pytorch-vsumm-reinforce icon indicating copy to clipboard operation
pytorch-vsumm-reinforce copied to clipboard

Published 20 hours ago • verified publisher icon KaiyangZhou

Summary not generating!!!

Open bjameslondon123 opened this issue 5 years ago • 12 comments

I have Trained and tested the model but am not able to visualize it. In the summary2video.py there is some typeError popping up stating TypeError: 'KeysViewHDF5' object is not subscriptable.

I have a folder named Videoframes containing several jpg images labelled 000001.jpg and so on....

bjameslondon123 avatar Jan 30 '20 09:01 bjameslondon123

Hi, @bjameslondon123

Did you input a h5 path correctly?? and where line did you face error??

SinDongHwan avatar Jan 31 '20 03:01 SinDongHwan

Hey @SinDongHwan, Thanks for replying Yes i did give the correct results.h5 path, the error is appearing in the summary2video.py file, line 40, in key = h5_res.keys()[args.idx] TypeError: 'KeysViewHDF5' object is not subscriptable.

bjameslondon123 avatar Jan 31 '20 05:01 bjameslondon123

@bjameslondon123 did you solve this?

sylvainma avatar Apr 11 '20 18:04 sylvainma

Hi, I think it is caused by the difference between python2 and python3. Maybe you can rewrite the code 'key = h5_res.keys()[args.idx]' to 'key = list(h5_res.keys())[args.idx]' in line 40 at summary2video.py

pandababyer avatar Apr 20 '20 10:04 pandababyer

hi, kindly to help me am facing bellow error error: (-215:Assertion failed) !ssize.empty() in function 'cv::resize'

hayachiq avatar Feb 09 '21 20:02 hayachiq

hi, @hayachiq. Try to check a image variable, after read image. Maybe you didn't read images.

Good Luck

SinDongHwan avatar Feb 10 '21 00:02 SinDongHwan

thank you for replaying @SinDongHwan could you please help me where to find the images ? should they be downloaded with the dataset ?

hayachiq avatar Feb 12 '21 17:02 hayachiq

Hi, @hayachiq. Yes, You should download a original dataset(images). And maybe face a error about mismatching of image file names. So, you have to change a code this line Good Luck~!!

SinDongHwan avatar Feb 15 '21 00:02 SinDongHwan

Hi, @hayachiq. Yes, You should download a original dataset(images). And maybe face a error about mismatching of image file names. So, you have to change a code this line Good Luck~!!

I am unable to get from which folder to read images. Where they are placed and what needed to be changed in code.

sky-16 avatar Mar 05 '21 16:03 sky-16

Hi, @sky-16. before edit codes, check images file name.

SinDongHwan avatar Mar 06 '21 04:03 SinDongHwan

Hi, @sky-16. before edit codes, check images file name.

I have downloaded the dataset mentioned in #40. Is there any other dataset because in this one I can not find path for reading frames. Could you tell me in which folder to check for path for reading frames images? I followed the default instructions.

sky-16 avatar Mar 07 '21 04:03 sky-16

Hi, @sky-16. #40 is not original dataset. e.g. If you need to TVSum dataset, Go to TVSum dataset home page. And then Download original dataset. Others are same.

SinDongHwan avatar Mar 08 '21 00:03 SinDongHwan