search

26hzhang / VSLNet

Span-based Localizing Network for Natural Language Video Localization (ACL 2020)
MIT License
100 stars 17 forks source link

TypeError: int() argument must be a string, a bytes-like object or a number, not 'Tensor' #1

Closed onlyonewater closed 4 years ago

onlyonewater commented 4 years ago

Thank you very much for making the code open source.But I have a problem, when I use query_features = tf.transpose(query_features, perm=[0, 2, 1]) query_features = conv1d(query_features, dim=1, use_bias=True, reuse=reuse, name='dense'). There is an error as follows: TypeError: int() argument must be a string, a bytes-like object or a number, not 'Tensor'. I am very confused.

26hzhang commented 4 years ago

@onlyonewater are you using the same packages list in the instructions? I tested just now, it works. No such error reported. Could you provide more details of the error info?

onlyonewater commented 4 years ago

Thank you very much for your reply, I have solved this error. But I am confused that when I run the command python run_tacos.py --mode test, I got a higher result than in the paper. The results are as follows: Rank@1, IoU=0.3: 45.69 Rank@1, IoU=0.5: 32.67 Rank@1, IoU=0.7: 19.27 mean IoU : 32.30, So, What's wrong?