Skip to content

OpenVINO EP v2.0#3585

Merged
jywu-msft merged 286 commits intomicrosoft:masterfrom
intel:openvino-ep-v2
Apr 24, 2020
Merged

OpenVINO EP v2.0#3585
jywu-msft merged 286 commits intomicrosoft:masterfrom
intel:openvino-ep-v2

Conversation

@smkarlap
Copy link
Contributor

Description:
Version update for OpenVINO Execution Provider. Based on the latest OpenVINO release 2020.2 which now includes nGraph stack. Model conversion from ONNX to OpenVINO IR is now done using nGraph ONNX importer component.

Motivation and Context

  • Supports a larger set of ONNX operators
  • Supports inference of partial models using sub-graph partitioning
  • Supports models with symbolic dimensions in shapes.
  • Uses OpenVINO APIs with Long Term Support (LTS) providing extended support and backward compatibility.
  • Upto 40% reduction in memory footprint.
  • Avoids using the python components used by existing version EP thus avoiding expensive python interpreter startup and module conflicts arising from it.

ynimmaga and others added 30 commits November 14, 2019 14:05
Cleans the backend(intel_graph.*) code in the following ways:-

1. Minimize global usage: Since all the IR graphs need to be
re-generated on every Infer, it is bad practice to rely on globals
for their saving and usage as there would be multiple readers and
writers to the same global variable leading to incorrect usages or
contentions. This change replaces globals with locals where possible.
 This change also fixes an existing bug with due to
incorrect global usage.

2. Remove all unused functions.

3. Remove all unused headers and prepocessor directives.
Backend logic cleanup

See merge request iot-edge/onnxrt-unified-ep/onnxruntime!9
Signed-off-by: suryasidd <surya.siddharth.pemmaraju@intel.com>
Fix missed plugins.xml for python bindings

See merge request iot-edge/onnxrt-unified-ep/onnxruntime!10
Signed-off-by: suryasidd <surya.siddharth.pemmaraju@intel.com>
Signed-off-by: suryasidd <surya.siddharth.pemmaraju@intel.com>
Signed-off-by: suryasidd <surya.siddharth.pemmaraju@intel.com>
Signed-off-by: suryasidd <surya.siddharth.pemmaraju@intel.com>
Signed-off-by: suryasidd <surya.siddharth.pemmaraju@intel.com>
Added environment variable to enable debugging

See merge request iot-edge/onnxrt-unified-ep/onnxruntime!11
Signed-off-by: suryasidd <surya.siddharth.pemmaraju@intel.com>
Signed-off-by: suryasidd <surya.siddharth.pemmaraju@intel.com>
Unsqueeze fix

See merge request iot-edge/onnxrt-unified-ep/onnxruntime!12
Signed-off-by: suryasidd <surya.siddharth.pemmaraju@intel.com>
Overwrites the shape info of Model proto with the shape from
actual input data. Needed for inferring models with Dynamic
shapes.
Signed-off-by: suryasidd <surya.siddharth.pemmaraju@intel.com>
@HectorSVC
Copy link
Contributor

/azp run Linux CPU CI Pipeline,Linux GPU CI Pipeline,Linux GPU TensorRT CI Pipeline,MacOS CI Pipeline,Win CPU x86 CI Pipeline,Windows GPU CI Pipeline,Windows GPU TensorRT CI Pipeline,Win CPU x64 NoContribops CI Pipeline,MacOS NoContribops CI Pipeline,Linux CPU x64 NoContribops CI Pipeline,Windows CPU CI Pipeline,Linux OpenVINO CI Pipeline

@azure-pipelines
Copy link

Azure Pipelines successfully started running 10 pipeline(s).

@smkarlap
Copy link
Contributor Author

The above pipeline failure seems to be due to Inference Engine headers not being found, which is likely due to the OpenVINO paths not being set. The only commit that modified the way paths are set is the recent change that uses the variable OPENVINO_VERSION to set the paths. The docker image built using this changes seems to have all the paths set accurately when tested on my local workstation. However, platform differences between my workstation and the CI machine may be causing this to fail.
I have reverted that commit. Could you please run the pipelines now? If it fails, it will help narrow down the fault.

This is no longer used in recent ONNX update onnx/onnx@da13be2,
so this unset workaround is no longer necessary.
@HectorSVC
Copy link
Contributor

/azp run Linux OpenVINO CI Pipeline

@azure-pipelines
Copy link

Azure Pipelines successfully started running 1 pipeline(s).

Git Tag info for DLDT as well as install directory are set
using this value.

This reverts commit 9fa9c20.
@jywu-msft
Copy link
Member

/azp run Linux OpenVINO CI Pipeline

@azure-pipelines
Copy link

Azure Pipelines successfully started running 1 pipeline(s).

@jywu-msft
Copy link
Member

/azp run Linux CPU CI Pipeline,Linux GPU CI Pipeline,Linux GPU TensorRT CI Pipeline,MacOS CI Pipeline,Win CPU x86 CI Pipeline,Windows GPU CI Pipeline,Windows GPU TensorRT CI Pipeline,Win CPU x64 NoContribops CI Pipeline,MacOS NoContribops CI Pipeline,Linux CPU x64 NoContribops CI Pipeline,Windows CPU CI Pipeline

@azure-pipelines
Copy link

Azure Pipelines successfully started running 9 pipeline(s).

@jywu-msft
Copy link
Member

setup.py has conflict

@jywu-msft
Copy link
Member

/azp run Linux OpenVINO CI Pipeline

@azure-pipelines
Copy link

Azure Pipelines successfully started running 1 pipeline(s).

@HectorSVC
Copy link
Contributor

/azp run Linux CPU CI Pipeline,Linux GPU CI Pipeline,Linux GPU TensorRT CI Pipeline,MacOS CI Pipeline,Win CPU x86 CI Pipeline,Windows GPU CI Pipeline,Windows GPU TensorRT CI Pipeline,Win CPU x64 NoContribops CI Pipeline,MacOS NoContribops CI Pipeline,Linux CPU x64 NoContribops CI Pipeline,Windows CPU CI Pipeline

@azure-pipelines
Copy link

Azure Pipelines successfully started running 9 pipeline(s).

@jywu-msft jywu-msft merged commit 6d4f2f5 into microsoft:master Apr 24, 2020
@smkarlap smkarlap deleted the openvino-ep-v2 branch September 16, 2020 19:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

7 participants