Instructions to use HanzoHuang/YOLOv10-RKNN2-Ultralytics with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- YOLOv10
How to use HanzoHuang/YOLOv10-RKNN2-Ultralytics with YOLOv10:
from huggingface_hub import hf_hub_download from ultralytics import YOLOv10 # pick the weights file from this repo's "Files and versions" tab weights = hf_hub_download("HanzoHuang/YOLOv10-RKNN2-Ultralytics", "<weights>.pt") model = YOLOv10(weights) source = 'http://images.cocodataset.org/val2017/000000039769.jpg' model.predict(source=source, save=True) - Notebooks
- Google Colab
- Kaggle
Update RKNN2 model card and supported chips
Browse files
README.md
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: agpl-3.0
|
| 3 |
+
pipeline_tag: object-detection
|
| 4 |
+
tags:
|
| 5 |
+
- rknn
|
| 6 |
+
- rknn2
|
| 7 |
+
- rockchip
|
| 8 |
+
- yolov10
|
| 9 |
+
- rk3576
|
| 10 |
+
- rk3588
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# YOLOv10-RKNN2-Ultralytics
|
| 14 |
+
|
| 15 |
+
Ultralytics export for RK3576 and RK3588. YOLOv10 n, s, m, b, l, and x in FP16 and INT8 for RK3576 and RK3588. Choose a model under `models/detect/<checkpoint>/<chip>/<precision>/` and download both `model.rknn` and `metadata.yaml`.
|
| 16 |
+
|
| 17 |
+
These files were converted from the official Ultralytics assets. The Ultralytics exporter and RKNN Toolkit2 changed the model format and, for INT8, quantized the graph. The RKNN exports and original checkpoints are distributed under AGPL-3.0. See [LICENSE](LICENSE) and [checkpoint origins](source/weights.json).
|
| 18 |
+
|
| 19 |
+
## Reproduce the conversion
|
| 20 |
+
|
| 21 |
+
The exact input checkpoints, conversion scripts, package versions, and calibration hashes are in [source](tree/main/source). The Ultralytics 8.4.184 source distribution is available on [PyPI](https://pypi.org/project/ultralytics/8.4.184/#files). The converter uses [RKNN Toolkit2 2.3.2](https://github.com/airockchip/rknn-toolkit2/tree/v2.3.2), available separately from Rockchip.
|
| 22 |
+
|
| 23 |
+
Run `python source/convert-model.py source/weights/yolov10n.pt --chip rk3588 --precision fp16 --imgsz 640 --output yolov10n-rk3588`. For INT8, use `--precision int8 --data /path/to/coco128.yaml --calibration-limit 128`. The calibration manifest links to the source images. Point the YAML at the extracted COCO128 tree and its train2017 image directory.
|
| 24 |
+
|
| 25 |
+
The [model manifest](manifest.json) records file hashes, precision, and input size.
|