macOS

Alternatively, skip steps 1-3 if you prefer to manually install Python 3 and Git.

1. Install Xcode Command Line Tools

xcode-select --install

2. Install Homebrew

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

After you install it, follow the instructions from the Homebrew installation to set it up.

3. Install Python and dependencies

brew install python@3.9
brew install protobuf

Run each step below in a terminal. Note: If OpenVINO is installed globally, please do not run any of these commands in a terminal where setupvars.sh is sourced.

4. Create a Virtual Environment

Note: If you already installed openvino-dev and activated the openvino_env environment, you can skip to Step 6. If you use Anaconda, please see the Conda guide.

python3 -m venv openvino_env

5. Activate the Environment

source openvino_env/bin/activate

6. Clone the Repository

Note: Using the --depth=1 option for git clone reduces download size.

git clone --depth=1 /~https://github.com/openvinotoolkit/openvino_notebooks.git
cd openvino_notebooks

7. Install the Packages

This step installs OpenVINO and dependencies like Jupyter Lab. First, upgrade pip to the latest version. Then, install the required dependencies.

python -m pip install --upgrade pip wheel setuptools
pip install -r requirements.txt

8. Launch the Notebooks on Chrome or Firefox!

To launch a single notebook, like the Monodepth notebook

jupyter notebook notebooks/201-vision-monodepth/201-vision-monodepth.ipynb

To launch all notebooks in Jupyter Lab

jupyter lab notebooks

In Jupyter Lab, select a notebook from the file browser using the left sidebar. Each notebook is located in a subdirectory within the notebooks directory.

Troubleshooting

• Please note that support for M1/M2 is experimental, and the following notebooks may not work: 002-openvino-api, 004-hello-detection, 103-paddle-to-openvino, 104-model-tools, 105-language-quantize-bert, 107-speech-recognition-quantization, 110-ct-segmentation-quantize, 111-detection-quantization, 112-pytorch-post-training-quantization-nncf, 113-image-classification-quantization, 114-quantization-simplified-mode, 202-vision-superresolution, 203-meter-reader, 204-named-entity-recognition, 206-vision-paddlegan-anime, 207-vision-paddlegan-superresolution, 208-optical-character-recognition, 211-speech-to-text, 214-vision-paddle-classification, 217-vision-deblur, 220-yolov5-accuracy-check-and-quantization, 222-vision-image-colorization, 223-gpt2-text-prediction, 224-3D-segmentation-point-clouds, 226-yolov7-optimization, 227-whisper-subtitles-generation, 228-clip-zero-shot-image-classification, 230-yolo8-optimization, 301-tensorflow-training-openvino, 405-paddle-ocr-webcam, 406-3D-pose-estimation-webcam
• If you cannot run the above notebooks or experience bad performance, please see the community discussion page about using Rosetta* 2. For example, there are known issues with the use of model optimizer, and thus you may need to convert the model on x86 machines in prior.
• For Apple Silicons, we have some bugs and issues that may remain open. Please do report these to our OpenVION Contrib Repos if you find any issues.
• For the webcam notebooks, please use Chrome/Firefox as the browser. Otherwise, you will notice flickering with Safari.
• The built-in webcam performance is poor and has significant drop frame issues. The workaround is to use other webcams or create a virtual webcam with OBS (https://obsproject.com/).
• If you use Anaconda or Miniconda, see the Conda wiki page.