Known Issues and Limitations in the Model Optimizer

Model Optimizer for TensorFlow* should be run on Intel® hardware that supports the AVX instruction set

TensorFlow* provides only prebuilt binaries with AVX instructions enabled. When you're configuring the Model Optimizer by running the install_prerequisites or install_prerequisites_tf scripts, they download only those ones, which are not supported on hardware such as Intel® Pentium® processor N4200/5, N3350/5, N3450/5 (formerly known as Apollo Lake).

To run the Model Optimizer on this hardware, you should compile TensorFlow binaries from source as described at the TensorFlow website.

Another option is to run the Model Optimizer to generate an IR on hardware that supports AVX to and then perform inference on hardware without AVX.

Multiple OpenMP Loadings

If the application uses the Inference Engine with third-party components that depend on Intel OpenMP, multiple loadings of the libiomp library may occur and cause OpenMP runtime initialization conflicts. This may happen, for example, if the application uses Intel® Math Kernel Library (Intel® MKL) through the “Single Dynamic Library” ( mechanism and calls Intel MKL after loading the Inference Engine plugin. The error log looks as follows:

OMP: Error #15: Initializing, but found already initialized.
OMP: Hint: This means that multiple copies of the OpenMP runtime have been linked into the program. That is dangerous, since it can degrade performance or cause incorrect results. The best thing to do is to ensure that only a single OpenMP runtime is linked into the process, e.g. by avoiding static linking of the OpenMP runtime in any library. As an unsafe, unsupported, undocumented workaround you can set the environment variable KMP_DUPLICATE_LIB_OK=TRUE to allow the program to continue to execute, but that may cause crashes or silently produce incorrect results. For more information, please see

Possible workarounds:

Old proto compiler breaks protobuf library

With python protobuf library version 3.5.1 the following incompatibility can happen. The known case is for Cent OS 7.4

The error log looks as follows:

File "../lib64/python3.5/site-packages/google/protobuf/", line 829, in _new_
return _message.default_pool.AddSerializedFile(serialized_pb)
TypeError: expected bytes, str found

Possible workaround is to upgrade default protobuf compiler (libprotoc 2.5.0) to newer version, for example libprotoc 2.6.1.

Post-Training Optimization Tool fails on some IRs, generated in the OpenVINO 2020.1 version

For the 2020.1 release, several operations were added to opset1 by mistake. In 2020.2 this issue was fixed and the operations were moved to opset2. However running Post-Training Optimization tool from 2020.2 or higher version with an IR generated with the 2020.1 version may cause tool crash.

The error log looks as follows:

File "..\model-optimizer\mo\utils\ir_reader\", line 89, in save_restored_graph
prepare_emit_ir(graph, precision, path, name, meta_info=meta_data)
File "..\model-optimizer\mo\pipeline\", line 223, in prepare_emit_ir
for_graph_and_each_sub_graph_recursively(graph, OpVersioning().find_and_replace_pattern)
File "..\model-optimizer\mo\middle\", line 58, in for_graph_and_each_sub_graph_recursively
File "..\model-optimizer\extensions\back\", line 165, in find_and_replace_pattern
'please use another'.format(name))
mo.utils.error.Error: Node ... has `version` attribute set to `opset1`, but it is a reserved word, please use another

Possible workarounds: