YOLO for PYNQ-Z1 and Intel/Movidius Neural Compute Stick (NCS)
Using this repo on your PYNQ-Z1
To use this code on your PYNQ-Z1, just follow these steps:
Install NCSDK in API-mode on your PYNQ-Z1 as explained here: Setting up the PYNQ-Z1 for the Intel Movidius NCS
Clone this repo onto your PYNQ-Z1 in this directory:
Boot the PYNQ-Z1, open Jupyter in a web browser (http://pynq:9090) and open one of the notebooks
- Camera App is working.
- YOLOv1 Tiny is working.
Protobuf Model files
Download Pretrained Caffe Models to ./weights/
- Compile .prototxt and corresponding .caffemodel (with the same name) to get NCS graph file. For example: "mvNCCompile prototxt/yolo_tiny_deploy.prototxt -w weights/yolo_tiny_deploy.caffemodel -s 12"
- The compiled binary file "graph" has to be in main folder after this step.
Single Image Script
- Run "yolo_example.py" to process a single image. For example: "python3 py_examples/yolo_example.py images/dog.jpg" to get detections as below.
Camera Input Script
- Run "object_detection_app.py" to process a videos from your camera. For example: "python3 py_examples/object_detection_app.py" to get camera detections as below.
- Modify script arguments if needed.
- Press "q" to exit app.