license: agpl-3.0
tags:
- rknn
- rknn2
- rockchip
- rk3576
- rk3588
YOLOv8-RKNN2
Looking for the Ultralytics RKNN exports? See YOLOv8-RKNN2-Ultralytics.
YOLOv8 RKNN exports for RK3576 and RK3588, compiled with RKNN Toolkit2 2.3.2. Available tasks: detection, classification, oriented bounding boxes, pose, and instance segmentation.
Choose a model
Models are organized by toolkit version, task, variant, target chip, and precision:
models/v<toolkit-version>/<task>/<variant>/<chip>/<precision>/
Download model.rknn and metadata.yaml from the same directory. The manifest lists the available models and checksums.
Conversion
The repository includes Rockchip-guided and custom exports. Detection includes n, s, and m variants in FP16 and INT8 for both chips. The checkpoints were exported with Rockchip's RKOPT detector head and compiled with the Model Zoo converter. Checkpoint URLs, exporter commits, scripts, and ONNX graphs are in source.
Source and license
Original checkpoints, conversion scripts, and provenance are in source. The models and checkpoints are distributed under AGPL-3.0; see LICENSE and checkpoint details.