Human Activity Recognition STM32 Model Zoo

January 22, 2026 ยท View on GitHub

Directory components:

  • config_file_examples includes the sample minimalistic configuration files to run different operation modes, such as training, deployment, evaluation and benchmarking. These files can be used directly to launch one of the operation modes with little to no editions.
  • datasets placeholder for the human activity recognition datasets.
  • docs contains all readmes and documentation specific to the human activity recognition use case.
  • tf contains all the source code to build, train, evaluate, benchmark, quantize and deploy your human_activity_recognition model on your STM32 target.

Quick & easy examples:

The operation_mode top-level attribute specifies the operations or the service you want to execute. This may be single operation or a set of chained operations.

You can refer to following readme tutorials that provide typical examples of operation modes:

All .yaml configuration examples can be found in the config_file_examples folder.

The different values of the operation_mode attribute and the corresponding operations are described in the table below. In the names of the chain modes, 't' stands for training, and 'b' for benchmark on an STM32 board.

operation_mode attributeOperations
trainingTrain an HAR model
evaluationEvaluate the accuracy of a pretrained float model on a test or validation dataset
benchmarkingBenchmark a pretrained float model on an STM32 board
deploymentDeploy a pretrained float model on an STM32 board
chain_tbSequentially: train, and then benchmark an HAR model

You don't know where to start? You feel lost?

Don't forget to follow our tuto below for a quick ramp up :

Remember that minimalistic yaml files are available here to play with specific services, and that all pre-trained models in the STM32 model zoo are provided with their configuration .yaml file used to generate them. These are very good starting points to start playing with!