https://github.com/LinhanDai/yolov9-tensorrt
YOLOv9 Tensorrt deployment acceleration，provide two implementation methods: C++and Python🔥🔥🔥
https://github.com/LinhanDai/yolov9-tensorrt
cpp cuda-programming python tensorrt yolov9
Last synced: 3 months ago
JSON representation
YOLOv9 Tensorrt deployment acceleration，provide two implementation methods: C++and Python🔥🔥🔥
- Host: GitHub
- URL: https://github.com/LinhanDai/yolov9-tensorrt
- Owner: LinhanDai
- Created: 2024-02-23T08:15:40.000Z (over 1 year ago)
- Default Branch: master
- Last Pushed: 2024-02-29T05:28:26.000Z (over 1 year ago)
- Last Synced: 2024-10-27T20:18:53.041Z (8 months ago)
- Topics: cpp, cuda-programming, python, tensorrt, yolov9
- Language: C++
- Homepage:
- Size: 1.07 MB
- Stars: 31
- Watchers: 3
- Forks: 5
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
YOLOv9 tensorrt deployment
This repository provides an API for accelerating inference deployment, with two open interface implementation: C++ and Python. C++also provides the use of CUDA programming to accelerate YOLOv9 model preprocessing and post-processing
to pursue faster model inference speed🔥🔥🔥
## ⭐ Build
1. Export onnx
Clone [YOLOv9](https://github.com/WongKinYiu/yolov9) code repository, download the original model provided by the repository, or train your own model, such as [yolov9-c.pt](https://objects.githubusercontent.com/github-production-release-asset-2e65be/759338070/c8ca43f2-0d2d-4aa3-a074-426505bfbfb1?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAVCODYLSA53PQK4ZA%2F20240223%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20240223T073054Z&X-Amz-Expires=300&X-Amz-Signature=db76944695e398168b222b502bb019a301336e5b5dc74db31604699b8f837a9b&X-Amz-SignedHeaders=host&actor_id=45328395&key_id=0&repo_id=759338070&response-content-disposition=attachment%3B%20filename%3Dyolov9-c.pt&response-content-type=application%2Foctet-stream)
``` shell
# export onnx
python export.py --weights yolov9-c.pt --simplify --include "onnx"
```
2. Setup
Place the exported onnx file in the "yolov9-tensorrt/configs" folder and configure the relevant parameters through the "yolov9-tensorrt/configs/yolov9.yaml" file
``` shell
# move onnx
cd yolov9-Tensorrt
mv yolov9-c.onnx ./configs
```
Modify parameter configuration in configs/yolov9-yaml
``` shell
# modify configuration in configs/yolov9.yaml
confTreshold: 0.25 #Detection confidence threshold
nmsTreshold : 0.45 #nms threshold
maxSupportBatchSize: 1 #support max input batch size
quantizationInfer: "FP16" #support FP32 or FP16 quantization
onnxFile: "yolov9-c.onnx" # The currently used onnx model file
engineFile: "yolov9-c.engine" # Automatically generate file names for the Tensorrt inference engine
```
3. Build project
``` shell
mkdir build
cd build
cmake ..
make -j4
```
4. python API
Modify parameter configuration in configs/yolov9py-yaml
``` shell
# modify configuration in configs/yolov9py.yaml
confTreshold: 0.3 # detect treshold
nmsThreshold: 0.45 #nms treshold
quantizationInfer: "FP16" #FP32 or FP16
onnxFile: "yolov9-c.onnx" # The currently used onnx model file
engineFile: "yolov9-c.engine" # Automatically generate file names for the Tensorrt inference engine
```
## 🌠 Run demo
The first run will generate the inference engine ".engine" file in the configs folder. If the inference engine has already been generated, it will not be generated again
run c++ demo API
``` shell
# run images floder
./demo ../data
```
run python demo API
``` shell
# run images floder
python yolov9_trt.py --configs configs --yaml_file yolov9py.yaml --data data
```
## 👏 Acknowledgement
This project is based on the following awesome projects:
- [Yolov9](https://github.com/WongKinYiu/yolov9) - YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information.
- [TensorRT](https://github.com/NVIDIA/TensorRT/tree/release/8.6/samples) - TensorRT samples and api documentation.
## 🤗 Citation
```bibtex
@article{wang2024yolov9,
title={{YOLOv9}: Learning What You Want to Learn Using Programmable Gradient Information},
author={Wang, Chien-Yao and Liao, Hong-Yuan Mark},
booktitle={arXiv preprint arXiv:2402.13616},
year={2024}
}
```