IntptrMax.YoloSharp 1.1.19

YoloSharp

Train Yolo model in C# with TorchSharp. <br/> With the help of this project you won't have to transform .pt model to onnx, and can train your own model in C# and don't have to install python.

Feature

• Written in C# only, don't have to install python.
• Train and predict your own model.
• Support Yolov5, Yolov5u, Yolov8, Yolov11 and Yolov12 now.
• Support Predict, Segment, Obb, Pose(Key Points) and Classification now.
• Support n/s/m/l/x size.
• Support LetterBox and Mosaic4 method for preprocessing images.
• Support NMS with GPU.
• Support Load PreTrained models from ultralytics yolov5/yolov8/yolo11 and yolov12(converted).
• Support .Net6 or higher.

🔥Important News

2026/03/06
🚀 Add config for training and predict.
🚀 Add more metrics for val.

2026/02/03
🚀 Add Early Stop.
🚀 Add HSV transform.
🚀 Add Train Logs.

2026/01/20
🚀 YoloSharp support Mixed Precision Trainer (simple amp)
🚀 Tqdm supported.
🚀 Add BF16 Precision.

Models

You can download yolo pre-trained models here.
<details> <summary>Prediction Checkpoints</summary>

</details>

<details> <summary>Segmention Checkpoints</summary>

</details>

How to use

You can download the code or add it from nuget.

dotnet add package IntptrMax.YoloSharp

You can use it with the code below:

Yolo Task

string preTrainedModelPath = @"..\..\..\Assets\PreTrainedModels\yolov8n-obb.bin"; // Pretrained model path.
string predictImagePath = @"..\..\..\Assets\TestImage\trucks.jpg";

Mat predictImage = Cv2.ImRead(predictImagePath);

// Create a Yolo config
Config config = new Config
{
	DeviceType = DeviceType.CUDA,
	ScalarType = ScalarType.Float16,
	RootPath = @"..\..\..\Assets\DataSets\dotav1",
	TrainDataPath = "train.txt",
	ValDataPath = "val.txt",
	YoloType = YoloType.Yolov8,
	YoloSize = YoloSize.n,
	TaskType = TaskType.Obb,
	ImageProcessType = ImageProcessType.Mosiac,
	ImageSize = 640,
	BatchSize = 16,
	NumberClass = 15,
	PredictThreshold = 0.3f,
	IouThreshold = 0.7f,
	Workers = 4,
	Epochs = 100,
};

// Create a yolo task.
YoloTask yoloTask = new YoloTask(config);

// Load pre-trained model. If you don't want to use pre-trained model, skip the step.
yoloTask.LoadModel(preTrainedModelPath, skipNcNotEqualLayers: true);

// Train model
yoloTask.Train();

// Predict image, if the model is not trained or loaded, it will use random weight to predict.
List<YoloResult> predictResult = yoloTask.ImagePredict(predictImage);

<br/>

Use yolov8n pre-trained model to detect.

Use yolov8n-seg pre-trained model to detect.

Use yolov8n-obb pre-trained model to detect.

Use yolov8n-pose pre-trained model to detect.