Ultralytics 集成
Welcome to the Ultralytics Integrations page! This page provides an overview of our partnerships with various tools and platforms, designed to streamline your machine learning workflows, enhance dataset management, simplify model training, and facilitate efficient deployment.
观看： Ultralytics YOLO11 Deployment and Integrations
数据集集成
- Roboflow:促进Ultralytics 模型的无缝数据集管理，提供强大的注释、预处理和增强功能。
培训整合
Amazon SageMaker：利用 Amazon SageMaker 高效构建、训练和部署Ultralytics 模型，为 ML 生命周期提供一体化平台。
ClearML:自动化Ultralytics ML 工作流、监控实验并促进团队协作。
Comet ML：通过Ultralytics 跟踪、比较和优化机器学习实验，加强模型开发。
DVC：为Ultralytics 机器学习项目实施版本控制，有效同步数据、代码和模型。
Google Colab：使用Google Colab 在支持协作和共享的云环境中训练和评估Ultralytics 模型。
IBM Watsonx：了解 IBM Watsonx 如何利用其尖端的人工智能工具、轻松的集成和先进的模型管理系统简化Ultralytics 模型的训练和评估。
JupyterLab：了解如何使用 JupyterLab 的交互式可定制环境，轻松高效地训练和评估Ultralytics 模型。
Kaggle：探索如何使用 Kaggle 在基于云的环境中训练和评估Ultralytics 模型，该环境提供预装库、GPU 支持以及用于协作和共享的活跃社区。
MLFlow：简化Ultralytics 模型的整个 ML 生命周期，从实验、可重现性到部署。
Neptune:在这个专为 MLOps 设计的元数据存储中，通过Ultralytics 维护 ML 实验的全面日志。
Paperspace Gradient: Paperspace Gradient simplifies working on YOLO11 projects by providing easy-to-use cloud tools for training, testing, and deploying your models quickly.
Ray Tune：以任何比例优化Ultralytics 模型的超参数。
TensorBoard：可视化Ultralytics ML 工作流、监控模型指标并促进团队协作。
Ultralytics HUB：访问预先训练的Ultralytics 模型社区并为其做出贡献。
Weights & Biases (W&B）：监控实验、可视化指标、促进Ultralytics 项目的可重复性和协作。
VS Code: An extension for VS Code that provides code snippets for accelerating development workflows with Ultralytics and also for anyone looking for examples to help learn or get started with Ultralytics.
Albumentations: Enhance your Ultralytics models with powerful image augmentations to improve model robustness and generalization.
SONY IMX500: Optimize and deploy Ultralytics YOLOv8 models on Raspberry Pi AI Cameras with the IMX500 sensor for fast, low-power performance.
部署集成
CoreML: CoreML, developed by Apple, is a framework designed for efficiently integrating machine learning models into applications across iOS, macOS, watchOS, and tvOS, using Apple's hardware for effective and secure model deployment.
Gradio🚀 新功能：使用 Gradio 部署Ultralytics 模型，进行实时、交互式对象检测演示。
NCNN: Developed by Tencent, NCNN is an efficient neural network inference framework tailored for mobile devices. It enables direct deployment of AI models into apps, optimizing performance across various mobile platforms.
MNN: Developed by Alibaba, MNN is a highly efficient and lightweight deep learning framework. It supports inference and training of deep learning models and has industry-leading performance for inference and training on-device.
Neural Magic:利用量化感知训练（QAT）和剪枝技术优化Ultralytics 模型，使其性能更优、体积更小。
ONNX:一种开源格式，由 Microsoft创建的一种开源格式，用于促进人工智能模型在各种框架之间的转移，提高Ultralytics 模型的通用性和部署灵活性。
OpenVINO: Intel's toolkit for optimizing and deploying computer vision models efficiently across various Intel CPU and GPU platforms.
PaddlePaddle:PaddlePaddle 是百度的一个开源深度学习平台，可实现人工智能模型的高效部署，并专注于工业应用的可扩展性。
TF GraphDef:开发者 Google,GraphDef 是TensorFlow 用来表示计算图的格式，可以优化机器学习模型在不同硬件上的执行。
TF SavedModel: Developed by Google, TF SavedModel is a universal serialization format for TensorFlow models, enabling easy sharing and deployment across a wide range of platforms, from servers to edge devices.
TF.js：由 GoogleTF.js 允许基于 JavaScript 部署 ML 模型，以促进浏览器和 Node.js 中的机器学习。
TFLite：由 GoogleTFLite 是一个轻量级框架，用于在移动和边缘设备上部署机器学习模型，确保以最小的内存占用进行快速、高效的推理。
TFLite Edge TPU: Developed by Google for optimizing TensorFlow Lite models on Edge TPUs, this model format ensures high-speed, efficient edge computing.
TensorRT: Developed by NVIDIA, this high-performance deep learning inference framework and model format optimizes AI models for accelerated speed and efficiency on NVIDIA GPUs, ensuring streamlined deployment.
TorchScript:作为 PyTorchTorchScript 可在各种生产环境中高效执行和部署机器学习模型，而无需依赖Python 。
导出格式
我们还支持多种模型导出格式，以便在不同环境中部署。以下是可用的格式：
| 格式 | format 论据 | 模型 | 元数据 | 论据 |
|---|---|---|---|---|
| PyTorch | - | yolo11n.pt | ✅ | - |
| TorchScript | torchscript | yolo11n.torchscript | ✅ | imgsz, optimize, batch |
| ONNX | onnx | yolo11n.onnx | ✅ | imgsz, half, dynamic, simplify, opset, batch |
| OpenVINO | openvino | yolo11n_openvino_model/ | ✅ | imgsz, half, int8, batch |
| TensorRT | engine | yolo11n.engine | ✅ | imgsz, half, dynamic, simplify, workspace, int8, batch |
| CoreML | coreml | yolo11n.mlpackage | ✅ | imgsz, half, int8, nms, batch |
| TF SavedModel | saved_model | yolo11n_saved_model/ | ✅ | imgsz, keras, int8, batch |
| TF GraphDef | pb | yolo11n.pb | ❌ | imgsz, batch |
| TF 轻型 | tflite | yolo11n.tflite | ✅ | imgsz, half, int8, batch |
| TF 边缘TPU | edgetpu | yolo11n_edgetpu.tflite | ✅ | imgsz |
| TF.js | tfjs | yolo11n_web_model/ | ✅ | imgsz, half, int8, batch |
| PaddlePaddle | paddle | yolo11n_paddle_model/ | ✅ | imgsz, batch |
| MNN | mnn | yolo11n.mnn | ✅ | imgsz, batch, int8, half |
| NCNN | ncnn | yolo11n_ncnn_model/ | ✅ | imgsz, half, batch |
| IMX500 | imx | yolo11n_imx_model/ | ✅ | imgsz, int8 |
浏览链接，了解有关每个集成的更多信息，以及如何通过Ultralytics 充分利用这些集成。查看全文 export 中的详细信息 出口 page.
为我们的集成做出贡献
我们总是很高兴看到社区如何将Ultralytics YOLO 与其他技术、工具和平台集成！如果您已成功地将YOLO 与新系统集成，或有宝贵的见解可供分享，请考虑向我们的集成文档投稿。
通过编写指南或教程，您可以帮助扩展我们的文档，并提供有益于社区的实际示例。这是为围绕Ultralytics YOLO 不断发展的生态系统做出贡献的绝佳方式。
要做出贡献，请查看我们的贡献指南，了解如何提交 Pull Request (PR) 🛠️。我们热切期待您的贡献！
让我们携手合作，使Ultralytics YOLO 生态系统更加广阔、功能更加丰富🙏！
常见问题
Ultralytics HUB 是什么？它如何简化 ML 工作流程？
Ultralytics HUB is a cloud-based platform designed to make machine learning (ML) workflows for Ultralytics models seamless and efficient. By using this tool, you can easily upload datasets, train models, perform real-time tracking, and deploy YOLO11 models without needing extensive coding skills. You can explore the key features on the Ultralytics HUB page and get started quickly with our Quickstart guide.
如何将Ultralytics YOLO 模型与Roboflow 进行数据集管理？
将Ultralytics YOLO 模型与Roboflow 整合，可为标注、预处理和扩增提供强大的工具，从而加强数据集管理。要开始使用，请按照 Roboflow集成页面上的步骤开始操作。这种合作关系可确保高效的数据集处理，这对开发准确、强大的YOLO 模型至关重要。
能否使用 MLFlow 跟踪Ultralytics 模型的性能？
是的，您可以。将 MLFlow 与Ultralytics 模型集成后，您就可以跟踪实验、提高可重复性并简化整个 ML 生命周期。有关设置此集成的详细说明，请参阅MLFlow集成页面。该集成对于监控模型指标和高效管理 ML 工作流特别有用。
What are the benefits of using Neural Magic for YOLO11 model optimization?
Neural Magic optimizes YOLO11 models by leveraging techniques like Quantization Aware Training (QAT) and pruning, resulting in highly efficient, smaller models that perform better on resource-limited hardware. Check out the Neural Magic integration page to learn how to implement these optimizations for superior performance and leaner models. This is especially beneficial for deployment on edge devices.
如何使用 Gradio 部署Ultralytics YOLO 模型进行交互式演示？
To deploy Ultralytics YOLO models with Gradio for interactive object detection demos, you can follow the steps outlined on the Gradio integration page. Gradio allows you to create easy-to-use web interfaces for real-time model inference, making it an excellent tool for showcasing your YOLO model's capabilities in a user-friendly format suitable for both developers and end-users.
通过解决这些常见问题，我们旨在改善用户体验，为Ultralytics 产品的强大功能提供有价值的见解。纳入这些常见问题不仅可以增强文档的功能，还能为Ultralytics 网站带来更多的有机流量。
