GAIA-ML

GAIA-ML is an experimental compiler that turns ONNX models into optimized C++ kernels and graphs for AMD AI Engine v1, targeting the Versal Premium VP2802-2M. It supports Dense/MLP networks, supported CNN patterns, and bounded tree ensembles. Neural-network kernels support float32, int32, int16 and int8.

GAIA selects implementations, fuses supported operations, packs weights and partitions work across AIE tiles. It generates an ADF graph, kernels, test vectors, build files and reports describing resource use and estimated latency. Generating a project does not require Vitis. Building and simulating the emitted AIE code requires the AMD toolchain separately.

Install

From a checkout, with Python 3.9 or newer:

python3 -m venv .venv
source .venv/bin/activate
python -m pip install -e .
gaia_ml version

Compile a model

Inspect the model and target constraints, then generate a project:

gaia_ml inspect model.onnx --stage constraints

gaia_ml compile model.onnx \
  --output gaia_project \
  --schedule-objective latency \
  --data-file input.csv \
  --simulation-events 10

Each CSV row is one flattened input event; provide at least as many rows as --simulation-events. For dynamic inputs, supply a static --input-shape, for example --input-shape 1,3,18,18. Use a new output directory or --force to regenerate an existing project.

Useful compilation options:

The generated graph/, kernels/ and weights/ contain the AIE implementation. manifest.json records the selected schedule, constraints and latency estimates; execution_plan.json records resolved execution choices.

Build and validate

After loading your AMD Vitis environment:

cd gaia_project
make aie
make aiesim
cd ..
gaia_ml verify gaia_project

GAIA is not a general-purpose ONNX backend: operator, shape and resource limits apply. Integer conversion can change model accuracy. Latency estimates are planning predictions, not hardware guarantees; validate numerical outputs and measured timing for your model.

Documentation and development

• Run gaia_ml compile --help for all compilation options.
• See the compiler stages for IRs and optimization details.
• See release notes for version history.
• Run tests with python -m unittest discover -s tests.

Developed by the NGT WP2.1 Group. Core developer: Ioannis Xiotidis. Distributed under the MIT license.