Incorrect number of (training, validation) images when using a split of 0.5 and 12

Course Q&A TensorFlow Developer Professional Convolutional Neural Networks in TensorFlow
5

Removing objects while iterating through the same collection is a bad idea.
Please use a separate list for those images whose size is not zero.

The grader feedback means the following:

  1. Create a source directory with 12 images where 6 images are empty and invoke the split_data function with SPLIT_SIZE = .5.
  2. The expected outcome is that each training and validation directories will contain 3 images.

See this as well.