search

facebookresearch / ov-seg

This is the official PyTorch implementation of the paper Open-Vocabulary Semantic Segmentation with Mask-adapted CLIP.
Other
689 stars 61 forks source link

demo.py #22

Closed jixiaofeng123 closed 1 year ago

jixiaofeng123 commented 1 year ago

Hi, thanks for doing such interesting work, I am very interested in this, I ran it on an online demo with my own data and found the results very exciting. However, when running the demo.py file, I found that the result is not as good as the online demo, and I can't realize the arbitrary semantic segmentation without prompting, may I ask how to get the relevant code of the online demo. Looking forward to your reply!

Jeff-LiangF commented 1 year ago

Thanks for your interests. This is a very good finding. By default, online demo is using SAM as proposal generator which has finer grained and higher quality masks then MaskFormer proposal which is used in demo.py. So, it is expected you have better results with online demo.

If you are interested in online demo, you can directly git clone from huggingface. The pretrained models are also available in that repo.

Thanks! Feel free to reopen the issue if you have further questions.