[GitHub] [beam] ryanthompson591 commented on a diff in pull request #17368: [BEAM-13983] Sklearn Loader for RunInference

[GitBox <gi...@apache.org>]

ryanthompson591 commented on code in PR #17368:
URL: https://github.com/apache/beam/pull/17368#discussion_r851328207


##########
sdks/python/apache_beam/ml/inference/sklearn_loader.py:
##########
@@ -0,0 +1,73 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements.  See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You under the Apache License, Version 2.0
+# (the "License"); you may not use this file except in compliance with
+# the License.  You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+import abc
+import enum
+import pickle
+import sys
+from dataclasses import dataclass
+from typing import Any
+from typing import Iterable
+from typing import List
+
+import joblib
+import numpy
+
+import apache_beam.ml.inference.api as api
+import apache_beam.ml.inference.base as base
+import sklearn_loader
+from apache_beam.io.filesystems import FileSystems
+
+
+class SerializationType(enum.Enum):
+  PICKLE = 1
+  JOBLIB = 2
+
+
+class SKLearnInferenceRunner(base.InferenceRunner):
+  def run_inference(self, batch: List[numpy.array],
+                    model: Any) -> Iterable[numpy.array]:
+    # vectorize data for better performance
+    vectorized_batch = numpy.stack(batch, axis=0)
+    predictions = model.predict(vectorized_batch)
+    return [api.PredictionResult(x, y) for x, y in zip(batch, predictions)]
+
+  def get_num_bytes(self, batch: List[numpy.array]) -> int:
+    """Returns the number of bytes of data for a batch."""
+    return sum(sys.getsizeof(element) for element in batch)
+
+
+class SKLearnModelLoader(base.ModelLoader):
+  def __init__(
+      self,
+      serialization: SerializationType = SerializationType.PICKLE,
+      model_uri: str = ''):
+    self._serialization = serialization
+    self._model_uri = model_uri

Review Comment:
   I think it's fine to use different parameters. In my view, path communicates more a directory, and URI a single file.
   
   TFX-BSL is using saved_model_spec (which is a proto) that contains model_path. As far as I can tell, TF saves models to a path rather than a URI or file.
   https://cloud.google.com/blog/topics/developers-practitioners/using-tfx-inference-dataflow-large-scale-ml-inference-patterns
   
   What does pytorch do? Is it a single file or a path with a bunch of data?



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: github-unsubscribe@beam.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org