Github user thvasilo commented on a diff in the pull request:
https://github.com/apache/flink/pull/871#discussion_r34139499
--- Diff:
flink-staging/flink-ml/src/main/scala/org/apache/flink/ml/recommendation/ALS.scala
---
@@ -25,11 +25,15 @@ import org.apache.flink.api.scala._
import org.apache.flink.api.common.operators.Order
import org.apache.flink.core.memory.{DataOutputView, DataInputView}
import org.apache.flink.ml.common._
-import org.apache.flink.ml.pipeline.{FitOperation,
PredictDataSetOperation, Predictor}
+import org.apache.flink.ml.evaluation.RegressionScores
+import org.apache.flink.ml.math.{DenseVector, BLAS}
+import org.apache.flink.ml.pipeline._
import org.apache.flink.types.Value
import org.apache.flink.util.Collector
-import org.apache.flink.api.common.functions.{Partitioner =>
FlinkPartitioner, GroupReduceFunction, CoGroupFunction}
+import org.apache.flink.api.common.functions.{Partitioner =>
FlinkPartitioner,
+ GroupReduceFunction, CoGroupFunction}
+// TODO: Use only one BLAS interface
--- End diff --
I'm not sure if this belongs to this PR. We can get the BLAS operations
through ml.math.BLAS, like I did in the predict and evaluate operations, but
the lapack ops still need to be done through netlib. Should I change all the
BLAS operations to use ml.math.BLAS in this PR?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
