Ask you a question

[szySunzhouyu]

Recently read your thesis, there are problems in the process of running the code, ask you.
The following problem occurs when split data set in Image-Text-Embedding:
Index exceeds matrix dimensions.

Error in prepare_imdb (line 24)
Images = fullfile(imdb.images.data(train(1:end))) ;

Going beyond the border, I don’t know what’s wrong.

[layumi]

@szySunzhouyu
Sorry for the late response.
I am preparing the supplementary material for CVPR2019.
I will reply you soon at this weekend.

I really appreciate your understanding.
Thank you.

[szySunzhouyu]

It's ok. Looking forward to your reply

[layumi]

Hi @szySunzhouyu
Could you upload a screenshot of the error? Which dataset do you use?

[szySunzhouyu]

@layumi

The dataset I am using is Flickr30k

[layumi]

1. Could you print train = find(imdb.images.set == 1) ?
   Is it an array? What is the latest number in this array?

2. imdb.images.data should be a cell array with filename. How much cell in your imdb.images.data ?

[szySunzhouyu]

Train = find(imdb.images.set == 1) The latest number is 31783.
Imdb.images.data has 31,783 cells.
Sorry, I found the cause of this problem. This is caused by the deletion of a ‘/’.

Now I have a new problem. I am bothering you.

[layumi]

vl_argparse is a function in matconvnet
Please compile and install matconvnet.

I have included my Matconvnet in this repo, so you do not need to download it again.You just need to uncomment and modify some lines in gpu_compile.m and run it in Matlab. Try it~ (The code does not support cudnn 6.0. You may just turn off the Enablecudnn or try cudnn5.1)

If you fail in compilation, you may refer to http://www.vlfeat.org/matconvnet/install/

[szySunzhouyu]

OK,thanks