TensorFlow: How to freeze a model and serve it with a python API

THanks for a very nice article! Using persitent session I ran into out of memory after calling my endpoint just few times, however if I initialize session at every call it works fine, but slower of course. Do you have any tips how to handle memory release with persistent session?

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.