Add PyTorch Image Captioning BLIP + CLIP Batch

Author: Amar3tto

.github/workflows/beam_Inference_Python_Benchmarks_Dataflow.yml

@@ -93,6 +93,7 @@ jobs:
             ${{ github.workspace }}/.github/workflows/load-tests-pipeline-options/beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Sentiment_Batch_DistilBert_Base_Uncased.txt
             ${{ github.workspace }}/.github/workflows/load-tests-pipeline-options/beam_Inference_Python_Benchmarks_Dataflow_VLLM_Gemma_Batch.txt
             ${{ github.workspace }}/.github/workflows/load-tests-pipeline-options/beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Image_Object_Detection.txt
+            ${{ github.workspace }}/.github/workflows/load-tests-pipeline-options/beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Image_Object_Captioning.txt
       # The env variables are created and populated in the test-arguments-action as "<github.job>_test_arguments_<argument_file_paths_index>"
       - name: get current time
         run: echo "NOW_UTC=$(date '+%m%d%H%M%S' --utc)" >> $GITHUB_ENV
@@ -190,7 +191,7 @@ jobs:
             -Prunner=DataflowRunner \
             -PpythonVersion=3.10 \
             -PloadTest.requirementsTxtFile=apache_beam/ml/inference/torch_tests_requirements.txt \
-            '-PloadTest.args=${{ 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'
+            '-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-${{env.NOW_UTC}}.txt' \
       - name: run PyTorch Image Object Detection Faster R-CNN ResNet-50 Batch
         uses: ./.github/actions/gradle-command-self-hosted-action
         timeout-minutes: 180
@@ -201,4 +202,15 @@ jobs:
             -Prunner=DataflowRunner \
             -PpythonVersion=3.10 \
             -PloadTest.requirementsTxtFile=apache_beam/ml/inference/pytorch_image_object_detection_requirements.txt \
-            '-PloadTest.args=${{ env.beam_Inference_Python_Benchmarks_Dataflow_test_arguments_9 }} --mode=batch --job_name=benchmark-tests-pytorch-object_detection-batch-${{env.NOW_UTC}} --output_table=apache-beam-testing.beam_run_inference.result_torch_inference_object_detection_batch' \
+            '-PloadTest.args=${{ env.beam_Inference_Python_Benchmarks_Dataflow_test_arguments_9 }} --mode=batch --job_name=benchmark-tests-pytorch-image-object_detection-batch-${{env.NOW_UTC}} --output_table=apache-beam-testing.beam_run_inference.result_torch_inference_image_object_detection_batch' \
+      - name: run PyTorch Image Captioning BLIP + CLIP Batch
+        uses: ./.github/actions/gradle-command-self-hosted-action
+        timeout-minutes: 180
+        with:
+          gradle-command: :sdks:python:apache_beam:testing:load_tests:run
+          arguments: |
+            -PloadTest.mainClass=apache_beam.testing.benchmarks.inference.pytorch_image_captioning_benchmarks \
+            -Prunner=DataflowRunner \
+            -PpythonVersion=3.10 \
+            -PloadTest.requirementsTxtFile=apache_beam/ml/inference/pytorch_image_captioning_requirements.txt \
+            '-PloadTest.args=${{ env.beam_Inference_Python_Benchmarks_Dataflow_test_arguments_10 }} --mode=batch --job_name=benchmark-tests-pytorch-image-captioning-batch-${{env.NOW_UTC}} --output_table=apache-beam-testing.beam_run_inference.result_torch_inference_image_captioning_batch'

.github/workflows/load-tests-pipeline-options/beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Image_Object_Captioning.txt

@@ -0,0 +1,41 @@
+#  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
+--worker_machine_type=n1-standard-4
+--num_workers=50
+--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/pytorch_image_captioning_requirements.txt
+--publish_to_big_query=true
+--metrics_dataset=beam_run_inference
+--metrics_table=result_torch_inference_image_captioning_batch
+--input_options={}
+--influx_measurement=result_torch_inference_image_captioning_batch
+--device=GPU
+--mode=batch
+--input=gs://apache-beam-ml/testing/inputs/openimage_50k_benchmark.txt
+--blip_model_name=Salesforce/blip-image-captioning-base
+--blip_batch_size=4
+--num_captions=5
+--max_new_tokens=30
+--num_beams=5
+--clip_model_name=openai/clip-vit-base-patch32
+--clip_batch_size=8
+--clip_score_normalize=false
+--runner=DataflowRunner
+--experiments=use_runner_v2
+--worker_accelerator=type=nvidia-tesla-t4,count=1,install-nvidia-driver=true

.github/workflows/load-tests-pipeline-options/beam_Inference_Python_Benchmarks_Dataflow_Pytorch_Image_Object_Detection.txt

@@ -24,9 +24,9 @@
 --requirements_file=apache_beam/ml/inference/pytorch_image_object_detection_requirements.txt
 --publish_to_big_query=true
 --metrics_dataset=beam_run_inference
---metrics_table=result_torch_inference_object_detection_batch
+--metrics_table=result_torch_inference_image_object_detection_batch
 --input_options={}
---influx_measurement=result_torch_inference_object_detection_batch
+--influx_measurement=result_torch_inference_image_object_detection_batch
 --pretrained_model_name=fasterrcnn_resnet50_fpn
 --device=GPU
 --mode=batch

.test-infra/tools/refresh_looker_metrics.py

@@ -44,6 +44,7 @@
     ("85", ["268", "269", "270", "271", "272"]),  # PyTorch Sentiment Batch DistilBERT base uncased
     ("86", ["284", "285", "286", "287", "288"]),  # VLLM Batch Gemma
     #TODO: PyTorch Image Object Detection Faster R-CNN ResNet-50 Batch
+    #TODO: PyTorch Image Captioning BLIP + CLIP Batch
 ]