数据检查于 2026-08-13. Data may be stale - beyond the review cycle.
审查周期记录在方法论页面。 方法论
rockchip-linux/rknn-toolkit2
软件工具 Tier A edge_ai BSD-3-Clause
Rockchip 官方 RKNN 模型转换与推理工具包（RK3588 等 NPU）。
概览
Rockchip RKNN 工具包，官方仓库：https://github.com/rockchip-linux/rknn-toolkit2。
工程快照
用途与任务 6
- 最适合
-
- Edge AI deployment on Rockchip RK3588/RK3566/RK3568/RK3576 based companion computers
- Model conversion from ONNX/TensorFlow/PyTorch to RKNN format
- Quantization and optimization for Rockchip NPU inference
- 主要任务
-
- Convert trained models to RKNN format for Rockchip NPU
- Quantize models (INT8/INT16) for efficient inference
- Deploy inference pipelines on RK3588/RK3566/RK3568/RK3576 boards
技术栈与生态
- 资源类型
- 软件工具
- 生态系统
- edge_ai
许可与法律
- 许可证
-
BSD-3-Clause(推断)
生命周期与时效性 展开
- 维护状态
- 活跃维护
- 最新版本
- 未记录
- 最近活动
- 2026-08-10
- 最近检查
- 2026-08-13
- 验证状态
- 官方确认
解决什么问题
Engineers need a Rockchip-supported toolchain to convert, quantize, and deploy neural network models onto RK3588/RK3566/RK3568/RK3576 NPUs for edge AI on companion computers.
主要用例
- Convert trained models to RKNN format for Rockchip NPU
- Quantize models (INT8/INT16) for efficient inference
- Deploy inference pipelines on RK3588/RK3566/RK3568/RK3576 boards
次要用例
- Performance profiling and optimization on Rockchip NPU
- Integration with custom C++/Python applications on Linux
何时使用
考虑使用
- Target hardware is Rockchip RK3588, RK3566, RK3568, or RK3576
- License compatibility with BSD-3-Clause is acceptable
- Model architectures are supported by RKNN-Toolkit2 (check release notes)
- Quantization accuracy trade-offs are evaluated for the specific model
采用前需验证
- Model conversion success for your specific architecture and opset
- Quantization accuracy drop on representative validation set
- Inference latency and throughput on target RK NPU
- Compatibility with host OS (Linux) and Python/C++ API versions
从这里开始
采用检查清单
- 需验证 Model conversion success for your specific architecture and opset
- 需验证 Quantization accuracy drop on representative validation set
- 需验证 Inference latency and throughput on target RK NPU
- 需验证 Compatibility with host OS (Linux) and Python/C++ API versions
以上检查项只有在官方来源确认后才能标记为“已验证”；无法确认的保持未验证。
已知限制与未知项
已知限制
- Only supports Rockchip RKNPU targets (RK3588, RK3566, RK3568, RK3576)
- Model operator coverage limited to RKNN supported ops
- Quantization may require calibration data and tuning
- Toolchain runs on Linux host (x86_64/aarch64); Windows/macOS not officially supported
未公开验证
- Exact model support matrix per RKNN version
- Performance benchmarks for specific models on each RK NPU
- Long-term maintenance roadmap and release cadence
替代与相关工具
相关工具
- onnx-runtime — 集成
如何使用？
从下方记录的入口开始，然后对照技术清单进行验证。
技术清单
- 通过 许可证已识别 已记录: BSD-3-Clause
- 通过 维护信号 活跃维护
- 通过 验证状态 官方确认
- 通过 已附加来源证据 1 个来源记录
- 需复核 已记录最新版本 未记录
官方链接
元数据与治理
| 许可证 | BSD-3-Clause — 推断 |
|---|---|
| 商业化模式 | 未知 |
| 维护状态 | 活跃维护 |
| 验证状态 | 官方确认 — 已通过官方仓库 API 响应确认，证据见下方来源引用。 |
| 最新版本 | 未记录 |
| 最新发布 | 未记录 |
| 最近活动 | 2026-08-10 |
| 最近检查 | 2026-08-13 |
| 首次发现 | 未记录 |
数据集事实
以上事实仅来自官方数据集卡，未确认字段保持未知。
相关资源与依赖
- onnx-runtime — integrates with (confirmed)
近期动态
新增资源：rockchip-linux/rknn-toolkit2
New repository resource added by the sprint promote pipeline.
相关知识
指南
- Deploying AI Inference on UAV Edge Compute — A repeatable workflow for choosing and validating an inference runtime for an onboard UAV computer.
合集
- 边缘AI感知入门套件 — 一套用于在边缘部署工作之前对空中感知模型进行原型验证的起步技术栈。