BentoML

Hi <@U05KX7Q663Y> - you will need to specify the `methods` when saving the model

If you use `bentoml.xgboost.save_model` , then it will do it for you automatically

<@UKB4CLKP1>
Thank you, I have solved the problem. I'm still using the picklable_model, but I had to register a method called `predict` and then infer using `runner.predict.run`. If I only use `runner.run`, it doesn't work.
If following the documentation <https://docs.bentoml.com/en/latest/frameworks/picklable.html|here>, then it seems the documentation might be incorrect. Could you please check this documentation and let me know if it is correct or not?

<@U05KX7Q663Y> you should use `bentoml.xgboost` module which will set a few additional config for you for best performance

Picklable module is generally not recommended, unless you have a custom python implementation of a model

`runnable#__call__` will be mapped to `runner.run()`, whereas `runnable#preddict` will map to `runner.predict.run` - this is as expected