Sahil Wadhwa
Sahil Wadhwa
@kvignesh1420 In the code we used 50 epochs for the spatial stream, but I am pretty sure, we got the results before that. @stillbreeze Do you know the specs for...
Post some code so that we can help.
But still @kvignesh1420 , how did u solve it? It would be better for other people if you could share your views or code.
Basically we stack consecutive 10 optical flow images and form a single 10*2 input (x,y) . If a video has less than 10 frames then we discard that video IMO....
@sudonto yes you can train them jointly but due to lack of resources we had to train them separately.
yes, training and testing after some(predefined) epochs.
Yes this is the test data been prepared. It is just the number of videos processed/prepared for testing AFAIK.
preds = model.predict(X_test) does the predictions of test videos.
@pribadihcr is it resolved now?
There are same. Batch/chunk is nothing but a subset of your main dataset. TF/Theano all work in batches. SGD and parameters updates happen in batches. Batch training is much faster...