Repo contains:

1. Training a simple object detection model (MobileNetv2)
2. Quantising TFLite Model
3. Generating Autotiler Code for AI-Deck implementation

1. Clone repository into aideck-gap8-examples/examples/ai/ folder
2. From a terminal with the docker container, or gap_sdk dev environment, in the aideck-gap8-examples/ folder, execute:
   • docker run --rm -v ${PWD}:/module aideck-with-autotiler tools/build/make-example examples/ai/detection clean model build image
3. From another terminal (outside of the container), use the cfloader to flash the example if you have the gap8 bootloader flashed AIdeck. Change the [CRAZYFLIE URI] with your crazyflie URI like radio://0/90/2M/E7E7E7E726.
   • cfloader flash examples/ai/detection/BUILD/GAP8_V2/GCC_RISCV_FREERTOS/target.board.devices.flash.img deck-bcAI:gap8-fw -w radio://0/90/2M/E7E7E7E726
   • Flashing will take ~10 minutes depending on the model size and it may appear to be stuck at 99%. However, if the crazyflie automatically reboots at 99%, the flashing is usually completed and you can Ctrl+C in the terminal.
   • Reboot the Crazyflie.

When the example is flashing, you should see the GAP8 LED blink fast, which is the bootloader. The example itself can be noticed by a slow blinking LED. You should also receive the detection output in the cfclient console.

1. Download dataset from Roboflow in Pascal VOC format.
2. Extract them in aideck-gap8-examples/examples/ai/detection/images/.
3. Run python3 pascal_csv.py.

• Refer to model.ipynb to create detection_q.tflite model. Make sure tflite model is saved in model/ folder.