Closed Michael0x2a closed 7 years ago
I have manual saving in the character n-gram logistic regression model, although I'd like to make a minor tweak so it can save with arbitrary filenames in arbitrary paths (probably have to pass those in separately?).
We can also have automatic saving every X epochs while training, but I don't think that's all that useful unless training is so long that we're afraid things will break before training completes.
Implement code to save and load trained model. Probably dependent on #3
Tasks 2 and 3 can probably be combined, since they both use tensorflow?