Skip to content

Latest commit

 

History

History
71 lines (59 loc) · 3.81 KB

SSDMobileNetV2.md

File metadata and controls

71 lines (59 loc) · 3.81 KB

Tensorflow SSD MobileNet v2

Setup AI Model Efficiency Toolkit (AIMET)

Please install and setup AIMET before proceeding further. This evaluation was run using AIMET 1.22.2 for TensorFlow 1.15 i.e. please set release_tag="1.22.2" and AIMET_VARIANT="tf_gpu_tf115" in the above instructions.

Experiment Setup

Additional dependencies:

pip install pycocotools
pip install --upgrade tf_slim
pip install git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI

Download and install protoc

protoc is a standalone binary for the Google protobuf compiler.

  • we use protoc in the version 3.14.0, use the following commands to download and install protoc-3.14.0-linux-x86_64.
PROTOC_ZIP=protoc-3.14.0-linux-x86_64.zip
curl -OL https://github.com/protocolbuffers/protobuf/releases/download/v3.14.0/$PROTOC_ZIP
unzip protoc-3.14.0-linux-x86_64.zip
export PATH=<path to protoc>/protoc-3.14.0-linux-x86_64/bin:$PATH

Clone TensorFlow model zoo as FP32 source

git clone https://github.com/tensorflow/models.git
git checkout master
cd models/research
protoc object_detection/protos/*.proto --python_out=.

Append the repo location to your PYTHONPATH by doing the following:

export PYTHONPATH=<path to tensorflow models repo>/models/research:$PYTHONPATH

Model checkpoint for AIMET optimization

  • Downloading of model checkpoints is handled by evaluation script.
  • SSD MobileNet v2 checkpoint used for AIMET quantization can be downloaded from the Releases page.

Dataset

TFRecord format of 2017 COCO dataset is needed. There are two options for download and process MSCOCO dataset:

cd models/research/object_detection/dataset_tools
./download_and_preprocess_mscoco.sh <mscoco_dir>
  • Option 2: If COCO dataset is already available or you want to download COCO dataset separately 
python object_detection/dataset_tools/create_coco_tf_record.py --logtostderr --include_masks --train_image_dir=./MSCOCO_PATH/images/train2017/ --val_image_dir=./MSCOCO_PATH/images/val2017/ --test_image_dir=./MSCOCO_PATH/images/test2017/ --train_annotations_file=./MSCOCO_PATH/annotations/instances_train2017.json --val_annotations_file=./MSCOCO_PATH/annotations/instances_val2017.json --testdev_annotations_file=./MSCOCO_PATH/annotations/image_info_test2017.json --output_dir=./OUTPUT_DIR/

Note: The --include_masks option must be used.

Usage

  • ssd_mobilenet_v2_quanteval.py has two required arguments, an example usage is shown below
python ssd_mobiledet_v2_quanteval.py 
 --dataset-path <path to tfrecord dataset> \
 --annotation-json-file <path to instances json file>/instances_val2017.json \
 --model-to-eval < which model to evaluate.Two options are available: fp32 for evaluating original fp32 model, int8 for evaluating quantized int8 model.>

Quantization configuration

In the evaluation script included, we have manually configured the quantizer ops with the following assumptions:

  • Weight quantization: 8 bits, asymmetric quantization
  • Bias parameters are not quantized
  • Activation quantization: 8 bits, asymmetric quantization
  • Model inputs are quantized