ONNX
Curated by SignalOpen format framework to represent deep learning models
Overview
ONNX is an open format framework to represent deep learning models that is used for porting models from one framework to another. It supports multiple frameworks including Chainer, Caffe2, Microsoft Cognitive Toolkit, MXNet, and PyTorch. The models are also compatible with multiple hardware runtimes, libraries, and claims to improve the performance of the hardware. It has been developed and supported by AMD, ARM, NVIDIA, HUAWEI, Intel, Preferred Networks, and Qualcomm. Its partners include Facebook Open Source, Microsoft, and Amazon Web Services.
Key facts
- Founded
- 2017
- Headquarters
- Menlo Park, United States
- Stage
- Sign up to view
- Industry
- High Technology