AnandInguva commented on code in PR #24529:
URL: https://github.com/apache/beam/pull/24529#discussion_r1040077334
##########
examples/notebooks/beam-ml/run_inference_tensorflow_hub.ipynb:
##########
@@ -0,0 +1,307 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ "source": [
+ "##### Copyright 2022 Google Inc.\n",
+ "\n",
+ "Licensed under the Apache License, Version 2.0 (the \"License\").\n",
+ "<!--\n",
+ " Licensed to the Apache Software Foundation (ASF) under one\n",
+ " or more contributor license agreements. See the NOTICE file\n",
+ " distributed with this work for additional information\n",
+ " regarding copyright ownership. The ASF licenses this file\n",
+ " to you under the Apache License, Version 2.0 (the\n",
+ " \"License\"); you may not use this file except in compliance\n",
+ " with the License. You may obtain a copy of the License at\n",
+ "\n",
+ " http://www.apache.org/licenses/LICENSE-2.0\n";,
+ "\n",
+ " Unless required by applicable law or agreed to in writing,\n",
+ " software distributed under the License is distributed on an\n",
+ " \"AS IS\" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY\n",
+ " KIND, either express or implied. See the License for the\n",
+ " specific language governing permissions and limitations\n",
+ " under the License.\n",
+ "-->\n"
+ ],
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3ac8fc4a-a0ef-47b9-bd80-10801eebe13e"
+ },
+ "source": [
+ "# RunInference with Sentenced T5 Model from TensorFlow Hub\n",
+ "\n",
+ "\n",
+ "In this notebook, we walk through the use of the RunInference
transform with a [sentence encoder built on T5
model](https://tfhub.dev/google/sentence-t5/st5-base/1) and testing it locally
with Interactive Runner.\n"
+ ],
+ "id": "3ac8fc4a-a0ef-47b9-bd80-10801eebe13e"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3402ecc9-28d6-4226-99b1-147a2d23b7a9"
+ },
+ "source": [
+ "## Install and import packages."
+ ],
+ "id": "3402ecc9-28d6-4226-99b1-147a2d23b7a9"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "b22e51ad-cbf2-4dfe-a1c6-3ad17d5a3749",
+ "scrolled": true
+ },
+ "outputs": [],
+ "source": [
+ "!pip install apache_beam[gcp]==2.41.0\n",
+ "!pip install apache-beam[interactive]==2.41.0\n",
Review Comment:
Replace both of apache_beam installs with `pip install
apache_beam[gcp,interactive]`
##########
examples/notebooks/beam-ml/run_inference_tensorflow_hub.ipynb:
##########
@@ -0,0 +1,307 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ "source": [
+ "##### Copyright 2022 Google Inc.\n",
Review Comment:
For example, like this notebook:
https://github.com/apache/beam/blob/3de2f85b3439a7084b60eb600a56bdc4cd36525c/examples/notebooks/beam-ml/run_inference_pytorch.ipynb
##########
examples/notebooks/beam-ml/run_inference_tensorflow_hub.ipynb:
##########
@@ -0,0 +1,307 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ "source": [
+ "##### Copyright 2022 Google Inc.\n",
+ "\n",
+ "Licensed under the Apache License, Version 2.0 (the \"License\").\n",
+ "<!--\n",
+ " Licensed to the Apache Software Foundation (ASF) under one\n",
+ " or more contributor license agreements. See the NOTICE file\n",
+ " distributed with this work for additional information\n",
+ " regarding copyright ownership. The ASF licenses this file\n",
+ " to you under the Apache License, Version 2.0 (the\n",
+ " \"License\"); you may not use this file except in compliance\n",
+ " with the License. You may obtain a copy of the License at\n",
+ "\n",
+ " http://www.apache.org/licenses/LICENSE-2.0\n";,
+ "\n",
+ " Unless required by applicable law or agreed to in writing,\n",
+ " software distributed under the License is distributed on an\n",
+ " \"AS IS\" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY\n",
+ " KIND, either express or implied. See the License for the\n",
+ " specific language governing permissions and limitations\n",
+ " under the License.\n",
+ "-->\n"
+ ],
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3ac8fc4a-a0ef-47b9-bd80-10801eebe13e"
+ },
+ "source": [
+ "# RunInference with Sentenced T5 Model from TensorFlow Hub\n",
+ "\n",
+ "\n",
+ "In this notebook, we walk through the use of the RunInference
transform with a [sentence encoder built on T5
model](https://tfhub.dev/google/sentence-t5/st5-base/1) and testing it locally
with Interactive Runner.\n"
+ ],
+ "id": "3ac8fc4a-a0ef-47b9-bd80-10801eebe13e"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3402ecc9-28d6-4226-99b1-147a2d23b7a9"
+ },
+ "source": [
+ "## Install and import packages."
+ ],
+ "id": "3402ecc9-28d6-4226-99b1-147a2d23b7a9"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "b22e51ad-cbf2-4dfe-a1c6-3ad17d5a3749",
+ "scrolled": true
+ },
+ "outputs": [],
+ "source": [
+ "!pip install apache_beam[gcp]==2.41.0\n",
+ "!pip install apache-beam[interactive]==2.41.0\n",
Review Comment:
Also, do we have a restriction on the Beam version? If not, we can remove
the pinning of version.
##########
examples/notebooks/beam-ml/run_inference_tensorflow_hub.ipynb:
##########
@@ -0,0 +1,307 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ "source": [
+ "##### Copyright 2022 Google Inc.\n",
+ "\n",
+ "Licensed under the Apache License, Version 2.0 (the \"License\").\n",
+ "<!--\n",
+ " Licensed to the Apache Software Foundation (ASF) under one\n",
+ " or more contributor license agreements. See the NOTICE file\n",
+ " distributed with this work for additional information\n",
+ " regarding copyright ownership. The ASF licenses this file\n",
+ " to you under the Apache License, Version 2.0 (the\n",
+ " \"License\"); you may not use this file except in compliance\n",
+ " with the License. You may obtain a copy of the License at\n",
+ "\n",
+ " http://www.apache.org/licenses/LICENSE-2.0\n";,
+ "\n",
+ " Unless required by applicable law or agreed to in writing,\n",
+ " software distributed under the License is distributed on an\n",
+ " \"AS IS\" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY\n",
+ " KIND, either express or implied. See the License for the\n",
+ " specific language governing permissions and limitations\n",
+ " under the License.\n",
+ "-->\n"
+ ],
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3ac8fc4a-a0ef-47b9-bd80-10801eebe13e"
+ },
+ "source": [
+ "# RunInference with Sentenced T5 Model from TensorFlow Hub\n",
+ "\n",
+ "\n",
+ "In this notebook, we walk through the use of the RunInference
transform with a [sentence encoder built on T5
model](https://tfhub.dev/google/sentence-t5/st5-base/1) and testing it locally
with Interactive Runner.\n"
+ ],
+ "id": "3ac8fc4a-a0ef-47b9-bd80-10801eebe13e"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3402ecc9-28d6-4226-99b1-147a2d23b7a9"
+ },
+ "source": [
+ "## Install and import packages."
+ ],
+ "id": "3402ecc9-28d6-4226-99b1-147a2d23b7a9"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "b22e51ad-cbf2-4dfe-a1c6-3ad17d5a3749",
+ "scrolled": true
+ },
+ "outputs": [],
+ "source": [
+ "!pip install apache_beam[gcp]==2.41.0\n",
+ "!pip install apache-beam[interactive]==2.41.0\n",
+ "!pip install tensorflow==2.10.0\n",
+ "!pip install tensorflow_text==2.10.0\n",
+ "!pip install keras==2.10.0\n",
+ "!pip install tfx_bsl==1.10.0\n",
+ "!pip install pillow==8.4.0"
+ ],
+ "id": "b22e51ad-cbf2-4dfe-a1c6-3ad17d5a3749"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "f313a508-59ea-47ed-86eb-c9c8e67785f2",
+ "scrolled": true
+ },
+ "outputs": [],
+ "source": [
+ "import os\n",
+ "import importlib\n",
+ "\n",
+ "import numpy as np\n",
+ "import tensorflow as tf\n",
+ "import tensorflow_hub as hub\n",
+ "import tensorflow_text\n",
+ "\n",
+ "from tensorflow import keras\n",
+ "\n",
+ "from typing import Any\n",
+ "from typing import Dict\n",
+ "from typing import Iterable\n",
+ "from typing import Optional\n",
+ "from typing import Sequence\n",
+ "\n",
+ "import apache_beam as beam\n",
+ "import apache_beam.runners.interactive.interactive_beam as ib\n",
+ "\n",
+ "from apache_beam.ml.inference.base import RunInference\n",
+ "from apache_beam.ml.inference.base import ModelHandler\n",
+ "from apache_beam.ml.inference.base import PredictionResult\n",
+ "from apache_beam.internal import pickler\n",
+ "from apache_beam.runners.runner import PipelineResult\n",
+ "from apache_beam.runners.interactive.interactive_runner import
InteractiveRunner\n",
+ "\n",
+ "from tfx_bsl.public.beam.run_inference import CreateModelHandler\n",
+ "from tfx_bsl.public.proto import model_spec_pb2"
+ ],
+ "id": "f313a508-59ea-47ed-86eb-c9c8e67785f2"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "74db0203-3d26-4bc4-8271-81fad9756297"
+ },
+ "source": [
+ "## Create a Keras Model from TensorFlow Hub Image"
+ ],
+ "id": "74db0203-3d26-4bc4-8271-81fad9756297"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "r1fgn5I_-mzA"
+ },
+ "outputs": [],
+ "source": [
+ "from google.colab import auth\n",
+ "auth.authenticate_user()"
+ ],
+ "id": "r1fgn5I_-mzA"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "2ff8e394-f577-4dea-bef9-a4f4528c1378"
+ },
+ "outputs": [],
+ "source": [
+ "PROJECT_ID = '<Project Id>'\n",
+ "GCS_BUCKET = '<GCS Bucket>'\n",
+ "\n",
+ "MODEL_PATH = f'{GCS_BUCKET}/st5-base/1'"
+ ],
+ "id": "2ff8e394-f577-4dea-bef9-a4f4528c1378"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "ccaede25-1c1a-4ec4-9296-25c9a2ac43d7"
+ },
+ "outputs": [],
+ "source": [
+ "inp = tf.keras.layers.Input(shape=[], dtype=tf.string,
name='input')\n",
+ "hub_url = \"https://tfhub.dev/google/sentence-t5/st5-base/1\"\n";,
+ "imported = hub.KerasLayer(hub_url)\n",
+ "outp = imported(inp)\n",
+ "model = tf.keras.Model(inp, outp)\n",
+ "\n",
+ "# Sentenced-T5 model returns a 768-dimensional vector for an English
text input.\n",
+ "# Note the 'input' that we will pass in as example's feature key
name.\n",
+ "model.summary()"
Review Comment:
Move the model.summary to a different cell for better visibility.
##########
examples/notebooks/beam-ml/run_inference_tensorflow_hub.ipynb:
##########
@@ -0,0 +1,307 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ "source": [
+ "##### Copyright 2022 Google Inc.\n",
+ "\n",
+ "Licensed under the Apache License, Version 2.0 (the \"License\").\n",
+ "<!--\n",
+ " Licensed to the Apache Software Foundation (ASF) under one\n",
+ " or more contributor license agreements. See the NOTICE file\n",
+ " distributed with this work for additional information\n",
+ " regarding copyright ownership. The ASF licenses this file\n",
+ " to you under the Apache License, Version 2.0 (the\n",
+ " \"License\"); you may not use this file except in compliance\n",
+ " with the License. You may obtain a copy of the License at\n",
+ "\n",
+ " http://www.apache.org/licenses/LICENSE-2.0\n";,
+ "\n",
+ " Unless required by applicable law or agreed to in writing,\n",
+ " software distributed under the License is distributed on an\n",
+ " \"AS IS\" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY\n",
+ " KIND, either express or implied. See the License for the\n",
+ " specific language governing permissions and limitations\n",
+ " under the License.\n",
+ "-->\n"
+ ],
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3ac8fc4a-a0ef-47b9-bd80-10801eebe13e"
+ },
+ "source": [
+ "# RunInference with Sentenced T5 Model from TensorFlow Hub\n",
+ "\n",
+ "\n",
+ "In this notebook, we walk through the use of the RunInference
transform with a [sentence encoder built on T5
model](https://tfhub.dev/google/sentence-t5/st5-base/1) and testing it locally
with Interactive Runner.\n"
Review Comment:
```suggestion
"In this notebook, we walk through the use of the RunInference
transform with a [sentence encoder built on T5
model](https://tfhub.dev/google/sentence-t5/st5-base/1) and run the notebook
locally with Interactive Runner.\n"
```
Can we also add a link to Beam's Interactive runner?
https://beam.apache.org/releases/pydoc/2.11.0/apache_beam.runners.interactive.interactive_runner.html
##########
examples/notebooks/beam-ml/run_inference_tensorflow_hub.ipynb:
##########
@@ -0,0 +1,307 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ "source": [
+ "##### Copyright 2022 Google Inc.\n",
+ "\n",
+ "Licensed under the Apache License, Version 2.0 (the \"License\").\n",
+ "<!--\n",
+ " Licensed to the Apache Software Foundation (ASF) under one\n",
+ " or more contributor license agreements. See the NOTICE file\n",
+ " distributed with this work for additional information\n",
+ " regarding copyright ownership. The ASF licenses this file\n",
+ " to you under the Apache License, Version 2.0 (the\n",
+ " \"License\"); you may not use this file except in compliance\n",
+ " with the License. You may obtain a copy of the License at\n",
+ "\n",
+ " http://www.apache.org/licenses/LICENSE-2.0\n";,
+ "\n",
+ " Unless required by applicable law or agreed to in writing,\n",
+ " software distributed under the License is distributed on an\n",
+ " \"AS IS\" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY\n",
+ " KIND, either express or implied. See the License for the\n",
+ " specific language governing permissions and limitations\n",
+ " under the License.\n",
+ "-->\n"
+ ],
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3ac8fc4a-a0ef-47b9-bd80-10801eebe13e"
+ },
+ "source": [
+ "# RunInference with Sentenced T5 Model from TensorFlow Hub\n",
+ "\n",
+ "\n",
+ "In this notebook, we walk through the use of the RunInference
transform with a [sentence encoder built on T5
model](https://tfhub.dev/google/sentence-t5/st5-base/1) and testing it locally
with Interactive Runner.\n"
+ ],
+ "id": "3ac8fc4a-a0ef-47b9-bd80-10801eebe13e"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3402ecc9-28d6-4226-99b1-147a2d23b7a9"
+ },
+ "source": [
+ "## Install and import packages."
+ ],
+ "id": "3402ecc9-28d6-4226-99b1-147a2d23b7a9"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "b22e51ad-cbf2-4dfe-a1c6-3ad17d5a3749",
+ "scrolled": true
+ },
+ "outputs": [],
+ "source": [
+ "!pip install apache_beam[gcp]==2.41.0\n",
+ "!pip install apache-beam[interactive]==2.41.0\n",
+ "!pip install tensorflow==2.10.0\n",
+ "!pip install tensorflow_text==2.10.0\n",
+ "!pip install keras==2.10.0\n",
+ "!pip install tfx_bsl==1.10.0\n",
+ "!pip install pillow==8.4.0"
+ ],
+ "id": "b22e51ad-cbf2-4dfe-a1c6-3ad17d5a3749"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "f313a508-59ea-47ed-86eb-c9c8e67785f2",
+ "scrolled": true
+ },
+ "outputs": [],
+ "source": [
+ "import os\n",
+ "import importlib\n",
+ "\n",
+ "import numpy as np\n",
+ "import tensorflow as tf\n",
+ "import tensorflow_hub as hub\n",
+ "import tensorflow_text\n",
+ "\n",
+ "from tensorflow import keras\n",
+ "\n",
+ "from typing import Any\n",
+ "from typing import Dict\n",
+ "from typing import Iterable\n",
+ "from typing import Optional\n",
+ "from typing import Sequence\n",
+ "\n",
+ "import apache_beam as beam\n",
+ "import apache_beam.runners.interactive.interactive_beam as ib\n",
+ "\n",
+ "from apache_beam.ml.inference.base import RunInference\n",
+ "from apache_beam.ml.inference.base import ModelHandler\n",
+ "from apache_beam.ml.inference.base import PredictionResult\n",
+ "from apache_beam.internal import pickler\n",
+ "from apache_beam.runners.runner import PipelineResult\n",
+ "from apache_beam.runners.interactive.interactive_runner import
InteractiveRunner\n",
+ "\n",
+ "from tfx_bsl.public.beam.run_inference import CreateModelHandler\n",
+ "from tfx_bsl.public.proto import model_spec_pb2"
+ ],
+ "id": "f313a508-59ea-47ed-86eb-c9c8e67785f2"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "74db0203-3d26-4bc4-8271-81fad9756297"
+ },
+ "source": [
+ "## Create a Keras Model from TensorFlow Hub Image"
+ ],
+ "id": "74db0203-3d26-4bc4-8271-81fad9756297"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "r1fgn5I_-mzA"
+ },
+ "outputs": [],
+ "source": [
+ "from google.colab import auth\n",
+ "auth.authenticate_user()"
+ ],
+ "id": "r1fgn5I_-mzA"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "2ff8e394-f577-4dea-bef9-a4f4528c1378"
+ },
+ "outputs": [],
+ "source": [
+ "PROJECT_ID = '<Project Id>'\n",
Review Comment:
We are not using this variable in the notebook. Can we remove it?
##########
examples/notebooks/beam-ml/run_inference_tensorflow_hub.ipynb:
##########
@@ -0,0 +1,307 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ "source": [
+ "##### Copyright 2022 Google Inc.\n",
+ "\n",
+ "Licensed under the Apache License, Version 2.0 (the \"License\").\n",
+ "<!--\n",
+ " Licensed to the Apache Software Foundation (ASF) under one\n",
+ " or more contributor license agreements. See the NOTICE file\n",
+ " distributed with this work for additional information\n",
+ " regarding copyright ownership. The ASF licenses this file\n",
+ " to you under the Apache License, Version 2.0 (the\n",
+ " \"License\"); you may not use this file except in compliance\n",
+ " with the License. You may obtain a copy of the License at\n",
+ "\n",
+ " http://www.apache.org/licenses/LICENSE-2.0\n";,
+ "\n",
+ " Unless required by applicable law or agreed to in writing,\n",
+ " software distributed under the License is distributed on an\n",
+ " \"AS IS\" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY\n",
+ " KIND, either express or implied. See the License for the\n",
+ " specific language governing permissions and limitations\n",
+ " under the License.\n",
+ "-->\n"
+ ],
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3ac8fc4a-a0ef-47b9-bd80-10801eebe13e"
+ },
+ "source": [
+ "# RunInference with Sentenced T5 Model from TensorFlow Hub\n",
+ "\n",
+ "\n",
+ "In this notebook, we walk through the use of the RunInference
transform with a [sentence encoder built on T5
model](https://tfhub.dev/google/sentence-t5/st5-base/1) and testing it locally
with Interactive Runner.\n"
+ ],
+ "id": "3ac8fc4a-a0ef-47b9-bd80-10801eebe13e"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3402ecc9-28d6-4226-99b1-147a2d23b7a9"
+ },
+ "source": [
+ "## Install and import packages."
+ ],
+ "id": "3402ecc9-28d6-4226-99b1-147a2d23b7a9"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "b22e51ad-cbf2-4dfe-a1c6-3ad17d5a3749",
+ "scrolled": true
+ },
+ "outputs": [],
+ "source": [
+ "!pip install apache_beam[gcp]==2.41.0\n",
+ "!pip install apache-beam[interactive]==2.41.0\n",
+ "!pip install tensorflow==2.10.0\n",
+ "!pip install tensorflow_text==2.10.0\n",
+ "!pip install keras==2.10.0\n",
+ "!pip install tfx_bsl==1.10.0\n",
+ "!pip install pillow==8.4.0"
+ ],
+ "id": "b22e51ad-cbf2-4dfe-a1c6-3ad17d5a3749"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "f313a508-59ea-47ed-86eb-c9c8e67785f2",
+ "scrolled": true
+ },
+ "outputs": [],
+ "source": [
+ "import os\n",
+ "import importlib\n",
+ "\n",
+ "import numpy as np\n",
+ "import tensorflow as tf\n",
+ "import tensorflow_hub as hub\n",
+ "import tensorflow_text\n",
+ "\n",
+ "from tensorflow import keras\n",
+ "\n",
+ "from typing import Any\n",
+ "from typing import Dict\n",
+ "from typing import Iterable\n",
+ "from typing import Optional\n",
+ "from typing import Sequence\n",
+ "\n",
+ "import apache_beam as beam\n",
+ "import apache_beam.runners.interactive.interactive_beam as ib\n",
+ "\n",
+ "from apache_beam.ml.inference.base import RunInference\n",
+ "from apache_beam.ml.inference.base import ModelHandler\n",
+ "from apache_beam.ml.inference.base import PredictionResult\n",
+ "from apache_beam.internal import pickler\n",
+ "from apache_beam.runners.runner import PipelineResult\n",
+ "from apache_beam.runners.interactive.interactive_runner import
InteractiveRunner\n",
+ "\n",
+ "from tfx_bsl.public.beam.run_inference import CreateModelHandler\n",
+ "from tfx_bsl.public.proto import model_spec_pb2"
+ ],
+ "id": "f313a508-59ea-47ed-86eb-c9c8e67785f2"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "74db0203-3d26-4bc4-8271-81fad9756297"
+ },
+ "source": [
+ "## Create a Keras Model from TensorFlow Hub Image"
+ ],
+ "id": "74db0203-3d26-4bc4-8271-81fad9756297"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "r1fgn5I_-mzA"
+ },
+ "outputs": [],
+ "source": [
+ "from google.colab import auth\n",
+ "auth.authenticate_user()"
+ ],
+ "id": "r1fgn5I_-mzA"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "2ff8e394-f577-4dea-bef9-a4f4528c1378"
+ },
+ "outputs": [],
+ "source": [
+ "PROJECT_ID = '<Project Id>'\n",
+ "GCS_BUCKET = '<GCS Bucket>'\n",
Review Comment:
Also, can you provide instructions for the users on how they need to use the
`GCS_BUCKET`?
for example: Download the model from `some_link` and place it in the
`GCS_BUCKET`
##########
examples/notebooks/beam-ml/run_inference_tensorflow_hub.ipynb:
##########
@@ -0,0 +1,307 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ "source": [
+ "##### Copyright 2022 Google Inc.\n",
+ "\n",
+ "Licensed under the Apache License, Version 2.0 (the \"License\").\n",
+ "<!--\n",
+ " Licensed to the Apache Software Foundation (ASF) under one\n",
+ " or more contributor license agreements. See the NOTICE file\n",
+ " distributed with this work for additional information\n",
+ " regarding copyright ownership. The ASF licenses this file\n",
+ " to you under the Apache License, Version 2.0 (the\n",
+ " \"License\"); you may not use this file except in compliance\n",
+ " with the License. You may obtain a copy of the License at\n",
+ "\n",
+ " http://www.apache.org/licenses/LICENSE-2.0\n";,
+ "\n",
+ " Unless required by applicable law or agreed to in writing,\n",
+ " software distributed under the License is distributed on an\n",
+ " \"AS IS\" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY\n",
+ " KIND, either express or implied. See the License for the\n",
+ " specific language governing permissions and limitations\n",
+ " under the License.\n",
+ "-->\n"
+ ],
+ "id": "634de4ce-07f7-4003-8ff6-d3f58b41a79d"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3ac8fc4a-a0ef-47b9-bd80-10801eebe13e"
+ },
+ "source": [
+ "# RunInference with Sentenced T5 Model from TensorFlow Hub\n",
+ "\n",
+ "\n",
+ "In this notebook, we walk through the use of the RunInference
transform with a [sentence encoder built on T5
model](https://tfhub.dev/google/sentence-t5/st5-base/1) and testing it locally
with Interactive Runner.\n"
+ ],
+ "id": "3ac8fc4a-a0ef-47b9-bd80-10801eebe13e"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3402ecc9-28d6-4226-99b1-147a2d23b7a9"
+ },
+ "source": [
+ "## Install and import packages."
+ ],
+ "id": "3402ecc9-28d6-4226-99b1-147a2d23b7a9"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "b22e51ad-cbf2-4dfe-a1c6-3ad17d5a3749",
+ "scrolled": true
+ },
+ "outputs": [],
+ "source": [
+ "!pip install apache_beam[gcp]==2.41.0\n",
+ "!pip install apache-beam[interactive]==2.41.0\n",
+ "!pip install tensorflow==2.10.0\n",
+ "!pip install tensorflow_text==2.10.0\n",
+ "!pip install keras==2.10.0\n",
+ "!pip install tfx_bsl==1.10.0\n",
+ "!pip install pillow==8.4.0"
+ ],
+ "id": "b22e51ad-cbf2-4dfe-a1c6-3ad17d5a3749"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "f313a508-59ea-47ed-86eb-c9c8e67785f2",
+ "scrolled": true
+ },
+ "outputs": [],
+ "source": [
+ "import os\n",
+ "import importlib\n",
+ "\n",
+ "import numpy as np\n",
+ "import tensorflow as tf\n",
+ "import tensorflow_hub as hub\n",
+ "import tensorflow_text\n",
+ "\n",
+ "from tensorflow import keras\n",
+ "\n",
+ "from typing import Any\n",
+ "from typing import Dict\n",
+ "from typing import Iterable\n",
+ "from typing import Optional\n",
+ "from typing import Sequence\n",
+ "\n",
+ "import apache_beam as beam\n",
+ "import apache_beam.runners.interactive.interactive_beam as ib\n",
+ "\n",
+ "from apache_beam.ml.inference.base import RunInference\n",
+ "from apache_beam.ml.inference.base import ModelHandler\n",
+ "from apache_beam.ml.inference.base import PredictionResult\n",
+ "from apache_beam.internal import pickler\n",
+ "from apache_beam.runners.runner import PipelineResult\n",
+ "from apache_beam.runners.interactive.interactive_runner import
InteractiveRunner\n",
+ "\n",
+ "from tfx_bsl.public.beam.run_inference import CreateModelHandler\n",
+ "from tfx_bsl.public.proto import model_spec_pb2"
+ ],
+ "id": "f313a508-59ea-47ed-86eb-c9c8e67785f2"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "74db0203-3d26-4bc4-8271-81fad9756297"
+ },
+ "source": [
+ "## Create a Keras Model from TensorFlow Hub Image"
+ ],
+ "id": "74db0203-3d26-4bc4-8271-81fad9756297"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "r1fgn5I_-mzA"
+ },
+ "outputs": [],
+ "source": [
+ "from google.colab import auth\n",
+ "auth.authenticate_user()"
+ ],
+ "id": "r1fgn5I_-mzA"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "2ff8e394-f577-4dea-bef9-a4f4528c1378"
+ },
+ "outputs": [],
+ "source": [
+ "PROJECT_ID = '<Project Id>'\n",
+ "GCS_BUCKET = '<GCS Bucket>'\n",
+ "\n",
+ "MODEL_PATH = f'{GCS_BUCKET}/st5-base/1'"
+ ],
+ "id": "2ff8e394-f577-4dea-bef9-a4f4528c1378"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "ccaede25-1c1a-4ec4-9296-25c9a2ac43d7"
+ },
+ "outputs": [],
+ "source": [
+ "inp = tf.keras.layers.Input(shape=[], dtype=tf.string,
name='input')\n",
+ "hub_url = \"https://tfhub.dev/google/sentence-t5/st5-base/1\"\n";,
+ "imported = hub.KerasLayer(hub_url)\n",
+ "outp = imported(inp)\n",
+ "model = tf.keras.Model(inp, outp)\n",
+ "\n",
+ "# Sentenced-T5 model returns a 768-dimensional vector for an English
text input.\n",
+ "# Note the 'input' that we will pass in as example's feature key
name.\n",
+ "model.summary()"
+ ],
+ "id": "ccaede25-1c1a-4ec4-9296-25c9a2ac43d7"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "29803d5b-93b9-41fc-b414-f7c737c5d7bc"
+ },
+ "source": [
+ "## Save the model with a TF function definition for RunInference()"
+ ],
+ "id": "29803d5b-93b9-41fc-b414-f7c737c5d7bc"
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "id": "704abfca-5956-4fc1-9717-4c6d5bf2db8e"
+ },
+ "outputs": [],
+ "source": [
+ "RAW_DATA_PREDICT_SPEC = {\n",
+ " 'input': tf.io.FixedLenFeature([], tf.string),\n",
+ "}\n",
+ "\n",
+ "@tf.function(input_signature=[tf.TensorSpec(shape=[None],
dtype=tf.string)])\n",
+ "def call(serialized_examples):\n",
+ " features = tf.io.parse_example(serialized_examples,
RAW_DATA_PREDICT_SPEC)\n",
+ " return model(features)\n",
+ "\n",
+ "tf.saved_model.save(model, MODEL_PATH, signatures={'serving_default':
call})"
+ ],
+ "id": "704abfca-5956-4fc1-9717-4c6d5bf2db8e"
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "7b56569d-e540-44ed-a46a-9cec886522f6"
+ },
+ "source": [
+ "## Create and test a RunInference pipeline locally"
Review Comment:
Where are we testing the output? the heading looks misleading if we are not
actually testing the output
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
