Save trained models

[milesgranger]

Hello,
First thank you for the work done with this library! I'm still quite green with Rust in general, so perhaps I'm missing something obvious, but it would be nice if one could save trained models for later predictions. I didn't see anything in the documentation relating to this.

Thanks again! :)

[shybyte]

I wonder if just adding #[derive(Serialize, Deserialize)] from https://github.com/serde-rs/serde to all relevant structs would be an appropriate implementation of this feature.

[shybyte]

This issue may be a duplicate of #122