Add support for XGBoost UBJSON in FIL

[hcho3]

Closes #6008

[dantegd]

Just tested and this fixes things perfectly for me

[jameslamb]

Great, thanks! I did not try re-running the notebook yet, but I trust that this works.

Won't block the PR over this since this is time-sensitive for the release... but I think we should have some unit tests covering these new if-else branches, to detect such issues earlier in the future. I'd be happy to do that on 24.10 if you'd like, would be a good opportunity for me to learn a little more about FIL.

[jameslamb]

Oh wait.... doesn't the notebook need to be updated?

Here it's saying model_type="xgboost"

cuml/notebooks/forest_inference_demo.ipynb

Line 278 in 50d1a74

" model_type='xgboost'\n",

But I think rapidsai/docker CI is going to still be broken, because it'll pull in xgboost=2.1.* and with that, the appropriate model format is "xgboost_ubjson".

cross-ref: rapidsai/docker#699

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB