ssd-resnet34

Running the ssd-resnet34 test on TensorFlow

Ensure the virtual environment is prepared as described in MLPerf Inference

Activate the virtual environment first and then

cd ~
mkdir models
cd models
ck install package --tags=object-detection,dataset,coco,val # Choose [0] the 2017 data
wget https://zenodo.org/record/3345892/files/tf_ssd_resnet34_22.1.zip
unzip tf_ssd_resnet34_22.1.zip
cp tf_ssd_resnet34_22.1/resnet34_tf.22.1.pb .
cd ~/src/inference/vision/classification_and_detection
python tools/ssd-nhwc.py ~/models/resnet34_tf.22.1.pb
rm ~/models/resnet34_tf.22.1.pb
mv ~/models/resnet34_tf.22.1.pb.patch ~/models/resnet34_tf.22.1.pb
export DATA_DIR=~/CK-TOOLS/dataset-coco-2017-val
export MODEL_DIR=${HOME}/models
./run_local.sh tf ssd-resnet34 --data-format NHWC

Alternatively if you are on a machine in Linaro’s Cambridge colo then you can do this

Mount the NFS drive with the already downloaded data Cambridge Colo data server

export DATA_DIR=/mnt/datasets/data/CK-TOOLS/dataset-coco-2017-val
export MODEL_DIR=/mnt/datasets/data/models
cd ~/src/inference/vision/classification_and_detection
./run_local.sh tf ssd-resnet34 --data-format NHWC