- Install dependencies (assumes Ubuntu - Debian-ish env)
sudo apt install libcamera-dev
sudo apt install ninja-build
# for test_server.py
pip install opencv-python
pip install flask
- Build and run:
git clone <this repo>
cd porch
porch~$ meson setup build
porch~$ ninja -C build # or `meson compile -C build`
porch~$ cd build
porch/build~$ ./porch
- TensorFlow 2 Detection Model Zoo
https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf2_detection_zoo.md
git https://github.com/tensorflow/models.git
wget http://download.tensorflow.org/models/object_detection/tf2/20200711/<MODEL_NAME>.tar.gz
tar -xzvf path/to/<MODEL_NAME>.tar.gz -C destination_directory/
replace MODEL_NAME with desired model from tf2 model zoo. Define variables pipline_config, model_dir, and PATH_TO_LABELS with correct filepaths
More TF links:
https://tutorials-raspberrypi.com/using-tensorflow-lite-with-google-coral-tpu-on-raspberry-pi-4/
https://github.com/google-coral/edgetpu-platforms
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent