Hello, I have this dataset contains the labels of english letters and its points, I want to train yolov3 with that dataset so I need to use anchor images, but before that I need to convert this file into original images. I found this script to convert but it didn’t help me so much. Do anyone know any other way to train yolov3 with that dataset? Or help me to convert this file into original images.
There seem to be a few scripts which are loading the data and visualizing a few examples using
E.g. have a look at this script, which loads the data using
After loading, you can directly index the
.iloc and reshape the numpy array to
[28, 28] in order to get the appropriate shaped images.
If you are planning on creating a custom
Dataset, you could load the data in the
__init__ method and get each sample in