Multiple-instance-learning icon indicating copy to clipboard operation
Multiple-instance-learning copied to clipboard

Published 20 hours ago • verified publisher icon Epiphqny

what are the datas in 'img_tag.txt' and './data/zh_vocab.pkl' look like ?

Open kangkang59812 opened this issue 6 years ago • 6 comments
trafficstars

Could you give me some piece or files to show?

kangkang59812 avatar May 20 '19 10:05 kangkang59812

Could you give me some piece or files to show?

I have updated the README file, you can refer to that.

YuqingWang1029 avatar May 20 '19 12:05 YuqingWang1029

@Epiphqny thanks! By the way, in 'img_tag.txt, the number of tags is 1032? just as the last dim in your code. And are these tags the 1032-top hign frequency ?

kangkang59812 avatar May 21 '19 05:05 kangkang59812

I have made my own 'img_tag.txt' and './data/zh_vocab.pkl' , but maybe different with yours. Could you show me the first complete data in data_loader ? In my work, MIL is the first step, so I want to simplily realize it. for i, (images, targets, lengths) in enumerate(data_loader): thanks for your generosity!

kangkang59812 avatar May 21 '19 09:05 kangkang59812

realy need what's 'targets' looks like! thanks!

kangkang59812 avatar May 23 '19 05:05 kangkang59812

Hey @kangkang59812 have you figured out what it looks like? I also would highly appreciate knowing how img_tag.txt and zh_vocab.pkl look like...

aurschmi avatar Nov 09 '20 16:11 aurschmi

Hey @kangkang59812 have you figured out what it looks like? I also would highly appreciate knowing how img_tag.txt and zh_vocab.pkl look like...

No. I just did my work following the paper about MIML.

kangkang59812 avatar Nov 10 '20 06:11 kangkang59812