AI Potluck
Infrastructure / Core ML frameworks & libraries

ONNX

LF AI & Data (Linux Foundation)

ONNX (Open Neural Network Exchange) is an open standard format and specification for representing machine learning models, defining an extensible computation graph model plus built-in operators and standard data types. It enables interoperability so models can move between frameworks like PyTorch, TensorFlow, and scikit-learn. It is governed under the Linux Foundation as part of LF AI & Data.

Verified live 2026-06-22 via primary sources. Apache-2.0 license body confirmed; full source public on GitHub.

Openness

5 high confidence
5.0
license
Apache-2.0(OSI)
source
public
core-gated
ungated

Apache-2.0 license body confirmed; full source public on GitHub.

Adoption

5 high confidence
5.0

Re-scored 4->5. 20.65M PyPI downloads in the last 30 days. The map's prevailing usage_volume level-5 floor is >10M/mo (cf. pydantic-ai ~31M, llama-index ~10.18M at level 5; the companion onnxruntime is already level 5 at ~74.7M); the prior level 4 applied the stricter ml_frameworks-batch band. ONNX is the de facto cross-framework model-exchange standard and clears the floor.

Capability

5 high confidence
5.0

Defines the graph model, operators, and data types used as the de-facto interchange standard.

Unchanged since 2026-07-30 (last edited, not re-checked)