PyTorch

This directory contains steps and artifacts to run a PyTorch sample workload on Gramine. Specifically, this example uses a pre-trained model for image classification.

The workload reads an image from a file on disk input.jpg and runs the classifier to detect what object is present in the image. For the image shipping with the workload, the output should be

[('Labrador retriever', 41.58518600463867), ('golden retriever', 16.59165382385254), ('Saluki, gazelle hound', 16.286855697631836), ('whippet', 2.853914976119995), ('Ibizan hound, Ibizan Podenco', 2.3924756050109863)]

With high probability (41%) the classifier detected the image to contain a Labrador retriever.

Pre-requisites

The following steps should suffice to run the workload on a stock Ubuntu 18.04 installation.

• sudo apt install libnss-mdns libnss-myhostname to install additional DNS-resolver libraries.
• sudo apt install python3-pip lsb-release to install pip and lsb_release. The former is required to install additional Python packages while the latter is used by the Makefile.
• pip3 install --user torchvision pillow to install the torchvision and pillow Python packages and their dependencies (usually in $HOME/.local). WARNING: This downloads several hundred megabytes of data!
• python3 download-pretrained-model.py to download and save the pre-trained model. WARNING: this downloads about 200MB of data!

Build

Run make to build the non-SGX version and make SGX=1 to build the SGX version.

Run

Execute any one of the following commands to run the workload:

• natively: python3 pytorchexample.py
• Gramine w/o SGX: gramine-direct ./pytorch ./pytorchexample.py
• Gramine with SGX: gramine-sgx ./pytorch ./pytorchexample.py