ML.NET-API（一）_常用代码片段

合集 - 机器视觉-Yolo(8)

1.详细Anaconda+Pycharm配置YoloV5环境笔记2021-09-032.Pytorch学习笔记-（二）搭建Pytorch环境（Pytorch2.1+CUDA12.1+Anaconda3_2023+Pycharm2023）2023-10-093.yolov5-OSError: [WinError 1455] 页面文件太小，无法完成操作。 Error loading "C:\Anaconda3\lib\site-packages\torch\lib\caffe2_detectron_ops_gpu.dll"2021-09-074.YoloV5_RuntimeError: CUDA out of memory. Tried to allocate 100.00 MiB (GPU 0; 2.00 GiB total capacity; 1.15 GiB already allocated; 0 bytes free; 1.19 GiB reserved in total by PyTorch)2021-09-075.C#调用YoloV5与YoloV8示例2024-05-116.C#实现控制台传参调用YoloV5模型进行人体识别2024-07-08

7.ML.NET-API（一）_常用代码片段2024-07-12

8.ML.NET-模型生成器工具（二）-对象检测教程2024-07-12

收起

1、训练图像

//Append ImageClassification trainer to your pipeline containing any preprocessing transforms
pipeline
    .Append(mlContext.MulticlassClassification.Trainers.ImageClassification(featureColumnName: "Image")
    .Append(mlContext.Transforms.Conversion.MapKeyToValue("PredictedLabel");

// 训练模型
var model = pipeline.Fit(trainingData);

// 预测
var predictedData = model.Transform(newData).GetColumn<string>("PredictedLabel");

2、训练文本

// Define training pipeline using TextClassification trainer
var pipeline =
    mlContext.Transforms.Conversion.MapValueToKey("Label","Sentiment")
        .Append(mlContext.MulticlassClassification.Trainers.TextClassification(sentence1ColumnName: "Text"))
        .Append(mlContext.Transforms.Conversion.MapKeyToValue("PredictedLabel"));

// 训练模型
var model = pipeline.Fit(trainingData);

// 预测
var predictedData = model.Transform(newData).GetColumn<string>("PredictedLabel");

3、句子相似性

// Define your pipeline
var pipeline = mlContext.Regression.Trainers.SentenceSimilarity(sentence1ColumnName: "Sentence", sentence2ColumnName: "Sentence2");

// 训练模型
var model = pipeline.Fit(trainingData);

// 预测
var score = model.Transform(newData).GetColumn<float>("Score");

4、使用预先训练的TensorFlow模型

// Load TensorFlow model
TensorFlowModel tensorFlowModel = mlContext.Model.LoadTensorFlowModel(_modelPath);

//Append ScoreTensorFlowModel transform to your pipeline containing any preprocessing transforms
pipeline.Append(tensorFlowModel.ScoreTensorFlowModel(outputColumnName: "Prediction/Softmax", inputColumnName:"Features"))

// 训练
ITransformer model = pipeline.Fit(dataView);

// 预测方式一
var predictions = pipeline.Fit(data).GetColumn<float[]>(TinyYoloModelSettings.ModelOutput);
// 预测方式二：
var predictions = model.Transform(dataView).GetColumn<float>("Prediction/Softmax");

5、使用预先训练的ONNX模型

// Append ApplyOnnxModel transform to pipeline containing any preprocessing transforms
pipeline.Append((modelFile: modelLocation, outputColumnNames: new[] { TinyYoloModelSettings.ModelOutput }, inputColumnNames: new[] { TinyYoloModelSettings.ModelInput })

// 训练
var model = pipeline.Fit(data);

// 预测方式一
var predictions = pipeline.Fit(data).GetColumn<float[]>(TinyYoloModelSettings.ModelOutput);
// 预测方式二：
IDataView scoredData = model.Transform(imageDataView);

6、使用‘模型生成器工具’的代码进行预测

MLModel1.Predict(sampleData);           // 预测
MLModel1.PredictAllLabels(sampleData);  // 预测标签