I am working on a ranking model and am pretty new to the world of tensor flow. As such I have been trying to adapt some Tensorflow ranking examples to work with my model which is trying to rank runners in a race. However I am having some issues getting my model to build. The model I am following uses the the tfrecords example list with context file format and assume my current issue is how I am generating that dataset. The error I get is
ValueError: Feature finish_position is not in features dictionary. However, I am not sure how to inspect what the is in the features dictionary and what I am doing wrong to not have it in there. Can anyone assist me in debugging? At this point I am not really concerned about an optimal model (I am hardly including any features), rather just looking for it to build and iterate on.
Here is my python script
My data files are listed in there but pulled in via the Keras get file util.
I also have these 2 other files.
- https://stranlabs.s3.us-west-1.amazonaws.com/tfrecords/tfexport.py This is the file I am using to pull files from my database
- https://stranlabs.s3.us-west-1.amazonaws.com/tfrecords/tfview.py This is the file I am trying to use to inspect the contents of my tfrecords files.
Can anyone help get me on the right path of getting my model to compile?