port pytorch/onnx yolov5 model to run on a Jetson Nano
ipynb is for testing pytorch code and exporting onnx models using Google Colab
python code runs numpy/tensorrt implementation on Jetson Nano
├── python
│ ├── lib
| ├── demo.py
| ├── Processor.py
| ├── Visualizer.py
| ├── classes.py
| └── models
| ├── yolov5s-simple-32.trt
| ├── yolov5s-simple-16.trt
| └── yolov5s-simple.onnx
│ └── export_tensorrt.py
- convert yolov5 onnx model to tensorrt
- pre-process image
- run inference against input using tensorrt engine
- post process output (forward pass)
- apply nms thresholding on candidate boxes
- visualize results
python3 export_tensorrt.py --help
usage: export_tensorrt.py [-h] [-m MODEL] [-fp FLOATINGPOINT] [-o OUTPUT]
compile Onnx model to TensorRT
optional arguments:
-h, --help show this help message and exit
-m MODEL, --model MODEL
onnx file location inside ./lib/models
-fp FLOATINGPOINT, --floatingpoint FLOATINGPOINT
floating point precision. 16 or 32
-o OUTPUT, --output OUTPUT
name of trt output file
python3 demo.py --image=./path/to/image.jpg --model=./path/to/model.trt
for now, only testing initial inference performance
nms, and post processing are slow rn
| model | fp precision | input size | time (ms) |
|---|---|---|---|
| small-simple | 32 | 640x640x3 | 221 ms |
| small-simple | 16 | 640x640x3 | ? |