YOLOv8-TensorRT/docs/Segment.md

YOLOv8-seg Model with TensorRT

The yolov8-seg model conversion route is : YOLOv8 PyTorch model -> ONNX -> TensorRT Engine

Notice !!! We don't support TensorRT API building !!!

Export Modified ONNX model

You can export your onnx model by ultralytics API and the onnx is also modify by this repo.

python3 export-seg.py \
--weights yolov8s-seg.pt \
--opset 11 \
--sim \
--input-shape 1 3 640 640 \
--device cuda:0

Description of all arguments

• --weights : The PyTorch model you trained.
• --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.

This onnx model doesn't contain postprocessing.

Export Engine by TensorRT Python api

You can export TensorRT engine from ONNX by build.py .

Usage:

python3 build.py \
--weights yolov8s-seg.onnx \
--fp16  \
--device cuda:0 \
--seg

Description of all arguments

• --weights : The ONNX model you download.
• --fp16 : Whether to export half-precision engine.
• --device : The CUDA deivce you export engine.
• --seg : Whether to export seg engine.

Export Engine by Trtexec Tools

You can export TensorRT engine by trtexec tools.

Usage:

/usr/src/tensorrt/bin/trtexec \
--onnx=yolov8s-seg.onnx \
--saveEngine=yolov8s-seg.engine \
--fp16

Inference

Infer with python script

You can infer images with the engine by infer-seg.py .

Usage:

python3 infer-seg.py \
--engine yolov8s-seg.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.

Infer with C++

You can infer segment engine with c++ in csrc/segment/simple .

Build:

Please set you own librarys in CMakeLists.txt and modify you own config in main.cpp such as CLASS_NAMES, COLORS, MASK_COLORS and postprocess parameters .

int topk = 100;
int seg_h = 160; // yolov8 model proto height
int seg_w = 160; // yolov8 model proto width
int seg_channels = 32; // yolov8 model proto channels
float score_thres = 0.25f;
float iou_thres = 0.65f;

export root=${PWD}
cd src/segment/simple
mkdir build
cmake ..
make
mv yolov8-seg ${root}
cd ${root}

Notice !!!

If you have build OpenCV(>=4.7.0) by yourself, it provides a new api cv::dnn::NMSBoxesBatched . It is a gread api about efficient in-class nms . It will be used by default!

!!!

Usage:

# infer image
./yolov8-seg yolov8s-seg.engine data/bus.jpg
# infer images
./yolov8-seg yolov8s-seg.engine data
# infer video
./yolov8-seg yolov8s-seg.engine data/test.mp4 # the video path

Export Orin ONNX model by ultralytics

You can leave this repo and use the original ultralytics repo for onnx export.

1. ONNX -> TensorRT

You can export your onnx model by ultralytics API.

yolo export model=yolov8s-seg.pt format=onnx opset=11 simplify=True

or run this python script:

from ultralytics import YOLO

# Load a model
model = YOLO("yolov8s-seg.pt")  # load a pretrained model (recommended for training)
success = model.export(format="onnx", opset=11, simplify=True)  # export the model to onnx format
assert success

Then build engine by Trtexec Tools.

You can export TensorRT engine by trtexec tools.

Usage:

/usr/src/tensorrt/bin/trtexec \
--onnx=yolov8s-seg.onnx \
--saveEngine=yolov8s-seg.engine \
--fp16

2. Direct to TensorRT (NOT RECOMMAND!!)

Usage:

yolo export model=yolov8s-seg.pt format=engine device=0

or run python script:

from ultralytics import YOLO

# Load a model
model = YOLO("yolov8s-seg.pt")  # load a pretrained model (recommended for training)
success = model.export(format="engine", device=0)  # export the model to engine format
assert success

After executing the above script, you will get an engine named yolov8s-seg.engine .

Inference with c++

You can infer with c++ in csrc/segment/normal .

Build:

Please set you own librarys in CMakeLists.txt and modify CLASS_NAMES and COLORS in main.cpp.

Besides, you can modify the postprocess parameters such as num_labels and score_thres and iou_thres and topk in main.cpp.

int topk = 100;
int seg_h = 160; // yolov8 model proto height
int seg_w = 160; // yolov8 model proto width
int seg_channels = 32; // yolov8 model proto channels
float score_thres = 0.25f;
float iou_thres = 0.65f;

And build:

export root=${PWD}
cd src/segment/normal
mkdir build
cmake ..
make
mv yolov8-seg ${root}
cd ${root}

Usage:

# infer image
./yolov8-seg yolov8s-seg.engine data/bus.jpg
# infer images
./yolov8-seg yolov8s-seg.engine data
# infer video
./yolov8-seg yolov8s-seg.engine data/test.mp4 # the video path

Refuse To Use PyTorch for segment Model Inference !!!

It is the same as detection model. you can get more information in infer-seg-without-torch.py,

Usage:

python3 infer-seg-without-torch.py \
--engine yolov8s-seg.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.