standard visualization for Flickr30k about caption-guided-saliency HOT 12 CLOSED

hi
5GBs? Why two? The code doesn't use multi GPU setup in any way.

Yes, it might be so. I don't remember exactly the requirements but try to decrease the batch size. Also, note that there is

I'm quite busy right now, sorry. Maybe in a few weeks. Btw it should be quite straightforward to figure out how to visualize it using msr-vtt visualization.

from caption-guided-saliency.

I'm so sorry to bother you.When I was training MSR-VTT, the code was working.I've decreased batch_size,it does not work.

from caption-guided-saliency.

No worries. I would expect that at least twice more memory is needed for Flickr30k model (simply because of explicitly unrolled LSTM). Try to decrease batch size and/or LSTM hidden state size (in config).

from caption-guided-saliency.

You are a genius.thank you very much for your help.

from caption-guided-saliency.

I've implemented standard visualization for Flickr30k based on your code, but I do not know if it's completely correct, can you check it out?
visualization.zip

from caption-guided-saliency.

@xingtengfei Sorry, the archive seems to be broken, I cannot open it. Could you check?

from caption-guided-saliency.

I created a new file visua_flickr30k

from caption-guided-saliency.

@xingtengfei Did you forget to attach it? :)

from caption-guided-saliency.

I am stupid. haha,I forgot to pull requests

from caption-guided-saliency.

my code is ok?

from caption-guided-saliency.

Sorry, I didn't have time to look into it. I'll do it tomorrow.

from caption-guided-saliency.

I looked through the code. It seems you've tried to keep the code for Flickr30k to be as close as possible to the one we published for MSR-VTT. Thanks for that but here's what we need to take into account:

First of all notice the discrepancy in frame preprocessing in training and visualization (current code):

1. Training: the shorter side is scaled to 256px then 224px center crop is taken (it is done to mimic previously published papers)
2. Visualization: frames are scaled to 400x300 (frame extraction) then scaled again to 299x299 (without preserving aspect ratio). The reason for this - to get 'saliency map' for the whole frame not only for the central part.
   For visualization purposes, we extract features again from all frames in specified input video. Thus, resulting dimensionality is Tx8x8x2048 (training dimensions are 26x1x2048, 8x8 feature maps are mean pooled as in InceptionV3 layer). In your case, you don't extract features again so the image should be properly cropped/scaled.

For Flickr30k training is performed using 8x8x2048 feature maps for every frame directly by unrolling them into 64x2048 'time sequence' as it is described in the paper.

Look into this gist based on your version of my code (see notes about 'normalization' in L170-180):
https://gist.github.com/ramanishka/ccf59b400d99e6aac452f50952525e2f

I'm closing this issue.

from caption-guided-saliency.