-
UMAP uses a sparse spectral clustering function for one of its initialization strategies. UMAP creates a sparse adjacency matrix from the weighted nearest neighbors, but a more general-use implementat…
-
For accuracy tests, it would be helpful to provide an easy way to load more realistic datasets. These should include (like the Python #1791 issue):
- [ ] sklearn datasets
- [ ] 1-2 interesting exa…
-
DBSCAN is already doing this, as it uses the eps value to determine the the edges of the adjacency graph. It should not be too hard to perform the same batching that DBSCAN is performing, whilst thres…
-
It would be nice to update the thresholds used in RF tests (both single GPU and MNMG) to a more realistic value. It seems like the RF accuracy has improved from what it was before and the old threshol…
-
**Is your feature request related to a problem? Please describe.**
Probability calibration is needed for SVC to enable prediction of class probabilities (predict_proba). Currently [CalibratedClassifi…
-
Dask uses Python's `logging` module for logging. In particular, [all logs pushed to `logging` on the workers' side are forwarded to a central location.](https://docs.dask.org/en/latest/debugging.html#…
hcho3 updated
3 years ago
-
Hi!
`cudf.read_json` supports passing multiple files to it, which is much more performant than reading json files individually, and then merging them.
It would be very useful for certain workloa…
-
- [x] Add SHAP package to CI and pass/fix any issues
- [x] Additional real world dataset tests and stress tests
- [ ] Finish LassoLarsIC
- [x] Review and benchmark tag usage
- [x] Better handling …
-
It would be great to have a `sample_weight` option in the `cuml.LogisticRegression` `fit()` method like in `sklearn`. That way one could easily implement custom penalty methods.
-
Linking issue from RAFT: https://github.com/rapidsai/raft/issues/97
Other things to consider:
1. RMM also has adapters for thrust device vectors which is especially useful when trying to use the t…