Only training limited images out of a dataset

I have a dataset of 500 images - I realise this is still quite a small dataset. However, the main issue I need to overcome is whenever I train this dataset it is only ever training 8 images at a time. See attached screenshot. This is my own dataset but the images are all the same dimension, crop and colour profile.

Am I missing something obvious I haven't executed or a default I need to change in the utils.py?

many thanks

Start with the dataset-loading and training configuration in utils.py, then reproduce the report using the 500-image dataset described in the issue. Check why training processes only 8 images at a time and document the relevant setting or observed limitation; confirm the result with the training output and screenshot behavior.

Written by the indexing model from the issue text.