intel / onnxruntime

ONNX Runtime: cross-platform, high performance scoring engine for ML models
MIT License
56 stars 22 forks source link

Export/Import OV Compiled blobs as EPContext Models #356

Closed sspintel closed 4 months ago

sspintel commented 4 months ago

This PR adds a new provider option export_ep_ctx_blob to let OVEP export an 'optimized' onnx model with the compiled blob as one of the parameters of the EPContext contrib op of ORT.

sfatimar commented 4 months ago

@sspintel can you please upload a small design document and unit test document with this, Please also include guidance to validation team for system tests to be conducted for review

sspintel commented 4 months ago

@sfatimar It is yet to be tested on Windows.

sspintel commented 4 months ago

Merging PR to kick-off automated builds