How can I save the progress of a randomized grid search?

I am trying to find the best model for a classification task using RandomizedSearchCV to sample a range of hyperparameters. The problem I have is that I am running this on a big server in a closed network that is prone to crashing; so I am unable to run the search for long enough to sample the hyperparameter space I have defined. Ideally I'd like to save each combination of hyperparameters run somehow so when I restart the session it can pick up where it left off - meaning the server crashes don't put me back to square one.

This looks like it might be harder to implement than I thought (feature request here from 2018) and I can't see how to use the suggested CalibratedClassifierCV. Maybe I need to use the verbose option and log it somehow then parse it on restarting? Has anyone ever come up against a similar problem? Thanks in advance.

How many English words
do you know?
Test your English vocabulary size, and measure
how many words do you know
Online Test
Powered by Examplum