Running the resnet50 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=image-classification,dataset,imagenet,aux
ck install package --tags=image-classification,dataset,imagenet,val
cp ~/CK-TOOLS/dataset-imagenet-ilsvrc2012-aux-from.dividiti/val.txt ~/CK-TOOLS/dataset-imagenet-ilsvrc2012-val-min/val_map.txt
wget https://zenodo.org/record/2535873/files/resnet50_v1.pb
export DATA_DIR=${HOME}/CK-TOOLS/dataset-imagenet-ilsvrc2012-val-min
export MODEL_DIR=${HOME}/models
cd ~/src/inference/vision/classification_and_detection
./run_local.sh tf resnet50
Alternatively if on a machine in Linaro’s Cambridge colo
Mount the NFS drive with the already downloaded data Cambridge Colo data server
export DATA_DIR=/mnt/datasets/data/CK-TOOLS/dataset-imagenet-ilsvrc2012-val-min export MODEL_DIR=/mnt/datasets/data/models cd ~/src/inference/vision/classification_and_detection ./run_local.sh tf resnet50