ayankumarbhunia / on-the-fly-fgsbir Goto Github PK

Thank you for your contribution to the field of SBIR. I down ShoeV2 and ChairV2 from http://sketchx.eecs.qmul.ac.uk/downloads/ ，but the number of their sketch are 6648 and 1275. The number in your paper are 6730 and 2000. Where can I get your datasets? Besides, can I get your pre pretraining model? My e-mails is [email protected]. Thank you!

Excuse me , I wanted to get an image index sequence in the test so that I can look for photos in the file.Because I am repeating your work . I don't know how to realize in your code , so I look for your help . Thank you !

Hello, thanks for the great research you have done, the code provided in your Gtthub repository is not complete, can you provide the complete source code?

Getting error No Module named encoding....please help to resolve

where can i get chairV2 setdata？i just find QMUL-shoe-V2
thanks

Net_Basic_V1 line 5
import encoding
where i can get this？

In 2021, i trained the basemodel and RL, and i got the similar result. But now, After the system of server is reinstalled，i retrain basemodel and RL , m@A and m@B decreased by four or five percentage points. if i used the based model that trained last year, i can still obtain the similar result, I really can't find the reason.

we guess it may be the reason for the pretrained Inception v3, which is updated by model library of pytorch.

if you have the inception v3 model that you used, please send me ([email protected]) , thank you very much