Skip to content

Latest commit

 

History

History
53 lines (33 loc) · 1.49 KB

README.md

File metadata and controls

53 lines (33 loc) · 1.49 KB

docker-edgetpu-compiler

The Edge TPU Compiler (edgetpu_compiler) is a command line tool that compiles a TensorFlow Lite model (.tflite file) into a file that's compatible with the Edge TPU (_edgetpu.tflite file).

Because the Edge TPU compiler currently only runs on Debian based Linux systems, we can run the compiler through a Docker container instead when on Windows or MacOS.

Requirements

Verify that Docker is installed by running command:

docker version

This container has been tested and verified on Windows 10 and Docker CE 19.03.8

Build container

docker build --tag edgetpu_compiler https://github.com/tomassams/docker-edgetpu-compiler.git

Run compiler

The easiest is to run the command from the directory containing your .tflite model.

This command will output a compiled model (_edgetpu.tflite file) to the same directory.

Windows - PowerShell

docker run -it --rm -v ${pwd}:/home/edgetpu edgetpu_compiler edgetpu_compiler YOUR_MODEL_FILE.tflite

MacOS and Linux - Terminal

docker run -it --rm -v $(pwd):/home/edgetpu edgetpu_compiler edgetpu_compiler YOUR_MODEL_FILE.tflite

You can also run it from inside the container shell.

Replace /your/path with the path to where you keep the .tflite model, e.g. D:\model or /d/model:

docker run -it --rm -v /your/host/path:/home/edgetpu edgetpu_compiler
edgetpu_compiler YOUR_MODEL_FILE.tflite