huacd / rknn_model_zoo

简体中文 | English

RKNN Model Zoo

Description

RKNN Model Zoo is developed based on the RKNPU SDK toolchain and provides deployment examples for current mainstream algorithms. Include the process of exporting the RKNN model and using Python API and CAPI to infer the RKNN model.

• Support RK3562, RK3566, RK3568, RK3588 , RK3576 platforms.
• Limited support RV1103, RV1106
• Support RK1808, RV1109, RV1126 platforms.

Dependency library installation

RKNN Model Zoo relies on RKNN-Toolkit2 for model conversion. The Android compilation tool chain is required when compiling the Android demo, and the Linux compilation tool chain is required when compiling the Linux demo. For the installation of these dependencies, please refer to the Quick Start documentation at https://github.com/airockchip/rknn-toolkit2/tree/master/doc.

• Please note that the Android compilation tool chain recommends using version r18 or r19. Using other versions may encounter the problem of Cdemo compilation failure.

Model support

In addition to exporting the model from the corresponding respository, the models file are available on https://console.zbox.filez.com/l/8ufwtG (key: rknn).

Category	Name	Dtype	Model Download Link	Support platform
Classification	mobilenet	FP16/INT8	mobilenetv2-12.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RV1103|RV1106 RK1808|RK3399PRO RV1109|RV1126
Classification	resnet	FP16/INT8	resnet50-v2-7.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Object Detection	yolov5	FP16/INT8	./yolov5s_relu.onnx ./yolov5n.onnx ./yolov5s.onnx ./yolov5m.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RV1103|RV1106 RK1808|RK3399PRO RV1109|RV1126
Object Detection	yolov6	FP16/INT8	./yolov6n.onnx ./yolov6s.onnx ./yolov6m.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Object Detection	yolov7	FP16/INT8	./yolov7-tiny.onnx ./yolov7.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Object Detection	yolov8	FP16/INT8	./yolov8n.onnx ./yolov8s.onnx ./yolov8m.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Object Detection	yolov8_obb	INT8	./yolov8n-obb.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Object Detection	yolov10	FP16/INT8	./yolov10n.onnx ./yolov10s.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RV1103|RV1106 RK1808|RK3399PRO RV1109|RV1126
Object Detection	yolox	FP16/INT8	./yolox_s.onnx ./yolox_m.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Object Detection	ppyoloe	FP16/INT8	./ppyoloe_s.onnx ./ppyoloe_m.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Object Detection	yolo_world	FP16/INT8	./yolo_world_v2s.onnx ./clip_text.onnx	RK3566|RK3568|RK3588|RK3562|RK3576
Body Pose	yolov8_pose	INT8	./yolov8n-pose.onnx	RK3566|RK3568|RK3588|RK3562|RK3576
Image Segmentation	deeplabv3	FP16/INT8	./deeplab-v3-plus-mobilenet-v2.pb	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Image Segmentation	yolov5_seg	FP16/INT8	./yolov5n-seg.onnx ./yolov5s-seg.onnx ./yolov5m-seg.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Image Segmentation	yolov8_seg	FP16/INT8	./yolov8n-seg.onnx ./yolov8s-seg.onnx ./yolov8m-seg.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Image Segmentation	ppseg	FP16/INT8	pp_liteseg_cityscapes.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Face Key Points	RetinaFace	INT8	RetinaFace_mobile320.onnx RetinaFace_resnet50_320.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Car Plate Recognition	LPRNet	FP16/INT8	./lprnet.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Text Detection	PPOCR-Det	FP16/INT8	../ppocrv4_det.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Text Recognition	PPOCR-Rec	FP16	../ppocrv4_rec.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Neural Machine Translation	lite_transformer	FP16	lite-transformer-encoder-16.onnx lite-transformer-decoder-16.onnx	RK3566|RK3568|RK3588|RK3562|RK3576 RK1808|RK3399PRO RV1109|RV1126
Image-Text Matching	clip	FP16	./clip_images.onnx ./clip_text.onnx	RK3566|RK3568|RK3588|RK3562|RK3576
Speech Recognition	whisper	FP16	whisper_encoder_base_20s.onnx whisper_decoder_base_20s.onnx	RK3566|RK3568|RK3588|RK3562|RK3576
Speech Classification	yamnet	FP16	yamnet_3s.onnx	RK3566|RK3568|RK3588|RK3562|RK3576

Model performance benchmark(FPS)

demo	model_name	inputs_shape	dtype	RK3566 RK3568	RK3562	RK3588 @single_core	RK3576 @single_core	RV1109	RV1126	RK1808
mobilenet	mobilenetv2-12	[1, 3, 224, 224]	INT8	188.4	276.8	453.0	475.1	212.9	322.3	170.3
resnet	resnet50-v2-7	[1, 3, 224, 224]	INT8	38.5	52.5	109.0	97.5	24.4	36.2	37.1
yolov5	yolov5s_relu	[1, 3, 640, 640]	INT8	26.1	32.5	64.3	65.5	20.2	29.2	37.2
	yolov5n	[1, 3, 640, 640]	INT8	39.9	47.1	76.5	112.7	36.3	53.2	61.2
	yolov5s	[1, 3, 640, 640]	INT8	19.5	23.1	45.9	56.5	13.6	20.0	28.2
	yolov5m	[1, 3, 640, 640]	INT8	8.8	10.5	20.3	23.9	5.8	8.5	13.3
yolov6	yolov6n	[1, 3, 640, 640]	INT8	48.3	55.3	102.0	109.5	37.8	56.8	66.8
	yolov6s	[1, 3, 640, 640]	INT8	15.2	16.7	35.3	34.0	10.8	16.3	24.1
	yolov6m	[1, 3, 640, 640]	INT8	7.3	8.2	17.5	17.3	5.6	8.3	11.5
yolov7	yolov7-tiny	[1, 3, 640, 640]	INT8	28.4	36.0	71.1	75.3	15.4	22.4	37.2
	yolov7	[1, 3, 640, 640]	INT8	4.7	5.6	11.2	12.7	3.3	4.8	7.4
yolov8	yolov8n	[1, 3, 640, 640]	INT8	34.5	39.7	67.0	89.5	24.0	35.4	42.3
	yolov8s	[1, 3, 640, 640]	INT8	15.4	17.6	36.0	40.7	8.9	13.1	19.1
	yolov8m	[1, 3, 640, 640]	INT8	6.7	8.0	15.7	16.5	3.9	5.8	9.1
yolov8_obb	yolov8n-obb	[1, 3, 640, 640]	INT8	34.6	40.4	67.2	91.4	25.1	37.3	42.8
yolov10	yolov10n	[1, 3, 640, 640]	INT8	20.0	32.9	55.7	80.6	/	/	/
	yolov10s	[1, 3, 640, 640]	INT8	9.9	16.2	32.2	39.6	/	/	/
yolox	yolox_s	[1, 3, 640, 640]	INT8	15.1	17.8	35.4	42.2	10.6	15.7	23.0
	yolox_m	[1, 3, 640, 640]	INT8	6.7	7.9	15.5	17.5	4.6	6.8	10.7
ppyoloe	ppyoloe_s	[1, 3, 640, 640]	INT8	7.0	17.8	21.8	43.6	11.2	16.4	21.1
	ppyoloe_m	[1, 3, 640, 640]	INT8	3.7	8.3	10.3	18.1	5.2	7.7	9.4
yolo_world	yolo_world_v2s	[1, 3, 640, 640]	INT8	5.8	10.8	16.7	23.4	/	/	/
	clip_text	[1, 20]	FP16	21.0	50.2	68.9	54.7	/	/	/
yolov8_pose	yolov8n-pose	[1, 3, 640, 640]	INT8	20.9	29.5	48.9	68.0	/	/	/
deeplabv3	deeplab-v3-plus-mobilenet-v2	[1, 513, 513, 1]	INT8	12.0	19.9	34.4	38.9	10.1	13.0	4.4
yolov5_seg	yolov5n-seg	[1, 3, 640, 640]	INT8	31.7	38.0	63.5	88.7	28.6	42.2	49.6
	yolov5s-seg	[1, 3, 640, 640]	INT8	14.9	17.6	35.1	41.5	9.6	14.0	22.5
	yolov5m-seg	[1, 3, 640, 640]	INT8	6.8	8.1	15.9	18.1	4.7	6.8	10.8
yolov8_seg	yolov8n-seg	[1, 3, 640, 640]	INT8	27.8	32.1	55.1	71.2	18.6	27.6	32.9
	yolov8s-seg	[1, 3, 640, 640]	INT8	11.9	13.6	27.5	30.6	6.6	9.8	14.6
	yolov8m-seg	[1, 3, 640, 640]	INT8	5.2	6.1	12.2	12.7	3.1	4.6	6.9
ppseg	ppseg_lite_1024x512	[1, 3, 512, 512]	INT8	6.1	8.8	33.2	29.5	18.4	27.1	20.9
RetinaFace	RetinaFace_mobile320	[1, 3, 320, 320]	INT8	117.7	175.7	168.7	495.3	144.8	212.5	198.5
	RetinaFace_resnet50_320	[1, 3, 320, 320]	INT8	17.9	24.4	45.1	56.1	14.6	20.8	24.6
LPRNet	lprnet	[1, 3, 24, 94]	INT8	47.3	144.2	234.7	136.9	30.6	47.6	30.1
PPOCR-Det	ppocrv4_det	[1, 3, 480, 480]	INT8	22.8	27.1	46.2	64.5	11.0	16.1	14.2
PPOCR-Rec	ppocrv4_rec	[1, 3, 48, 320]	FP16	19.3	50.8	63.2	95.7	1.0	1.6	6.7
lite_transformer	lite-transformer-encoder-16	embedding-256, token-16	FP16	321.8	659.9	830.3	804.2	22.7	35.4	98.3
	lite-transformer-decoder-16	embedding-256, token-16	FP16	121.2	216.4	312.5	265.5	48.0	65.8	109.9
clip	clip_images	[1, 3, 224, 224]	FP16	1.9	3.0	6.1	6.3	/	/	/
	clip_text	[1, 20]	FP16	21.4	50.3	68.3	54.7	/	/	/
whisper	encoder+decoder+NPU-outside process	20s audio	FP16	RTF 1.263	RTF 0.392	RTF 0.215	RTF 0.222	/	/	/
yamnet	yamnet_3s	3s audio	FP16	RTF 0.014	RTF 0.008	RTF 0.003	RTF 0.005	/	/	/

• This performance data are collected based on the maximum NPU frequency of each platform.
• This performance data calculate the time-consuming of model inference. Does not include the time-consuming of pre-processing and post-processing if not specified.
• / means currently not support.

Compile Demo

For Linux develop board:

./build-linux.sh -t <target> -a <arch> -d <build_demo_name> [-b <build_type>] [-m]
    -t : target (rk356x/rk3588/rk3576/rv1106/rk1808/rv1126)
    -a : arch (aarch64/armhf)
    -d : demo name
    -b : build_type(Debug/Release)
    -m : enable address sanitizer, build_type need set to Debug
Note: 'rk356x' represents rk3562/rk3566/rk3568, 'rv1106' represents rv1103/rv1106, 'rv1126' represents rv1109/rv1126

# Here is an example for compiling yolov5 demo for 64-bit Linux RK3566.
./build-linux.sh -t rk356x -a aarch64 -d yolov5

For Android development board:

# For Android develop boards, it's require to set path for Android NDK compilation tool path according to the user environment
export ANDROID_NDK_PATH=~/opts/ndk/android-ndk-r18b
./build-android.sh -t <target> -a <arch> -d <build_demo_name> [-b <build_type>] [-m]
    -t : target (rk356x/rk3588/rk3576)
    -a : arch (arm64-v8a/armeabi-v7a)
    -d : demo name
    -b : build_type (Debug/Release)
    -m : enable address sanitizer, build_type need set to Debug

# Here is an example for compiling yolov5 demo for 64-bit Android RK3566.
./build-android.sh -t rk356x -a arm64-v8a -d yolov5

Release Notes

Version	Description
2.1.0	New demo release, including yolov8_pose, yolov8_obb, yolov10, yolo_world, clip, whisper, yamnet RK1808, RV1109, RV1126 platform support of these demo will be added in next version.
2.0.0	Add new support for RK3576 for all demo. Full support for RK1808, RV1109, RV1126 platform.
1.6.0	New demo release, including object detection, image segmentation, OCR, car plate detection&recognition etc. Full support for RK3566, RK3568, RK3588, RK3562 platforms. Limited support for RV1103, RV1106 platforms.
1.5.0	Yolo detection demo release.

Environment dependencies

All demos in RKNN Model Zoo are verified based on the latest RKNPU SDK. If using a lower version for verification, the inference performance and inference results may be wrong.

Version	RKNPU2 SDK	RKNPU1 SDK
2.1.0	>=2.1.0	>=1.7.5
2.0.0	>=2.0.0	>=1.7.5
1.6.0	>=1.6.0	-
1.5.0	>=1.5.0	>=1.7.3

RKNPU Resource

• RKNPU2 SDK: https://github.com/airockchip/rknn-toolkit2
• RKNPU1 SDK: https://github.com/airockchip/rknn-toolkit

License

Apache License 2.0