YOLOv10 Nano | Luxonis

Our new model ZOO works with DepthAI V3. Find out more in our documentation.

1+ Likes

Model Details

Model Description

YOLOv10n model is a convolutional neural network designed for object detection. It introduces a new approach to real-time object detection, addressing both the post-processing and model architecture deficiencies found in previous YOLO versions. By eliminating non-maximum suppression (NMS) and optimizing various model components, YOLOv10 achieves state-of-the-art performance with significantly reduced computational overhead. Extensive examples demonstrate its superior accuracy-latency trade-offs across multiple model scales. It was trained on the COCO data set. It is capable of detecting objects of 80 classes.

Developed by: Tsinghua University
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 dataset with the input resolution of 640×640. Results are taken from .

Model	AP
	38.5

Technical Specifications

Input/Output Details

Input:
- Name: image
  - Info: NCHW BGR un-normalized image
Outputs:
- Name: multiple (see NN archive)
  - Info: Unprocessed outputs of a multitude of detections

Model Architecture

Backbone: an enhanced version of CSPNet (Cross Stage Partial Network)
Neck: includes PAN (Path Aggregation Network)
Head: One-to-One head

Please consult the for more information on model architecture.

Throughput

Model variant: yolov10-nano:coco-512x288

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

• Params (M): 2.299 • GFLOPs: 1.435

Platform	Precision	Throughput (infs/sec)	Power Consumption (W)
RVC2	FP16	28.73	N/A
RVC4	INT8	618.98	3.08

* 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/yolov10-nano: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/yolov10-nano:coco-512x288

General object detection model
License	GNU Affero General Public License v3.0 Commercial use
Downloads	174
Tasks	Object Detection
Model Types	ONNX

Model Variants

Name	Version	Available For	Created At	Deploy
		RVC2, RVC4	10 months ago