dotnet/machinelearning-samples
Machinelearning Samples
该仓库是用于在 .NET 生态系统中构建和集成机器学习模型的参考实现、模板和示例库。它提供了一系列使用 ML.NET 框架实现机器学习工作流的实践演示。
该项目强调通过 Open Neural Network Exchange (ONNX) 格式集成预训练模型，允许在托管应用程序中执行外部机器学习逻辑。它包含了加载和执行这些标准化模型的具体示例，以确保跨平台兼容性。
示例涵盖了多种监督学习任务，包括文本情感分类、用于目标检测的图像和视频分析，以及时间序列预测。它还提供了网络异常检测的实现，以及用于超参数优化和数据转换流水线的工具。
Features
- .NET Machine Learning Integrations - Integrates machine learning models and predictive capabilities specifically within the .NET ecosystem.
- Data Transformation Pipelines - Implements frameworks for filtering, cleaning, and modifying data before it is used in model generation.
- Image and Video Analysis - Implements machine learning workflows to classify image content and detect objects within real-time video streams.
- External Model Loading - Provides capabilities for importing and initializing pre-trained models from the standardized ONNX open format.
- ONNX Model Runtimes - Provides runtimes that load and execute ONNX models for cross-framework inference.
- Supervised Learning - Implements various classification and regression tasks using models trained on labeled data.
- Time Series Forecasting - Implements models and architectures designed for predicting future values in temporal data sequences.
- Real-Time Object Detection - Provides tools for identifying and locating specific objects within live video streams or webcam footage.
- Custom Data Transform Extensions - Provides mechanisms for defining and registering custom data preprocessing logic within machine learning pipelines.
- Feature Extraction - Provides tools for converting raw visual data into meaningful numerical representations for analysis.
- Image Classification - Implements systems that assign labels or categories to images based on their visual content.
- Hyperparameter Optimization - Implements automated methods for searching and selecting the best configuration parameters for a model.
- Dataset-Driven Model Generators - Provides capabilities to generate high-quality machine learning models and the source code required to execute them from datasets.
- Sentiment Classifiers - Provides implementation examples for building neural network architectures that classify the emotional tone of text data.
- Anomaly Detection - Provides reference implementations for identifying unusual patterns in network traffic and system logs.
- AI and Machine Learning - Interactive UI samples for sentiment analysis and prediction.
