You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Answers #3232.
Explicitly specify `batch_size` as parameter to MNMG KNN models in order to make it visible in the documentation.
Authors:
- viclafargue <viclafargue@nvidia.com>
- Corey J. Nolet <cjnolet@gmail.com>
Approvers:
- Corey J. Nolet
- John Zedlewski
URL: #3246
As noted in https://stackoverflow.com/questions/64785654/why-is-cuml-predict-method-for-knearestneighbors-taking-so-long-with-dask-cudf , the
batch_size
parameter is important for throughput. This is currently an undocumented parameter, as theKNeighborsRegressorMG
class accepts it but the daskKNeighborsRegressor
accepts it only through**kwargs
.We should document all relevant parameters for MNMG KNN.
The text was updated successfully, but these errors were encountered: