stm32ai-modelzoo

STMicroelectronics – STM32 model zoo

Welcome to STM32 model zoo!

🎉 We are excited to announce that the STM32 AI model zoo now includes comprehensive PyTorch support, joining TensorFlow and ONNX. It now features a vast library of PyTorch models, all seamlessly integrated with our end-to-end workflows. Whether you want to train, evaluate, quantize, benchmark, or deploy, you’ll find everything you need – plus the flexibility to choose between PyTorch, TensorFlow, and ONNX. Dive into the expanded STM32 model zoo and take your AI projects further than ever on STM32 devices.

---

The STM32 AI model zoo is a collection of reference machine learning models that are optimized to run on STM32 microcontrollers. Available on GitHub, this is a valuable resource for anyone looking to add AI capabilities to their STM32-based projects.

• A large collection of application-oriented models ready for re-training
• Pre-trained models on reference datasets, available across several frameworks.

Scripts to easily retrain, quantize, evaluate or benchmark any model from user datasets as well as application code examples automatically generated from user AI model can be found in the stm32ai-modelzoo-services GitHub

These models can be useful for quick deployment if you are interested in the categories that they were trained. We also provide training scripts to do transfer learning or to train your own model from scratch on your custom dataset.

Performance metrics on reference STM32 MCU, NPU, and MPU are provided for both float and quantized models.

What’s new in releases :

4.1:

• Support of STEdgeAI Core v4.0.0.
• Updated Audio Event Detection (AED) to support deployment on STM32U3.
• Added support for the YOLO26 model.
• Multiple bug fixes and overall quality improvements.

4.0:

• Major PyTorch support for Image Classification (IC) and Object Detection (OD)
• Support of STEdgeAI Core v4.0.0
• New training and evaluation scripts for PyTorch models
• Expanded model selection and improved documentation
• Unified workflow for TensorFlow and PyTorch
• Performance and usability improvements
• New use cases: Face Detection (FD), Arc Fault Detection (AFD), Re-Identification (ReID)
• New mixed precision models (Weights 4-bits, Activations 8-bits) for IC and OD use cases
• Support for Keras 3.8.0, TensorFlow 2.18.0, PyTorch 2.7.1, and ONNX 1.16.1
• Python software architecture rework
• Docker-based setup available, with a ready-to-use image including the full software stack.

3.2:

• Support for STEdgeAI Core v2.2.0 (STM32Cube.AI v10.2.0).
• Support of X-Linux-AI v6.1.0 support for MPU.
• New use cases added: StyleTransfer and FastDepth.
• New models added: Face Detection, available in the Object Detection use case, and Face Landmarks, available in the Pose Estimation use case.
• Architecture and codebase clean-up.

3.1:

• Included additional models support (yolov11n, st_yoloxn variants)

3.0:

• Included additional models compatible with the STM32N6570-DK board.
• Expanded models in all use cases.
• Expanded use case support to include Instance Segmentation and Speech Enhancement.
• Added Pytorch support through the speech enhancement Use Case.
• Model Zoo hosted on Hugging Face

2.1:

• Included additional models compatible with the STM32MP257F-EV1 board.
• Expanded use case support to include Pose Estimation and Semantic Segmentation.

2.0:

• An aligned and uniform architecture for all the use case

Find below a summary of available use cases

Use Case	Quick definition	Suitable Targets for deployment	Smart example
Image Classification	Classifies the content of an image within a predefined set of classes.	STM32H747I-DISCO NUCLEO-H743ZI2 STM32MP257F-EV1 STM32N6570-DK
Object Detection	Detects, locates and estimates the occurences probability of predefined objects from input images.	STM32H747I-DISCO STM32MP257F-EV1 STM32N6570-DK
Face Detection	Detects, locates and estimates the occurences probability of predefined faces and keypoints from input images.	STM32H747I-DISCO STM32N6570-DK
Pose Estimation	Detects key points on some specific objects (people, hand, face, …).	STM32MP257F-EV1 STM32N6570-DK
Semantic Segmentation	Associates a label to every pixel in an image to recognize a collection of pixels that form distinct categories.	STM32MP257F-EV1 STM32N6570-DK
Instance Segmentation	Associates a label to every pixel in an image to recognize a collection of pixels that form distinct categories or instances of each category.	STM32N6570-DK
Depth Estimation	Predict the distance to objects from an image as a pixel-wise depth map.	STM32N6570-DK
Neural Style Transfer	applies the artistic style of one image to the content of another image.	STM32N6570-DK
Audio Event Detection	Detection of a specific audio events.	B-U585I-IOT02A B-U585I-IOT02A STM32N6570-DK
Speech Enhancement	Enhancement of the audio perception in a noisy environment.	STM32N6570-DK
Human Activity Recognition	Recognizes various activities like walking, running, …	B-U585I-IOT02A
Hand Posture Recognition	Recognizes a set of hand postures using Time of Flight (ToF) sensor	NUCLEO-F401RE
Arc Fault Detection	Detects electrical arc faults from current signals.	B-U585I-IOT02A
Re-Identification	Reidentifies a person or object across different images or video frames.	STM32N6570-DK

Hugging Face host

The Model Zoo Dashboard is hosted in a Docker environment under the STMicroelectronics Organization. This dashboard is developed using Dash Plotly and Flask, and it operates within a Docker container. It can also run locally if Docker is installed on your system. The dashboard provides the following features:

• Training: Train machine learning models. • Evaluation: Evaluate the performance of models. • Benchmarking: Benchmark your model using ST Edge AI Developer Cloud • Visualization: Visualize model performance and metrics. • User Configuration Update: Update and modify user configurations directly from the dashboard. • Output Download: Download model results and outputs.

You can also find our models on Hugging Face under the STMicroelectronics Organization. Each model from the STM32AI Model Zoo is represented by a model card on Hugging Face, providing all the necessary information about the model and linking to dedicated scripts.

Before you start

The model zoo repo is using the git lfs, so the users need to install and set up the git lfs before cloning the repo. This can be done by following the instructions below.

• On Ubuntu:

  sudo apt-get install git-lfs
  

• On Windows: Download the Git LFS extension here.

Once downloaded and installed, set up Git LFS for your user account by running the following command:

git lfs install

You should see the message Git LFS initialized. if the command runs successfully.

NOTE: If you do not see the message Git LFS initialized., visit the GitHub documentation page for more details and support.

Clone the Repository

In order to use the config files examples from the modelzoo-services, without the need to change the model_path, it is needed to clone the modelzoo below in the same folder as the modelzoo services.

git clone https://github.com/STMicroelectronics/stm32ai-modelzoo.git