lvapeab/abivirnet

Attention Bidirectional Video Recurrent Net

abivirnet's Issues

About generate_descriptions_lists.py

Dear Author,

Because I don't know the exactly format of the caption.id.en used in the generate_description_lists.py script. I can only speculate about the outputs format of 'train_descriptions.txt' and 'train_descriptions_counts.npy'.

Here is the speculative results.
'train_description.txt': each row is all the descritption of a video, like:
'a dog is running' 'a dog is walking' 'a dog in the ground'
'a cat is eating' 'a cat is drinking'

'train_descriptions_counts.npy'' : each row is the description number of a video, like:
3
2

Is the speculative results right? Could you please upload some description example of files used in your code, such as: caption.id.en, train_list.txt, train_descriptions.txt, train_descriptions_counts.npy and so on?

Looking forward for your reply.

Thank you!

For Single video as an input

hi
After training, how we have to give input as single video and generate output, where we can store result

running error about main.py

Dear Author,

It occurred a problem "KeyError : 'references' " in code "gts = extra_vars[split]['references']" in file evaluation.py when I ran main.py. After a long time for training, the programme stooped here.

I would like to know why this problem occurred and how can I fix it? On the other hand, how could I skip training stage directly run evaluation stage? Because every time I modify the code, I need to run main.py and waiting for training, although the training is correct.

I am new to this area, and looking forward to your apply.

Thank you!

Can you provide the data to be prepared? I'm really hard to find.

Hi，
Great work, thank you so much for sharing the code.
Can you provide the data in here? It's really hard for me to find the data.

As follows:

./Images
    video_[video_id]
        [num_image].jpg
        [num_image].jpg
./Annotations
    test_list.txt
    train_list.txt
    val_list.txt
    captions.id.en
./Features
    test_[name_feat].csv
    train_[name_feat].csv
    val_[name_feat].csv

I am a beginner, looking forward to your reply.
Thank you very much for your help.

Where from to download the data and generate features to be used in the code?

I have already downloaded the MSVD dataset from http://www.cs.utexas.edu/users/ml/clamp/videoDescription/
but its format is nowhere similar to the one to be inputted in the subsample features.

wish to make some changes: different output & no pre-determined images features

Hey!
Great work, thank you so much for sharing the code.

I have a huge dataset of videos with a grade between 0 to 1 for each video.
This main thing is that there aren't annotated features on the frames.
I wish to alter your code in a way that i can predict the grade of a given video.

i think that a single grade can be treated as two classes of good&bad, or something such as this.

what do you think? can you focus me and aiming me on how to implement it using your code?

thanks a lot in advance!