X

Model Details

Model Description

YOLOv6n model is a convolutional neural network designed for object detection. It is highly effective and accurate, making it ideal for real-time applications on edge devices. It was trained on the COCO data set. It is capable of detecting objects of 80 classes.

Developed by: Meituan Vision AI Department
Shared by:
Model type: Computer Vision
License:
Resources for more information:

Training Details

Training Data

is a large-scale object detection, segmentation, and captioning dataset.

Testing Details

Metrics

Results of the mAP and speed are evaluated on COCO val2017 dataset with the input resolution of 640×640. Results are taken from .

Model	mAP	Params ^(M)
	35.9	4.3

Technical Specifications

Input/Output Details

Input:
- Name: image
  - Info: NCHW BGR un-normalized image
Output:
- Name: output1_yolov6r2
  - Info: Unprocessed output of the first channel
- Name: output2_yolov6r2
  - Info: Unprocessed output of the second channel
- Name: output3_yolov6r2
  - Info: Unprocessed output of the third channel

Model Architecture

Backbone: EfficientRep backbone
Neck: Rep-PAN neck
Head: Efficient decoupled head that is anchor-free

Please consult the for more information on model architecture.

Throughput

Model variant: yolov6-nano:r2-coco-512x288

• Input shape: [1, 3, 288, 512] • Output shapes: [[1, 85, 36, 64], [1, 85, 18, 32], [1, 85, 9, 16]]

• Params (M): 4.317 • GFLOPs: 2.032

Platform	Precision	Throughput (infs/sec)	Power Consumption (W)
RVC2	FP16	63.61	N/A
RVC4	INT8	798.38	3.38

Model variant: yolov6-nano:coco-416x416

Platform	Precision	Throughput (infs/sec)	Power Consumption (W)
RVC2	FP16	67.41	N/A

* Benchmarked with , using 2 threads (and the DSP runtime in balanced mode for RVC4).

* Parameters and FLOPs are obtained from the package.

Quantization

RVC4 version of the model was quantized using a custom dataset. This was created by taking a full 128-image dataset.

Utilization

Models converted for RVC Platforms can be used for inference on OAK devices. DepthAI pipelines are used to define the information flow linking the device, inference model, and the output parser (as defined in model head(s)). Below, we present the most crucial utilization steps for the particular model. Please consult the docs for more information.

Install DAIv3 library:

pip install depthai

Define model:

model_description = dai.NNModelDescription(
    "luxonis/yolov6-nano:r2-coco-512x288"
)

nn = pipeline.create(dai.node.DetectionNetwork).build(
    <CameraNode>, model_description
)

The model is automatically parsed by DAI and it outputs the message (bounding boxes, labels, and scores of the detected objects).

Get model output(s):

while pipeline.isRuning():
    nn_output: dai.ImgDetections = parser_output_queue.get()

Example

You can quickly run the model using our script. It automatically downloads the model, creates a DepthAI pipeline, runs the inference, and displays the results using our DepthAI visualizer tool. To try it out, run:

python3 main.py \
    --model luxonis/yolov6-nano:r2-coco-512x288

You can also check the example with .

Notes

There is a special .onnx model for RVC3 because of the unsupported reducemax operation on RVC3.
The post-processing operations were pruned. For RVC3, the objectness tensor is set to ones because of the unsupported reducemax operation on RVC3.

General object Detection model
License	GNU General Public License v3.0 Commercial use
Downloads	132782
Tasks	Object Detection
Model Types	IR ONNX

Name	Version	Available For	Created At	Deploy
		RVC2	About 1 year ago
		RVC3	About 1 year ago
		RVC2, RVC4	About 1 year ago