transform the image/mask pairs into tfrecord

Question

With respect to the data format given by Carvana Image Masking Challenge, how can we transform it to the tfrecord format that can be feed into the Deeplab V3 model, which can support VOC and Cityscape, which usually have mask annotation files in XML/text format.

Answer 1

I'm not familiar with the Carvana dataset but maybe this link will help you create a TFRecords file.