why only 1000+ images in voc_2007_train_first_split?

yanxp / MetaR-CNN

Meta R-CNN : Towards General Solver for Instance-level Low-shot Learning

https://yanxp.github.io/metarcnn.html

177 stars 23 forks source link

why only 1000+ images in voc_2007_train_first_split? #45

Open squirrel16 opened 4 years ago

squirrel16 commented 4 years ago

As far as I understand, the trainval list of voc2007 consists of 5011 images. If 5 of 20 classes are removed (for all three kinds of splits), there are still 15 classes, maybe 3000~4000 images are correct.
So my question is, how to generate the trainval list for the few-shot training? I'm looking forward to your reply.

buncybunny commented 3 years ago

Hey have you solved the problem? I'm having a same issue here :(