Fix noise in the generated output

xiezhy6 / GP-VTON

Official Implementation for CVPR2023 paper "GP-VTON: Towards General Purpose Virtual Try-on via Collaborative Local-Flow Global-Parsing Learning"

475 stars 83 forks source link

Closed sriramun closed 10 months ago

sriramun commented 1 year ago

Setting model_gen to eval mode. This ensures consistent behaviour across different batch sizes.

Solves the following issues: #36, #21, #7

chenyg59 commented 11 months ago

thanks!it works!