Skip to content

Benchmark models

Larq Compute Engine provides prebuilt binaries to benchmark Larq converted models on Android phones or 64-bit ARM based systems like the Raspberry Pi.

  1. Download the prebuilt benchmarking binary from the latest release:

    wget https://github.com/larq/compute-engine/releases/download/v0.6.0/lce_benchmark_model_aarch64 -O lce_benchmark_model
    

  2. Make the binary executable:

    chmod +x lce_benchmark_model
    

  3. Benchmark the converted .tflite model:

    ./lce_benchmark_model --graph=quicknet.tflite --num_threads=4
    
    Add --help to the command for a detailed description of the available benchmarking options.

  1. Install the Android Debug Bridge (adb) on your host machine. E.g. on macOS you can install it via brew:

    brew cask install android-platform-tools
    

  2. Follow the instructions here to enable "USB debugging" on your Android phone.

  3. Connect your phone and run the following command to confirm that your host computer recognises your phone:

    adb devices
    

  4. Download the prebuilt Android benchmarking binary from the latest release:

    wget https://github.com/larq/compute-engine/releases/download/v0.6.0/lce_benchmark_model_android_arm64 -O lce_benchmark_model
    

  5. Transfer the LCE inference binary to your phone:

    adb push lce_benchmark_model /data/local/tmp
    

  6. Transfer the converted .tflite model file to your phone:

    adb push quicknet.tflite /data/local/tmp
    

  7. Make the binary executable:

    adb shell chmod +x /data/local/tmp/lce_benchmark_model
    

  8. Benchmark the model:

    adb shell /data/local/tmp/lce_benchmark_model \
        --graph=/data/local/tmp/quicknet.tflite \
        --num_threads=4
    
    Add --help to the command for a detailed description of the available benchmarking options.