LTW.YoloDotNet.DirectML.Execution.Provider
1.0.1-aotfix.1
dotnet add package LTW.YoloDotNet.DirectML.Execution.Provider --version 1.0.1-aotfix.1
NuGet\Install-Package LTW.YoloDotNet.DirectML.Execution.Provider -Version 1.0.1-aotfix.1
<PackageReference Include="LTW.YoloDotNet.DirectML.Execution.Provider" Version="1.0.1-aotfix.1" />
<PackageVersion Include="LTW.YoloDotNet.DirectML.Execution.Provider" Version="1.0.1-aotfix.1" />
<PackageReference Include="LTW.YoloDotNet.DirectML.Execution.Provider" />
paket add LTW.YoloDotNet.DirectML.Execution.Provider --version 1.0.1-aotfix.1
#r "nuget: LTW.YoloDotNet.DirectML.Execution.Provider, 1.0.1-aotfix.1"
#:package LTW.YoloDotNet.DirectML.Execution.Provider@1.0.1-aotfix.1
#addin nuget:?package=LTW.YoloDotNet.DirectML.Execution.Provider&version=1.0.1-aotfix.1&prerelease
#tool nuget:?package=LTW.YoloDotNet.DirectML.Execution.Provider&version=1.0.1-aotfix.1&prerelease
信息
YoloDotNet 使用模块化执行器在不同硬件后端上运行推理。 每个执行器都针对特定平台或加速器,并且可能需要额外的系统级依赖, 例如运行时、驱动程序或操作系统支持。
仅安装 NuGet 包并不总是足够——正确配置取决于所选执行器与目标系统。
本文档介绍 DirectML 执行器 的安装与使用。
核心库要求
所有执行器都依赖核心 YoloDotNet 包。
NuGet 包
dotnet add package YoloDotNet
执行器 -- DirectML
DirectML 执行器通过 ONNX Runtime 使用 Microsoft DirectML, 在 Windows 上启用 GPU 加速推理。
DirectML 基于 DirectX 12 提供与硬件厂商无关的 GPU 后端, 可在多种 GPU 上运行推理,而无需厂商专属 SDK。
该执行器尤其适合:
- Windows 桌面应用
- 没有 CUDA 或 OpenVINO 的系统
- 希望使用统一 GPU API 的环境
⚠️ 注意
此执行器依赖 Windows 通过 DirectX 12 提供的 DirectML 支持。
除受支持的 Windows 版本和 GPU 驱动外,无需额外 SDK。
环境要求
- Windows 10 或更高版本(x64)
- 支持 DirectX 12 的 GPU(NVIDIA、AMD 或 Intel)
- 最新 GPU 驱动
- 与 DirectML 执行兼容的 ONNX 模型
安装
NuGet 包
dotnet add package YoloDotNet.ExecutionProvider.DirectML
YoloDotNet.ExecutionProvider.DirectML v1.0 需要 YoloDotNet 4.1 版本。
无需单独安装 DirectML 或 DirectX SDK。DirectML 已包含在 Windows 与 ONNX Runtime 中。
使用示例
using YoloDotNet;
using YoloDotNet.ExecutionProvider.DirectML;
using var yolo = new Yolo(new YoloOptions
{
ExecutionProvider = new DirectMLExecutionProvider(
model: "path/to/model.onnx",
gpuId: 0
),
// ...其他选项
});
性能与兼容性说明
- DirectML 性能高度依赖 GPU 驱动质量和模型结构。
- 在 NVIDIA 硬件上,如果追求最高吞吐,CUDA 执行器可能有更高性能。
DirectML 相比厂商专属优化,更强调 广泛兼容性与部署简便性。
故障排查
如果推理初始化失败:
- 确认 GPU 支持 DirectX 12。
- 确保 GPU 驱动为最新版本。
- 确认仅引用了 一个执行器 包。
- 验证 ONNX 模型受 ONNX Runtime DirectML 支持。
| Product | Versions Compatible and additional computed target framework versions. |
|---|---|
| .NET | net8.0 is compatible. net8.0-android was computed. net8.0-browser was computed. net8.0-ios was computed. net8.0-maccatalyst was computed. net8.0-macos was computed. net8.0-tvos was computed. net8.0-windows was computed. net9.0 was computed. net9.0-android was computed. net9.0-browser was computed. net9.0-ios was computed. net9.0-maccatalyst was computed. net9.0-macos was computed. net9.0-tvos was computed. net9.0-windows was computed. net10.0 is compatible. net10.0-android was computed. net10.0-browser was computed. net10.0-ios was computed. net10.0-maccatalyst was computed. net10.0-macos was computed. net10.0-tvos was computed. net10.0-windows was computed. |
-
net10.0
- LTW.YoloDotNet.Core (>= 4.2.1-AotFix.1)
- Microsoft.ML.OnnxRuntime.DirectML (>= 1.23.0)
-
net8.0
- LTW.YoloDotNet.Core (>= 4.2.1-AotFix.1)
- Microsoft.ML.OnnxRuntime.DirectML (>= 1.23.0)
NuGet packages
This package is not used by any NuGet packages.
GitHub repositories
This package is not used by any popular GitHub repositories.
| Version | Downloads | Last Updated |
|---|---|---|
| 1.0.1-aotfix.1 | 41 | 5/6/2026 |