Add Inference Python Benchmarks Dataflow workflow (apache#28943)

damondouglas · Oct 16, 2023 · 214f99a · 214f99a
1 parent 09cc368
commit 214f99a
Show file tree

Hide file tree

Showing 6 changed files with 316 additions and 0 deletions.
diff --git a/.github/workflows/beam_Inference_Python_Benchmarks_Dataflow.yml b/.github/workflows/beam_Inference_Python_Benchmarks_Dataflow.yml
@@ -0,0 +1,144 @@
+# 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.
+
+name: Inference Python Benchmarks Dataflow
+
+on:
+  issue_comment:
+    types: [created]
+  schedule:
+    - cron: '50 3 * * *'
+  workflow_dispatch:
+
+#Setting explicit permissions for the action to avoid the default permissions which are `write-all` in case of pull_request_target event
+permissions:
+  actions: write
+  pull-requests: read
+  checks: read
+  contents: read
+  deployments: read
+  id-token: none
+  issues: read
+  discussions: read
+  packages: read
+  pages: read
+  repository-projects: read
+  security-events: read
+  statuses: read
+
+# This allows a subsequently queued workflow run to interrupt previous runs
+concurrency:
+  group: '${{ github.workflow }} @ ${{ github.event.issue.number || github.sha || github.head_ref || github.ref }}-${{ github.event.schedule || github.event.comment.body || github.event.sender.login }}'
+  cancel-in-progress: true
+
+env:
+  GRADLE_ENTERPRISE_ACCESS_KEY: ${{ secrets.GE_ACCESS_TOKEN }}
+  GRADLE_ENTERPRISE_CACHE_USERNAME: ${{ secrets.GE_CACHE_USERNAME }}
+  GRADLE_ENTERPRISE_CACHE_PASSWORD: ${{ secrets.GE_CACHE_PASSWORD }}
+
+jobs:
+  beam_Inference_Python_Benchmarks_Dataflow:
+    if: |
+      github.event_name == 'workflow_dispatch' ||
+      github.event_name == 'schedule' ||
+      github.event.comment.body == 'Run Inference Benchmarks'
+    runs-on: [self-hosted, ubuntu-20.04, main]
+    timeout-minutes: 900
+    name: ${{ matrix.job_name }} (${{ matrix.job_phrase }})
+    strategy:
+      matrix:
+        job_name: ["beam_Inference_Python_Benchmarks_Dataflow"]
+        job_phrase: ["Run Inference Benchmarks"]
+    steps:
+      - uses: actions/checkout@v3
+      - name: Setup repository
+        uses: ./.github/actions/setup-action
+        with:
+          comment_phrase: ${{ matrix.job_phrase }}
+          github_token: ${{ secrets.GITHUB_TOKEN }}
+          github_job: ${{ matrix.job_name }} (${{ matrix.job_phrase }})
+      - name: Setup Python environment
+        uses: ./.github/actions/setup-environment-action
+        with:
+          python-version: '3.8'
+      - name: Prepare test arguments
+        uses: ./.github/actions/test-arguments-action
+        with:
+          test-type: load
+          test-language: python
+          argument-file-paths: |
+            ${{ github.workspace }}/.github/workflows/load-tests-job-configs/beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Vision_Classification_Resnet_101.txt
+            ${{ github.workspace }}/.github/workflows/load-tests-job-configs/beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Imagenet_Classification_Resnet_152.txt
+            ${{ github.workspace }}/.github/workflows/load-tests-job-configs/beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Language_Modeling_Bert_Base_Uncased.txt
+            ${{ github.workspace }}/.github/workflows/load-tests-job-configs/beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Language_Modeling_Bert_Large_Uncased.txt
+            ${{ github.workspace }}/.github/workflows/load-tests-job-configs/beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Imagenet_Classification_Resnet_152_Tesla_T4_GPU.txt
+      # The env variables are created and populated in the test-arguments-action as "<github.job>_test_arguments_<argument_file_paths_index>"
+      - name: get current time
+        run: echo "NOW_UTC=$(date '+%m%d%H%M%S' --utc)" >> $GITHUB_ENV
+      - name: run Pytorch Vision Classification with Resnet 101
+        uses: ./.github/actions/gradle-command-self-hosted-action
+        timeout-minutes: 180
+        with:
+          gradle-command: :sdks:python:apache_beam:testing:load_tests:run
+          arguments: |
+            -PloadTest.mainClass=apache_beam.testing.benchmarks.inference.pytorch_image_classification_benchmarks \
+            -Prunner=DataflowRunner \
+            -PpythonVersion=3.8 \
+            -PloadTest.requirementsTxtFile=apache_beam/ml/inference/torch_tests_requirements.txt \
+            '-PloadTest.args=${{ env.beam_Inference_Python_Benchmarks_Dataflow_test_arguments_1 }} --job_name=benchmark-tests-pytorch-imagenet-python-101-${{env.NOW_UTC}} --output=gs://temp-storage-for-end-to-end-tests/torch/result_resnet101-${{env.NOW_UTC}}.txt' \
+      - name: run Pytorch Imagenet Classification with Resnet 152
+        uses: ./.github/actions/gradle-command-self-hosted-action
+        timeout-minutes: 180
+        with:
+          gradle-command: :sdks:python:apache_beam:testing:load_tests:run
+          arguments: |
+            -PloadTest.mainClass=apache_beam.testing.benchmarks.inference.pytorch_image_classification_benchmarks \
+            -Prunner=DataflowRunner \
+            -PpythonVersion=3.8 \
+            -PloadTest.requirementsTxtFile=apache_beam/ml/inference/torch_tests_requirements.txt \
+            '-PloadTest.args=${{ env.beam_Inference_Python_Benchmarks_Dataflow_test_arguments_2 }} --job_name=benchmark-tests-pytorch-imagenet-python-152-${{env.NOW_UTC}} --output=gs://temp-storage-for-end-to-end-tests/torch/result_resnet152-${{env.NOW_UTC}}.txt' \
+      - name: run Pytorch Language Modeling using Hugging face bert-base-uncased model
+        uses: ./.github/actions/gradle-command-self-hosted-action
+        timeout-minutes: 180
+        with:
+          gradle-command: :sdks:python:apache_beam:testing:load_tests:run
+          arguments: |
+            -PloadTest.mainClass=apache_beam.testing.benchmarks.inference.pytorch_language_modeling_benchmarks \
+            -Prunner=DataflowRunner \
+            -PpythonVersion=3.8 \
+            -PloadTest.requirementsTxtFile=apache_beam/ml/inference/torch_tests_requirements.txt \
+            '-PloadTest.args=${{ env.beam_Inference_Python_Benchmarks_Dataflow_test_arguments_3 }} --job_name=benchmark-tests-pytorch-language-modeling-bert-base-uncased-${{env.NOW_UTC}} --output=gs://temp-storage-for-end-to-end-tests/torch/result_bert_base_uncased-${{env.NOW_UTC}}.txt' \
+      - name: run Pytorch Langauge Modeling using Hugging Face bert-large-uncased model
+        uses: ./.github/actions/gradle-command-self-hosted-action
+        timeout-minutes: 180
+        with:
+          gradle-command: :sdks:python:apache_beam:testing:load_tests:run
+          arguments: |
+            -PloadTest.mainClass=apache_beam.testing.benchmarks.inference.pytorch_language_modeling_benchmarks \
+            -Prunner=DataflowRunner \
+            -PpythonVersion=3.8 \
+            -PloadTest.requirementsTxtFile=apache_beam/ml/inference/torch_tests_requirements.txt \
+            '-PloadTest.args=${{ env.beam_Inference_Python_Benchmarks_Dataflow_test_arguments_4 }} --job_name=benchmark-tests-pytorch-language-modeling-bert-large-uncased-${{env.NOW_UTC}} --output=gs://temp-storage-for-end-to-end-tests/torch/result_bert_large_uncased-${{env.NOW_UTC}}.txt' \
+      - name: run Pytorch Imagenet Classification with Resnet 152 with Tesla T4 GPU
+        uses: ./.github/actions/gradle-command-self-hosted-action
+        timeout-minutes: 180
+        with:
+          gradle-command: :sdks:python:apache_beam:testing:load_tests:run
+          arguments: |
+            -PloadTest.mainClass=apache_beam.testing.benchmarks.inference.pytorch_image_classification_benchmarks \
+            -Prunner=DataflowRunner \
+            -PpythonVersion=3.8 \
+            -PloadTest.requirementsTxtFile=apache_beam/ml/inference/torch_tests_requirements.txt \
+            '-PloadTest.args=${{ env.beam_Inference_Python_Benchmarks_Dataflow_test_arguments_5 }} --job_name=benchmark-tests-pytorch-imagenet-python-gpu-${{env.NOW_UTC}} --output=gs://temp-storage-for-end-to-end-tests/torch/result_resnet152_gpu-${{env.NOW_UTC}}.txt'
diff --git a/.../beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Imagenet_Classification_Resnet_152.txt b/.../beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Imagenet_Classification_Resnet_152.txt
@@ -0,0 +1,34 @@
+#  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.
+
+--region=us-central1
+--machine_type=n1-standard-2
+--num_workers=75
+--disk_size_gb=50
+--autoscaling_algorithm=NONE
+--staging_location=gs://temp-storage-for-perf-tests/loadtests
+--temp_location=gs://temp-storage-for-perf-tests/loadtests
+--requirements_file=apache_beam/ml/inference/torch_tests_requirements.txt
+--publish_to_big_query=true
+--metrics_dataset=beam_run_inference
+--metrics_table=torch_inference_imagenet_results_resnet152
+--input_options={}
+--influx_measurement=torch_inference_imagenet_resnet152
+--pretrained_model_name=resnet152
+--device=CPU
+--input_file=gs://apache-beam-ml/testing/inputs/openimage_50k_benchmark.txt
+--model_state_dict_path=gs://apache-beam-ml/models/torchvision.models.resnet152.pth
+--runner=DataflowRunner
diff --git a/...ce_Python_Benchmarks_Dataflow_Pytorch_Imagenet_Classification_Resnet_152_Tesla_T4_GPU.txt b/...ce_Python_Benchmarks_Dataflow_Pytorch_Imagenet_Classification_Resnet_152_Tesla_T4_GPU.txt
@@ -0,0 +1,36 @@
+#  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.
+
+--region=us-central1
+--machine_type=n1-standard-2
+--num_workers=75
+--disk_size_gb=50
+--autoscaling_algorithm=NONE
+--staging_location=gs://temp-storage-for-perf-tests/loadtests
+--temp_location=gs://temp-storage-for-perf-tests/loadtests
+--requirements_file=apache_beam/ml/inference/torch_tests_requirements.txt
+--publish_to_big_query=true
+--metrics_dataset=beam_run_inference
+--metrics_table=torch_inference_imagenet_results_resnet152_tesla_t4
+--input_options={}
+--influx_measurement=torch_inference_imagenet_resnet152_tesla_t4
+--pretrained_model_name=resnet152
+--device=GPU
+--experiments=worker_accelerator=type:nvidia-tesla-t4;count:1;install-nvidia-driver
+--sdk_container_image=us.gcr.io/apache-beam-testing/python-postcommit-it/tensor_rt:latest
+--input_file=gs://apache-beam-ml/testing/inputs/openimage_50k_benchmark.txt
+--model_state_dict_path=gs://apache-beam-ml/models/torchvision.models.resnet152.pth
+--runner=DataflowRunner
diff --git a/...beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Language_Modeling_Bert_Base_Uncased.txt b/...beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Language_Modeling_Bert_Base_Uncased.txt
@@ -0,0 +1,34 @@
+#  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.
+
+--region=us-central1
+--machine_type=n1-standard-2
+--num_workers=250
+--disk_size_gb=50
+--autoscaling_algorithm=NONE
+--staging_location=gs://temp-storage-for-perf-tests/loadtests
+--temp_location=gs://temp-storage-for-perf-tests/loadtests
+--requirements_file=apache_beam/ml/inference/torch_tests_requirements.txt
+--publish_to_big_query=true
+--metrics_dataset=beam_run_inference
+--metrics_table=torch_language_modeling_bert_base_uncased
+--input_options={}
+--influx_measurement=torch_language_modeling_bert_base_uncased
+--device=CPU
+--input_file=gs://apache-beam-ml/testing/inputs/sentences_50k.txt
+--bert_tokenizer=bert-base-uncased
+--model_state_dict_path=gs://apache-beam-ml/models/huggingface.BertForMaskedLM.bert-base-uncased.pth
+--runner=DataflowRunner
diff --git a/...eam_Inference_Python_Benchmarks_Dataflow_Pytorch_Language_Modeling_Bert_Large_Uncased.txt b/...eam_Inference_Python_Benchmarks_Dataflow_Pytorch_Language_Modeling_Bert_Large_Uncased.txt
@@ -0,0 +1,34 @@
+#  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.
+
+--region=us-central1
+--machine_type=n1-standard-2
+--num_workers=250
+--disk_size_gb=50
+--autoscaling_algorithm=NONE
+--staging_location=gs://temp-storage-for-perf-tests/loadtests
+--temp_location=gs://temp-storage-for-perf-tests/loadtests
+--requirements_file=apache_beam/ml/inference/torch_tests_requirements.txt
+--publish_to_big_query=true
+--metrics_dataset=beam_run_inference
+--metrics_table=torch_language_modeling_bert_large_uncased
+--input_options={}
+--influx_measurement=torch_language_modeling_bert_large_uncased
+--device=CPU
+--input_file=gs://apache-beam-ml/testing/inputs/sentences_50k.txt
+--bert_tokenizer=bert-large-uncased
+--model_state_dict_path=gs://apache-beam-ml/models/huggingface.BertForMaskedLM.bert-large-uncased.pth
+--runner=DataflowRunner
diff --git a/...gs/beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Vision_Classification_Resnet_101.txt b/...gs/beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Vision_Classification_Resnet_101.txt
@@ -0,0 +1,34 @@
+#  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.
+
+--region=us-central1
+--machine_type=n1-standard-2
+--num_workers=75
+--disk_size_gb=50
+--autoscaling_algorithm=NONE
+--staging_location=gs://temp-storage-for-perf-tests/loadtests
+--temp_location=gs://temp-storage-for-perf-tests/loadtests
+--requirements_file=apache_beam/ml/inference/torch_tests_requirements.txt
+--publish_to_big_query=true
+--metrics_dataset=beam_run_inference
+--metrics_table=torch_inference_imagenet_results_resnet101
+--input_options={}
+--influx_measurement=torch_inference_imagenet_resnet101
+--pretrained_model_name=resnet101
+--device=CPU
+--input_file=gs://apache-beam-ml/testing/inputs/openimage_50k_benchmark.txt
+--model_state_dict_path=gs://apache-beam-ml/models/torchvision.models.resnet101.pth
+--runner=DataflowRunner