How to use a trained tensorflow model?

Question

Python

Hello, I'm trying to parse with tensorflow. I found this repository on github for creating/training a retrieval-based model using UDC.
It seems like I set it up, started the training, which worked for about 12 hours, but I can’t figure out how to implement the model further in the chat as such, Google didn’t help.
After training the model in the ./runs folder, there are files with the name model.ckpt-1505.meta, as I understand it, these are training checkpoints / some training graphs, but I have no idea how to use them further.
Please explain what's what.

Answer 1

There is also a udc_predict.py file with an example of using the trained model for predictions