Hi,
I want to extract GoogleNet features. So, I followed the steps as per mentioned in readme.md file and specified 50 as my batch size in prototxt file and number of data mini-batches as 1000. This command I fired just to test how it takes all the images.
I am using CUB dataset which has 17882 images. So I thought extract_features command will stop much before 50000 count, however, it reached 50000 count and actually mentioned that "Extracted features of 50000 query images for feature blob pool5/7x7_s1".
How come it reached 50000? How extract_features functionality works?