Progress indicator on vgg.test for lots of data?

Part 1 (2017)
1

Hi everyone,

I’m trying to get the activations for a dense layer of VGG to solve an image similarity task and I need to call vgg.test() on around 2 million images.

It’s running fine but I’m not sure how long it’s going to take - is there a callback on vgg.test or some way to get a progress update?

[ * ] batches, preds = vgg.test(path_pics, batch_size = batch_size*2)
Found 2176945 images belonging to 1 classes.

Thanks!

2

I’m thinking of writing some glue code to move say 100k images at a time from a /todo/ folder into the path_pics folder and then merging the preds together at the end - is there a simple way to do this with keras functions / batch generators or should I just write some python code to move the images?