No public description

PiperOrigin-RevId: 929873126

Author: tensorflower-gardener

official/projects/waste_identification_ml/Deploy/pet_grading_cloud_deployment/constants.py

@@ -0,0 +1,61 @@
+# Copyright 2026 The TensorFlow Authors. All Rights Reserved.
+#
+# Licensed 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.
+
+"""Constants for the pet grading cloud deployment."""
+
+# Specify all the input paths
+DINOV3_REPO_DIR = './dinov3'
+CLASSIFIER_CHECKPOINT_PATH = './model_weights/best_pet_grading_model.pth'
+SAM3_CHECKPOINT_PATH = './model_weights/sam3_original_weights_sam3.pt'
+
+# SAM3 configurations.
+BATCH_SIZE = 10
+DETECTION_PROMPT = 'bottles and containers'
+BOTTLE_EXTRACTION_CONFIDENCE_THRESHOLD = 0.5
+BOTTLE_EXTRACTION_SCORE_THRESHOLD = 0.20
+BOTTLE_EXTRACTION_CONTAINMENT_THRESHOLD = 0.98
+BOTTLE_EXTRACTION_MAX_SHORT_SIDE = 1024
+BOTTLE_EXTRACTION_CROP_SIZE = (256, 256)
+
+
+# Tracking configurations.
+BYTETRACK_MINIMUM_IOU_THRESHOLD = 0.1
+BYTETRACK_MINIMUM_CONSECUTIVE_FRAMES = 2
+
+# Classifier configuration. Must match what was used at training time.
+DINOV3_MODEL_NAME = 'dinov3_vitl16'
+CLASSIFICATION_THRESHOLD = 50
+CLASS_NAMES = (
+    'brown_bottles_grade3',
+    'clean_PET_cold_drink_bottles_with_label_cap_ring_grade1',
+    'clean_PET_cold_drink_bottles_without_label_cap_ring_grade1',
+    'clean_PET_mango_juice_bottles_without_label_cap_ring_grade1',
+    'clean_PET_water_bottles_with_label_cap_ring_grade1',
+    'clean_PET_water_bottles_without_label_cap_ring_grade1',
+    'clean_jars_grade1',
+    'clean_liquor_bottles_without_label_cap_ring_grade1',
+    'coloured_PET_bottles_grade3',
+    'dirt_PET_cold_drink_bottles_with_label_cap_ring_grade3',
+    'dirt_PET_cold_drink_bottles_without_label_cap_ring_grade3',
+    'dirt_PET_mango_juice_bottles_without_label_cap_ring_grade3',
+    'dirt_PET_water_bottles_with_label_cap_ring_grade3',
+    'dirt_PET_water_bottles_without_label_cap_ring_grade3',
+    'dirt_jars_grade3',
+    'dirt_liquor_bottles_without_label_cap_ring_grade3',
+    'full_sleeved_bottles_grade3',
+    'green_bottles_grade3',
+    'liquor_bottles_with_label_cap_ring_grade3',
+    'non_food_bottles_grade3',
+    'partially_sleeved_mango_juice_bottles_with_label_cap_ring_grade3',
+)

official/projects/waste_identification_ml/Deploy/pet_grading_cloud_deployment/requirements.sh

@@ -0,0 +1,38 @@
+#!/bin/bash
+
+# Summary
+cat << EOF
+This script sets up the environment for running ML models by ensuring Bash
+execution, installing system dependencies, setting up a virtual environment,
+installing ML packages, and cloning TensorFlow Model Garden.
+EOF
+
+# Ensure the script is executed with /bin/bash
+if [ -z "$BASH_VERSION" ]; then
+  exec /bin/bash "$0" "$@"
+fi
+
+# update linux packages
+sudo apt-get update -y
+
+# Create a virtual environment and install packages
+sudo apt-get install -y python3-venv python3-pip
+
+# Clone dinvov3 repo
+git clone https://github.com/facebookresearch/dinov3.git
+
+# Download Model weights
+mkdir -p ./model_weights
+[ -f ./model_weights/sam3_original_weights_sam3.pt ] || wget -P ./model_weights https://storage.googleapis.com/tf_model_garden/vision/waste_identification_ml/dairy_product_packet_detection/sam3_original_weights_sam3.pt
+[ -f ./model_weights/best_pet_grading_model.pth ] || wget -P ./model_weights https://storage.googleapis.com/tf_model_garden/vision/waste_identification_ml/dairy_product_packet_detection/best_pet_grading_model.pth
+
+python3.10 -m venv myenv
+source myenv/bin/activate
+
+echo "Activated python environment, installing dependencies."
+
+pip install -r requirements.txt
+pip install numpy==1.26.4
+
+deactivate
+echo "Environment setup is complete."

official/projects/waste_identification_ml/Deploy/pet_grading_cloud_deployment/requirements.txt

@@ -0,0 +1,27 @@
+--extra-index-url https://download.pytorch.org/whl/cu124
+sam3 @ git+https://github.com/facebookresearch/sam3.git@967fdd651f71ca14949122fed4c918a778ca9334
+natsort
+absl-py==2.4.0
+einops==0.8.2
+google-api-core==2.30.3
+google-auth==2.49.2
+google-auth-oauthlib==1.3.1
+google-cloud-bigquery==3.41.0
+google-cloud-core==2.5.1
+google-cloud-storage==3.10.1
+huggingface_hub==1.11.0
+opencv-python>=4.8.0,<4.10.0
+pandas==2.3.3
+pandas-gbq==0.35.0
+pycocotools
+scikit-image
+scikit-learn
+scipy
+seaborn
+supervision
+matplotlib
+termcolor
+torch==2.6.0+cu124
+torchvision==0.21.0+cu124
+torchmetrics
+trackers

official/projects/waste_identification_ml/Deploy/pet_grading_cloud_deployment/run_images.sh

@@ -0,0 +1,40 @@
+#!/bin/bash
+
+cat << EOF
+This script automates the execution of an Circularnet pipeline for image
+processing.
+Steps Performed:
+ 1. Activates the Python virtual environment named 'myenv'.
+ 2. Validates successful activation of the virtual environment.
+ 3. Executes the 'pipeline_images.py' script with the following parameters:
+ Parameters:
+ --input_directory         : GCS directory where the input images are stored for
+                             inference.
+ --output_directory        : GCS directory where the model inference outputs will be
+                             saved.
+ --model_name              : Name of the model to download and use for inference.
+ --threshold               : Confidence threshold for detections during
+                             inference.
+ --project_id              : Google Cloud Project ID for BigQuery operations.
+ --bq_dataset_id           : BigQuery Dataset ID where results will be stored.
+ --bq_table_id             : BigQuery Table ID where results will be stored.
+ --overwrite               : If set to True, overwrites the pre-existing
+                             BigQuery table.
+EOF
+#Activate the virtual environment
+source myenv/bin/activate
+# Check if the virtual environment is activated
+if [[ "$VIRTUAL_ENV" != "" ]]; then
+    echo "Virtual environment 'myenv' activated successfully."
+else
+    echo "Failed to activate virtual environment. Exiting."
+    exit 1
+fi
+
+python inference_pipeline.py \
+	--input_directory=gs://circularnet_data/tmp/pet_pipeline_test \
+	--output_directory=gs://circularnet_data/tmp/pet_pipeline_test \
+	--project_id=waste-identification-ml-330916 \
+	--bq_dataset_id=pet_dataset \
+	--bq_table_id=pet_pipeline_test \
+	--overwrite=True