ONNX
ONNX is an open format for exchanging AI models across frameworks and runtimes.
1 article
Last mentioned ONNX, short for Open Neural Network Exchange, is an open model interchange format and specification announced by Microsoft and Facebook, now Meta, in 2017. It represents a model's computation graph and parameters for use across machine-learning frameworks and inference runtimes.
Unlike pickle, which serializes Python objects, or Safetensors, which primarily stores tensor data, ONNX represents both model structure and weights.
This entry is based on AIPOST articles and widely known facts. If something is wrong, please send us a correction request.