include the title. it must return only article i dont want any extra information or introductory text with article e.g: ” Here is rewritten article:” or “Here is the rewritten content:”
