Are you perhaps referring to the time taken for each node to execute? It's a custom node called ComfyUI-Dev-Utils. You can install it from the ComfyUI Manager.
Thanks for your question! As described on the model card page, this is a fine-tuned version of Flux 0.1 dev, and you can find more details directly on the page.
@@CodeCraftersCorner yeah mate I read the model card, was just very curious about the aesthetic scoring, that would be entirely separate from the training of the base mode. I dont think its actually possible to score it based on any sort of 'aesthetics'
They did make a lora. It is available in the same Huggingface page. We have to keep in mind that the models are for non-commercial research purposes only. So, best guess would be that they were experimenting.
thx, but cannot see the link to the fine-tuned models in /under descriptions pls link them, also quite bright the site is not in dark mode in your videos - keep up the great work still! :)
They are improved !! they look flat and impersonal as a bad heavily photoshoped image like we can find in fashion magazines. They do this in fashion to somehow erase the model and force your attention to the clothing. If a model is too special with too original features it's not good for business.
Thanks!
Thanks for watching!
Looks great...its possible using it on Forge? thanks
Thanks! Not for now. Will have to wait for a compatible version.
@@CodeCraftersCorner ok...thanks for answering
How can you add time to your workflow? maybe a time for the shnell version? In particular for those starting with "load diffusion model"
Are you perhaps referring to the time taken for each node to execute? It's a custom node called ComfyUI-Dev-Utils. You can install it from the ComfyUI Manager.
Thanks for the content
Thank you Sharvin
Thank you for your efforts..
Please make a video on flux dev lora training
There are multiple videos on the AI toolkit. I'll see if I can bring something new.
Thanks for the tutorial. Can this model be also used for Forge webui or is it only for ComfyUI?
For now, only ComfyUI.
@@CodeCraftersCorner thank you 🙏
How was the aesthetic quality determined?, which model was used and how was it implemented?
Thanks for your question! As described on the model card page, this is a fine-tuned version of Flux 0.1 dev, and you can find more details directly on the page.
@@CodeCraftersCorner yeah mate I read the model card, was just very curious about the aesthetic scoring, that would be entirely separate from the training of the base mode. I dont think its actually possible to score it based on any sort of 'aesthetics'
Why they did not just make a lora?
They did make a lora. It is available in the same Huggingface page. We have to keep in mind that the models are for non-commercial research purposes only. So, best guess would be that they were experimenting.
thx, but cannot see the link to the fine-tuned models in /under descriptions pls link them, also quite bright the site is not in dark mode in your videos - keep up the great work still! :)
Thank you for reporting it. I fixed it. Let me know if you cannot access it.
@@CodeCraftersCorner thx boss! :)
Is this latest model? because I had tested privious model but it is too bad quality
Yes, this is the latest model we have for now.
They are improved !! they look flat and impersonal as a bad heavily photoshoped image like we can find in fashion magazines. They do this in fashion to somehow erase the model and force your attention to the clothing. If a model is too special with too original features it's not good for business.
Thanks for your comment! It’s great to hear a different perspective. I hadn’t thought of it that way.
Туда сюда повертел. Ну такое... типа матовость, типа шумы, смазы нелепые, ну разве что по свету интереснее, а так качество фигня
Yes, the quality can be affected in scenarios other than fashion photos. Thanks for your feedback!