Tutorials

ML Tutorials

End-to-end examples tied to current executable sources. C++ tutorials live inTutorial/Ml; paired Python workloads live in Tutorial/Py.

Image Classification — Fashion-MNIST

Train a two-layer MLP on real IDX binary data. Covers OaModule, OaLinear, OaAdamW, implicit autograd, validation, timing, and checkpoint round-trip.

Text Generation — Character GRU

Predict every next character in a 16-token window. Switch between the current C++ tutorial and its direct Python binding port.

Build & Run

All tutorials build from the oa repo root:

cmake --preset release
cmake --build Build/Release --target TutorialMnistClassifierAg TutorialNlpCharGruAg -j

./Bin/Release/Tutorial/Ml/TutorialMnistClassifierAg
./Bin/Release/Tutorial/Ml/TutorialNlpCharGruAg

OA_PYTHON_BUILD_DIR=Build/Release PYTHONPATH=Tutorial/Py \
  Build/PythonVenv/bin/python Tutorial/Py/TutorialNlpCharGruAg.py