JudyYe / zero-shot-gcn

Zero-Shot Learning with GCN (CVPR 2018)
MIT License
917 stars 178 forks source link

Interpret results #5

Closed michael20at closed 4 years ago

michael20at commented 6 years ago

So sorry for opening another issue, but I'd really like this to work! Ok, making progress now, downloaded img-3-hops.txt, tested it out on feature n01498406 (deleted everythin else), getting some results ['0.0', '0.0', '0.0', '0.9', '3.8'], but how to interpret them? How to get the corresponding labels printed or how can I look them up?

How can I give it a picture and get a result in text form? (say cat or dog for example) Sorry if this questions sounds redundant, I'd really like to test this out. I've read the code pretty thoroughly, but no luck so far :-|

JudyYe commented 6 years ago

Yes. Sorry for replying late.

The number in the list corresponds to topk accuracy in top_retrv.