search

xiezhy6 / GP-VTON

Official Implementation for CVPR2023 paper "GP-VTON: Towards General Purpose Virtual Try-on via Collaborative Local-Flow Global-Parsing Learning"
465 stars 82 forks source link

uncorrect output of the second stage #35

Open liangdashuang opened 1 year ago

liangdashuang commented 1 year ago

iTerm2 mxKOVj upper___14651_00 png___03304_00

refer to the above image, the orders are real_image, clothes, p_tryon, warped_cloth, warped_prod_edge and p_rendered, the p_tryon and p_rendered image are not correct. i am wandering what is the reason?

my test code follows the followings: python3 -m torch.distributed.launch --nproc_per_node=8 --master_port=4736 test_tryon.py \ --name test_gpvtongen_vitonhd_unpaired_1109 \ --resize_or_crop None --verbose --tf_log --batchSize 12 --num_gpus 8 --label_nc 14 --launcher pytorch \ --PBAFN_gen_checkpoint 'checkpoints/gp-vton_gen_vitonhd_wskin_wgan_lrarms_1029/PBAFN_gen_epoch_201.pth' \ --dataroot /home/tiger/datazy/CVPR2023/Datasets/VITONHD \ --image_pairs_txt test_pairs_unpaired_1018.txt \ --warproot /home/tiger/datazy/CVPR2023/GP-VTON-partflow_final_kaiwu/sample/test_partflow_vitonhd_unpaired_1109

KIRAN-MK-211 commented 1 year ago

What is the command you used for test_warping.py?

liangdashuang commented 8 months ago

thank you, i have solved the problem, it's a problem about synbn, please refer to #37