OpenVino Runtime Exception. Unexpected: CPU plug-in doesn't support If operation with dynamic rank. Operation name: input.15

[dsakharuk]

Describe the issue

I built ONNX from source using OpenVino 2024.6. My model was converted from Pytorch to ONNX and it runs well with the default CPU execution provider. I am trying to use OpenVino to increase inference speed. When trying to run my model for inferencing I am getting the following error:

[OpenVINO-EP] Exception while Loading Network for graph: OpenVINOExecutionProvider_OpenVINO-EP-subgraph_1_0Exception from src/inference/src/cpp/core.cpp:104: Exception from src/inference/src/dev/plugin.cpp:53: Exception from src/plugins/intel_cpu/src/node.cpp:87: Unexpected: CPU plug-in doesn't support If operation with dynamic rank. Operation name: input.15

I am wondering if it is possible to enable OpenVino to fall back to CPU execution provider for unsupported operations or if there is any other solution to this error.

Thanks.

To reproduce

ONNX library was built using this line:

./build.sh --config Release --use_openvino CPU --build_shared_lib --compile_no_warning_as_error --parallel --skip_submodule_sync

I am using the following OpenVino options to run on one thread and one core:

std::unordered_map<std::string, std::string> options; options["device_type"] = "CPU"; options["precision"] = "FP32"; options["num_of_threads"] = "1"; options["num_streams"] = "1"; options["enable_opencl_throttling"] = "false"; session_options_.AppendExecutionProvider("OpenVINO", options);

Urgency

No response

Platform

Linux

OS Version

Centos 7

ONNX Runtime Installation

Built from Source

ONNX Runtime Version or Commit ID

1.21.0

ONNX Runtime API

C++

Architecture

X64

Execution Provider

OpenVINO

Execution Provider Library Version

2024.6

Model File

No response

Is this a quantized model?

No