Benjamin Bossan

> What about providing a [checksum](https://calmcode.io/scikit-save/checksum.html) on a pickle file? Still won't solve many of the issues. But you could at least implement a loader that forces the user to...

I'll take a look. It is probably impossible to give any safety guarantees, so I assume we would just rely on fickling (btw. what a bad library name for a...

I tried it out just to see some results. It didn't work on pytorch or skoch code, probably not too surprising. I also tested it on a simple sklearn pipeline...

Not sure if it fits here, but I was wondering about predicting labels instead of ints for classification. AFAIK, you can already fit a classifier with labels (e.g. `y=["setosa", "versicolor",...

> I think I rather show the model output there. If the user wants strings as outputs, they can train it themselves. The default could stay the same but if...

> they could add that to their pipeline as a last step kinda thing If that's easily possible, we should show how. I would guess that showing the label instead...

Despite merging #130, let's keep this open for now because the inference repos can't be set to private for now.

In general, I'm in favor of this, because it makes it easy to replicate everything. I would not go as far as to force this but it should be encouraged....

> but I'm not sure of the API to let this happen. You mean an API from the skops side to provide this functionality? I don't think it's really possible....

Hmm, wouldn't that result in users potentially uploading files that they may not have intended to be uploaded?