STMicroelectronics/ssdlite_mobilenetv2_pt

SSDLite MobileNetV2

Use case : Object detection

Model description

SSDLite MobileNetV2 is a lightweight single-shot object detection model optimized for real-time inference on mobile and edge devices.

It combines the SSDLite framework, a streamlined version of SSD for efficiency, with MobileNetV2 as the backbone. MobileNetV2 uses inverted residual blocks and linear bottlenecks, enabling a strong balance between accuracy and computational efficiency while keeping the model small and fast.
The SSDLite head predicts object locations and class probabilities in a single forward pass, making the model suitable for real-time detection in resource-constrained environments.

The ssdlite_mobilenetv2_pt variant is implemented in PyTorch and is widely used in applications requiring low latency, minimal memory footprint, and high energy efficiency, such as mobile vision apps and embedded systems.

Network information

Network information	Value
Framework	Torch
Quantization	Int8
Provenance	torchvision GitHub
Paper	SSDLite MobileNetV2

The model is quantized to int8 using ONNX Runtime and exported for efficient deployment.

Network inputs / outputs

For an image resolution of NxM and NC classes

Input Shape	Description
(1, W, H, 3)	Single NxM RGB image with UINT8 values between 0 and 255

Output Shape	Description
(1, 3000,(1+NC) and (1,3000,4))	Model returns two output vectors of bounding boxes where first output returns confidence for each class (+ background class) and second output returns bounding box coordinates (x1, y1, x2,y2)

Recommended Platforms

Platform	Supported	Recommended
STM32L0	[]	[]
STM32L4	[]	[]
STM32U5	[]	[]
STM32H7	[]	[]
STM32MP1	[]	[]
STM32MP2	[]	[]
STM32N6	[x]	[x]

Performances

Metrics

Measures are done with default STEdgeAI Core configuration with enabled input / output allocated option.

Reference NPU memory footprint based on COCO dataset (see Accuracy for details on dataset)

Model	Dataset	Format	Resolution	Series	Internal RAM (KiB)	External RAM (KiB)	Weights Flash (KiB)	STEdgeAI Core version
ssdlite_mobilenetv2_pt	COCO	Int8	300x300x3	STM32N6	2408.59	2109.38	5121.38	3.0.0

Reference NPU inference time based on COCO dataset (see Accuracy for details on dataset)

Model	Dataset	Format	Resolution	Board	Execution Engine	Inference time (ms)	Inf / sec	STEdgeAI Core version
ssdlite_mobilenetv2_pt	COCO	Int8	300x300x3	STM32N6570-DK	NPU/MCU	121.02	8.27	3.0.0

Reference NPU memory footprint based on COCO Person dataset (see Accuracy for details on dataset)

Model	Dataset	Format	Resolution	Series	Internal RAM (KiB)	External RAM (KiB)	Weights Flash (KiB)	STEdgeAI Core version
ssdlite_mobilenetv2_pt	COCO-Person	Int8	300x300x3	STM32N6	2374.34	2109.38	3758.63	3.0.0

Reference NPU inference time based on COCO Person dataset (see Accuracy for details on dataset)

Model	Dataset	Format	Resolution	Board	Execution Engine	Inference time (ms)	Inf / sec	STEdgeAI Core version
ssdlite_mobilenetv2_pt	COCO-Person	Int8	300x300x3	STM32N6570-DK	NPU/MCU	117.25	8.53	3.0.0

Reference NPU memory footprint based on VOC dataset (see Accuracy for details on dataset)

Model	Dataset	Format	Resolution	Series	Internal RAM (KiB)	External RAM (KiB)	Weights Flash (KiB)	STEdgeAI Core version
ssdlite_mobilenetv2_pt	VOC	Int8	300x300x3	STM32N6	2374.81	2109.38	4086.38	3.0.0

Reference NPU inference time based on VOC dataset (see Accuracy for details on dataset)

Model	Dataset	Format	Resolution	Board	Execution Engine	Inference time (ms)	Inf / sec	STEdgeAI Core version
ssdlite_mobilenetv2_pt	VOC	Int8	300x300x3	STM32N6570-DK	NPU/MCU	118.10	8.47	3.0.0

AP on COCO dataset

Dataset details: link , License CC BY 4.0, Number of classes: 80

Model	Format	Resolution	AP50
ssdlite_mobilenetv2_pt	Float	3x300x300	29.03
ssdlite_mobilenetv2_pt	Int8	3x300x300	28.60

* EVAL_IOU = 0.5, NMS_THRESH = 0.5, SCORE_THRESH = 0.001, MAX_DETECTIONS = 100

AP on COCO-Person dataset

Dataset details: link , License CC BY 4.0 , Number of classes: 1

Model	Format	Resolution	AP50
ssdlite_mobilenetv2_pt	Float	3x300x300	39.90
ssdlite_mobilenetv2_pt	Int8	3x300x300	39.71

* EVAL_IOU = 0.5, NMS_THRESH = 0.5, SCORE_THRESH = 0.001, MAX_DETECTIONS = 100

AP on VOC dataset

Dataset details: link , License , Number of classes: 20

Model	Format	Resolution	AP50
ssdlite_mobilenetv2_pt	Float	3x300x300	70.17
ssdlite_mobilenetv2_pt	Int8	3x300x300	70.07

* EVAL_IOU = 0.5, NMS_THRESH = 0.5, SCORE_THRESH = 0.001, MAX_DETECTIONS = 100

Retraining and Integration in a simple example:

Please refer to the stm32ai-modelzoo-services GitHub here

References

• COCO
  Lin et al., Microsoft COCO: Common Objects in Context

• PASCAL VOC
  Everingham et al., The PASCAL Visual Object Classes (VOC) Challenge

• SSDLite
  Howard et al., MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications (SSDLite section)

• MobileNetV2
  Sandler et al., MobileNetV2: Inverted Residuals and Linear Bottlenecks