Trained model weights

Hi, I’ve had a look at the basic_submission notebook and I’m not understanding one thing: if I train my model locally, and then I want to submit it without re-training it, is there a way to do it? How can I upload my model weights?

1 Like

Team said in Discord that you have to save your model in resources folder before submitting using the cli. However, I haven’t had a chance to try it yet because I got no CPU cloud quota until next Wednesday

Following up on this discussion, I’ve tested again this morning and now I actually have problems with the cli and crunch test --no-force-first-train trying to use my joblib model in resources that throws me IndexError: pop from empty list @enzo

Yes, but currently you can only sent the content of the resources directory by using the CLI.

I am not sure where this is coming from?
Could you maybe share the trace?

Here is the following:

@worldwide-jeremy’s problem was related to a library version difference between the local and cloud environment.

1 Like