search

sangyun884 / HR-VITON

Official PyTorch implementation for the paper High-Resolution Virtual Try-On with Misalignment and Occlusion-Handled Conditions (ECCV 2022).
831 stars 171 forks source link

Full body outfit #9

Closed armangrigoryan96 closed 1 year ago

armangrigoryan96 commented 2 years ago

Hi, thank you for the great work. Just want to ask is it possible to train the model only with the custom full-body outfit? Or is it designed only for images where a person is not completely visible?

Jayyyyyyyyyyyy commented 2 years ago

look at the training dataset and the demo results, the trained model that auther provided supports up-body in a pic only. I tried full-body, the result is not very effective yet.

armangrigoryan96 commented 2 years ago

@Jayyyyyyyyyyyy thank you for your observation. I meant training on my custom dataset with a full body outfit, not using the pretrained model. So is it possible to get as convincing results as for the demo if I train myself on the custom images?

martacosano commented 1 year ago

Hello @armangrigoryan96 ! Did you train with your own full body images? Did you achieve good results? Thank you!

armangrigoryan96 commented 1 year ago

Hello @armangrigoryan96 ! Did you train with your own full body images? Did you achieve good results? Thank you!

Kindly ask you to email me at armg187@gmail.com for more details. Sorry, I am not checking the messages on the GitHub page very often, better email me.

koo616 commented 1 year ago

I think I can't confirm the performance because I haven't trained it with a dataset including the full body. The recently released DressCode dataset contains the full body, so I think you can use it to train the model and check its performance :)