search

HumanAIGC / OutfitAnyone

Outfit Anyone: Ultra-high quality virtual try-on for Any Clothing and Any Person
5.26k stars 407 forks source link

How can I prepare for the data to test? #24

Closed PyBlackDev closed 5 months ago

PyBlackDev commented 6 months ago

Should I have to preprocess the pictures and clothes?

HumanAIGC commented 6 months ago

Not necessary. The model images are so far preset, so all you need to upload is clothing images.

PyBlackDev commented 6 months ago

demo demo5 Please help me to get the good result with this images

ghost commented 5 months ago

@PyBlackDev You don't need to stand in any pose. What non sense! They clearly mentioned to upload the clothes images with or without human. You cannot able to see on yourself. You kidding, buddy? Read carefully before you put your photo here.

Develop yourself or just wait till you become grand father to see the code running. They never release code. It's purely for research purpose and considering the malfunctions possibility, they never release to try on custom user images.

There are other repos and you stand in same pose to try your favourite garments. lol! \

ghost commented 5 months ago

For now, Enjoy your outfit with my result. Here, you go! image (21)

ghost commented 5 months ago

It's the result from their work. if I concentrate more, The result would be similar to their work. Well, Do you need same shoes?

If they release the code, many people would copy and develop same within days. It's simple.

PyBlackDev commented 5 months ago

Hi, @East91221 . Did you run this with your own code?

ghost commented 5 months ago

Yes, Dude. I followed the way they did their previous work.

ghost commented 5 months ago

The conditional Diffusion Model central to our approach processes images of the model, garments, and accompanying text prompts, using garment images as the control factor. Internally, the network segregates into two streams for independent processing of model and garment data. These streams converge within a fusion network that facilitates the embedding of garment details onto the model's feature representation. On this foundation, we have established Outfit Anyone, comprising two key elements: the Zero-shot Try-on Network for initial try-on imagery, and the Post-hoc Refiner for detailed enhancement of clothing and skin texture in the output images.

ghost commented 5 months ago

It's clear on how they followed. It become simple to replicate if they release the code. many AI developers would be replicating the same. They stick to their principles. They do NOT release as others would make money using for commercial purpose. they are done with the Try-On since many years which was NOT seen the light. They keep updating more features. That's All and they move on to other different development

ghost commented 5 months ago

@HumanAIGC Thankyou for still NOT releasing as others make money which is against to your principles.

PyBlackDev commented 5 months ago

Thank you, @East91221. I have tried with several repos from github but I was suffered from preprocessing image for input.

ghost commented 5 months ago

@PyBlackDev @HumanAIGC There are as many as 100's of repos related to virtual try-on. Some works great. Follow the paper of Outfit Anyone... You get best results. I assure you. Do NOT simply ask for code as they never release nor reply to unimportant queries.

PyBlackDev commented 5 months ago

Thanks a lot. Could you share how you did it?

ghost commented 5 months ago

@PyBlackDev Tryyyy yourself...All you need is free???? Good Bye!