# YOLOv8-TensorRT `YOLOv8` using TensorRT accelerate ! --- [![Build Status](https://img.shields.io/endpoint.svg?url=https%3A%2F%2Factions-badge.atrox.dev%2Fatrox%2Fsync-dotenv%2Fbadge&style=flat)](https://github.com/triple-Mu/YOLOv8-TensorRT) [![Python Version](https://img.shields.io/badge/Python-3.8--3.10-FFD43B?logo=python)](https://github.com/triple-Mu/YOLOv8-TensorRT) [![img](https://badgen.net/badge/icon/tensorrt?icon=azurepipelines&label)](https://developer.nvidia.com/tensorrt) [![C++](https://img.shields.io/badge/CPP-11%2F14-yellow)](https://github.com/triple-Mu/YOLOv8-TensorRT) [![img](https://badgen.net/github/license/triple-Mu/YOLOv8-TensorRT)](https://github.com/triple-Mu/YOLOv8-TensorRT/blob/main/LICENSE) [![img](https://badgen.net/github/prs/triple-Mu/YOLOv8-TensorRT)](https://github.com/triple-Mu/YOLOv8-TensorRT/pulls) [![img](https://img.shields.io/github/stars/triple-Mu/YOLOv8-TensorRT?color=ccf)](https://github.com/triple-Mu/YOLOv8-TensorRT) --- # Prepare the environment 1. Install `CUDA` follow [`CUDA official website`](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#download-the-nvidia-cuda-toolkit). 🚀 RECOMMENDED `CUDA` >= 11.4 2. Install `TensorRT` follow [`TensorRT official website`](https://developer.nvidia.com/nvidia-tensorrt-8x-download). 🚀 RECOMMENDED `TensorRT` >= 8.4 2. Install python requirements. ``` shell pip install -r requirements.txt ``` 3. Install [`ultralytics`](https://github.com/ultralytics/ultralytics) package for ONNX export or TensorRT API building. ``` shell pip install ultralytics ``` 5. Prepare your own PyTorch weight such as `yolov8s.pt` or `yolov8s-seg.pt`. ***NOTICE:*** Please use the latest `CUDA` and `TensorRT`, so that you can achieve the fastest speed ! If you have to use a lower version of `CUDA` and `TensorRT`, please read the relevant issues carefully ! # Normal Usage If you get ONNX from origin [`ultralytics`](https://github.com/ultralytics/ultralytics) repo, you should build engine by yourself. You can only use the `c++` inference code to deserialize the engine and do inference. You can get more information in [`Normal.md`](docs/Normal.md) ! Besides, other scripts won't work. # Export End2End ONNX with NMS You can export your onnx model by `ultralytics` API and add postprocess such as bbox decoder and `NMS` into ONNX model at the same time. ``` shell python3 export-det.py \ --weights yolov8s.pt \ --iou-thres 0.65 \ --conf-thres 0.25 \ --topk 100 \ --opset 11 \ --sim \ --input-shape 1 3 640 640 \ --device cuda:0 ``` #### Description of all arguments - `--weights` : The PyTorch model you trained. - `--iou-thres` : IOU threshold for NMS plugin. - `--conf-thres` : Confidence threshold for NMS plugin. - `--topk` : Max number of detection bboxes. - `--opset` : ONNX opset version, default is 11. - `--sim` : Whether to simplify your onnx model. - `--input-shape` : Input shape for you model, should be 4 dimensions. - `--device` : The CUDA deivce you export engine . You will get an onnx model whose prefix is the same as input weights. ### Just Taste First If you just want to taste first, you can download the onnx model which are exported by `YOLOv8` package and modified by me. [**YOLOv8-n**](https://triplemu.oss-cn-beijing.aliyuncs.com/YOLOv8/ONNX/yolov8n_nms.onnx?OSSAccessKeyId=LTAI5tN1dgmZD4PF8AJUXp3J&Expires=1772936700&Signature=r6HgJTTcCSAxQxD9bKO9qBTtigQ%3D) [**YOLOv8-s**](https://triplemu.oss-cn-beijing.aliyuncs.com/YOLOv8/ONNX/yolov8s_nms.onnx?OSSAccessKeyId=LTAI5tN1dgmZD4PF8AJUXp3J&Expires=1682936722&Signature=JjxQFx1YElcVdsCaMoj81KJ4a5s%3D) [**YOLOv8-m**](https://triplemu.oss-cn-beijing.aliyuncs.com/YOLOv8/ONNX/yolov8m_nms.onnx?OSSAccessKeyId=LTAI5tN1dgmZD4PF8AJUXp3J&Expires=1682936739&Signature=IRKBELdVFemD7diixxxgzMYqsWg%3D) [**YOLOv8-l**](https://triplemu.oss-cn-beijing.aliyuncs.com/YOLOv8/ONNX/yolov8l_nms.onnx?OSSAccessKeyId=LTAI5tN1dgmZD4PF8AJUXp3J&Expires=1682936763&Signature=RGkJ4G2XJ4J%2BNiki5cJi3oBkDnA%3D) [**YOLOv8-x**](https://triplemu.oss-cn-beijing.aliyuncs.com/YOLOv8/ONNX/yolov8x_nms.onnx?OSSAccessKeyId=LTAI5tN1dgmZD4PF8AJUXp3J&Expires=1673936778&Signature=3o%2F7QKhiZg1dW3I6sDrY4ug6MQU%3D) # Build End2End Engine from ONNX ### 1. Build Engine by TensorRT ONNX Python api You can export TensorRT engine from ONNX by [`build.py` ](build.py). Usage: ``` shell python3 build.py \ --weights yolov8s.onnx \ --iou-thres 0.65 \ --conf-thres 0.25 \ --topk 100 \ --fp16 \ --device cuda:0 ``` #### Description of all arguments - `--weights` : The ONNX model you download. - `--iou-thres` : IOU threshold for NMS plugin. - `--conf-thres` : Confidence threshold for NMS plugin. - `--topk` : Max number of detection bboxes. - `--fp16` : Whether to export half-precision engine. - `--device` : The CUDA deivce you export engine . You can modify `iou-thres` `conf-thres` `topk` by yourself. ### 2. Export Engine by Trtexec Tools You can export TensorRT engine by [`trtexec`](https://github.com/NVIDIA/TensorRT/tree/main/samples/trtexec) tools. Usage: ``` shell /usr/src/tensorrt/bin/trtexec \ --onnx=yolov8s.onnx \ --saveEngine=yolov8s.engine \ --fp16 ``` **If you installed TensorRT by a debian package, then the installation path of `trtexec` is `/usr/src/tensorrt/bin/trtexec`** **If you installed TensorRT by a tar package, then the installation path of `trtexec` is under the `bin` folder in the path you decompressed** # Build TensorRT Engine by TensorRT API Please see more information in [`API-Build.md`](docs/API-Build.md) ***Notice !!!*** We don't support YOLOv8-seg model now !!! # Inference ## 1. Infer with python script You can infer images with the engine by [`infer-det.py`](infer-det.py) . Usage: ``` shell python3 infer-det.py \ --engine yolov8s.engine \ --imgs data \ --show \ --out-dir outputs \ --device cuda:0 ``` #### Description of all arguments - `--engine` : The Engine you export. - `--imgs` : The images path you want to detect. - `--show` : Whether to show detection results. - `--out-dir` : Where to save detection results images. It will not work when use `--show` flag. - `--device` : The CUDA deivce you use. - `--profile` : Profile the TensorRT engine. ## 2. Infer with C++ You can infer with c++ in [`csrc/detect/end2end`](csrc/detect/end2end) . ### Build: Please set you own librarys in [`CMakeLists.txt`](csrc/detect/end2end/CMakeLists.txt) and modify `CLASS_NAMES` and `COLORS` in [`main.cpp`](csrc/detect/end2end/main.cpp). ``` shell export root=${PWD} cd csrc/detect/end2end mkdir -p build && cd build cmake .. make mv yolov8 ${root} cd ${root} ``` Usage: ``` shell # infer image ./yolov8 yolov8s.engine data/bus.jpg # infer images ./yolov8 yolov8s.engine data # infer video ./yolov8 yolov8s.engine data/test.mp4 # the video path ``` # TensorRT Segment Deploy Please see more information in [`Segment.md`](docs/Segment.md) # TensorRT Pose Deploy Please see more information in [`Pose.md`](docs/Pose.md) # DeepStream Detection Deploy See more in [`README.md`](csrc/deepstream/README.md) # Jetson Deploy Only test on `Jetson-NX 4GB`. See more in [`Jetson.md`](docs/Jetson.md) # Profile you engine If you want to profile the TensorRT engine: Usage: ``` shell python3 trt-profile.py --engine yolov8s.engine --device cuda:0 ``` # Refuse To Use PyTorch for Model Inference !!! If you need to break away from pytorch and use tensorrt inference, you can get more information in [`infer-det-without-torch.py`](infer-det-without-torch.py), the usage is the same as the pytorch version, but its performance is much worse. You can use `cuda-python` or `pycuda` for inference. Please install by such command: ```shell pip install cuda-python # or pip install pycuda ``` Usage: ``` shell python3 infer-det-without-torch.py \ --engine yolov8s.engine \ --imgs data \ --show \ --out-dir outputs \ --method cudart ``` #### Description of all arguments - `--engine` : The Engine you export. - `--imgs` : The images path you want to detect. - `--show` : Whether to show detection results. - `--out-dir` : Where to save detection results images. It will not work when use `--show` flag. - `--method` : Choose `cudart` or `pycuda`, default is `cudart`.