search

shenyunhang / APE

[CVPR 2024] Aligning and Prompting Everything All at Once for Universal Visual Perception
https://arxiv.org/abs/2312.02153
Apache License 2.0
490 stars 29 forks source link

I use APE-L-D，but it doesn't perform as well as the demo #38

Open MingyuLau opened 7 months ago

MingyuLau commented 7 months ago

Could you share which model weight is used for demo?

shenyunhang commented 6 months ago

The demo uses APE-L-D weight. Different weights have different configuration files.

The detailed implementation of the demo can be found at https://huggingface.co/spaces/shenyunhang/APE_demo/blob/main/app.py.