Illegal instruction error runnig model optimizer

[FPART1]

Dear all,

I'm getting error trying to optimize a model downloaded from OpenVino modelzoo, please find message below:

$ sudo python3 mo_tf.py --input_model /home/felipe/Desktop/projeto/inception-resnet-v2-tf/inception_resnet_v2.pb --log_level DEBUG
[ 2020-04-14 21:32:07,018 ] [ DEBUG ] [ main:118 ]  Namespace(batch=None, blobs_as_inputs=True, data_type='float', disable_fusing=None, disable_gfusing=None, disable_nhwc_to_nchw=False, disable_resnet_optimization=False, disable_weights_compression=False, enable_concat_optimization=False, extensions='/opt/intel/openvino_2020.2.120/deployment_tools/model_optimizer/extensions', finegrain_fusing=None, framework='tf', freeze_placeholder_with_value=None, generate_deprecated_IR_V2=False, generate_deprecated_IR_V7=False, generate_experimental_IR_V10=True, input=None, input_checkpoint=None, input_meta_graph=None, input_model='/home/felipe/Desktop/projeto/inception-resnet-v2-tf/inception_resnet_v2.pb', input_model_is_text=False, input_shape=None, keep_quantize_ops_in_IR=True, keep_shape_ops=False, log_level='DEBUG', mean_values=(), model_name=None, move_to_preprocess=False, output=None, output_dir='/opt/intel/openvino_2020.2.120/deployment_tools/model_optimizer/.', progress=False, reverse_input_channels=False, saved_model_dir=None, saved_model_tags=None, scale=None, scale_values=(), silent=False, stream_output=False, tensorboard_logdir=None, tensorflow_custom_layer_libraries=None, tensorflow_custom_operations_config_update=None, tensorflow_object_detection_api_pipeline_config=None, tensorflow_use_custom_operations_config=None, transformations_config=None)
[ 2020-04-14 21:32:07,018 ] [ DEBUG ] [ main:119 ]  Model Optimizer started
[ 2020-04-14 21:32:07,018 ] [ DEBUG ] [ main:134 ]  Output model name would be inception_resnet_v2{.xml, .bin}
Model Optimizer arguments:
Common parameters:
    - Path to the Input Model:     /home/felipe/Desktop/projeto/inception-resnet-v2-tf/inception_resnet_v2.pb
    - Path for generated IR:     /opt/intel/openvino_2020.2.120/deployment_tools/model_optimizer/.
    - IR output name:     inception_resnet_v2
    - Log level:     DEBUG
    - Batch:     Not specified, inherited from the model
    - Input layers:     Not specified, inherited from the model
    - Output layers:     Not specified, inherited from the model
    - Input shapes:     Not specified, inherited from the model
    - Mean values:     Not specified
    - Scale values:     Not specified
    - Scale factor:     Not specified
    - Precision of IR:     FP32
    - Enable fusing:     True
    - Enable grouped convolutions fusing:     True
    - Move mean values to preprocess section:     False
    - Reverse input channels:     False
TensorFlow specific parameters:
    - Input model in text protobuf format:     False
    - Path to model dump for TensorBoard:     None
    - List of shared libraries with TensorFlow custom layers implementation:     None
    - Update the configuration file with input/output node names:     None
    - Use configuration file used to generate the model with Object Detection API:     None
    - Use the config file:     None
Model Optimizer version:     2020.2.0-60-g0bc66e26ff
Illegal instruction

I tried to upgrade OpenVino to the latest version, but the error persists. After the upgrade I run the installation check script which was able to optimize the demo model:

========= Converting squeezenet1.1 to IR (FP16)
Conversion command: /home/felipe/.virtualenvs/cv/bin/python3 -- /opt/intel/openvino_2019.3.334/deployment_tools/model_optimizer/mo.py --framework=caffe --data_type=FP16 --output_dir=/home/felipe/openvino_models/ir/public/squeezenet1.1/FP16 --model_name=squeezenet1.1 '--input_shape=[1,3,227,227]' --input=data '--mean_values=data[104.0,117.0,123.0]' --output=prob --input_model=/home/felipe/openvino_models/models/public/squeezenet1.1/squeezenet1.1.caffemodel --input_proto=/home/felipe/openvino_models/models/public/squeezenet1.1/squeezenet1.1.prototxt

Model Optimizer arguments:
Common parameters:
    - Path to the Input Model:     /home/felipe/openvino_models/models/public/squeezenet1.1/squeezenet1.1.caffemodel
    - Path for generated IR:     /home/felipe/openvino_models/ir/public/squeezenet1.1/FP16
    - IR output name:     squeezenet1.1
    - Log level:     ERROR
    - Batch:     Not specified, inherited from the model
    - Input layers:     data
    - Output layers:     prob
    - Input shapes:     [1,3,227,227]
    - Mean values:     data[104.0,117.0,123.0]
    - Scale values:     Not specified
    - Scale factor:     Not specified
    - Precision of IR:     FP16
    - Enable fusing:     True
    - Enable grouped convolutions fusing:     True
    - Move mean values to preprocess section:     False
    - Reverse input channels:     False
Caffe specific parameters:
    - Path to Python Caffe* parser generated from caffe.proto:     /opt/intel/openvino_2019.3.334/deployment_tools/model_optimizer/mo/front/caffe/proto
    - Enable resnet optimization:     True
    - Path to the Input prototxt:     /home/felipe/openvino_models/models/public/squeezenet1.1/squeezenet1.1.prototxt
    - Path to CustomLayersMapping.xml:     Default
    - Path to a mean file:     Not specified
    - Offsets for a mean file:     Not specified
Model Optimizer version:     2019.3.0-375-g332562022
[ WARNING ]  
Detected not satisfied dependencies:
    protobuf: installed: 3.10.0, required: 3.6.1

Please install required versions of components or use install_prerequisites script
/opt/intel/openvino_2019.3.334/deployment_tools/model_optimizer/install_prerequisites/install_prerequisites_caffe.sh
Note that install_prerequisites scripts may install additional components.

[ SUCCESS ] Generated IR model.
[ SUCCESS ] XML file: /home/felipe/openvino_models/ir/public/squeezenet1.1/FP16/squeezenet1.1.xml
[ SUCCESS ] BIN file: /home/felipe/openvino_models/ir/public/squeezenet1.1/FP16/squeezenet1.1.bin
[ SUCCESS ] Total execution time: 8.56 seconds.

Could you please advice what is wrong?

My environment is:

Ubuntu 18.04 LTS

Python 3.6

Tensorflow 1.12

CUDA 10

OpenVino 2020.2

Thanks in advance,

Felipe

[JAIVIN_J_Intel]

Hi Felipe,

Use the --input_shape parameter with Model Optimizer command. You may run Model Optimizer with the --help option to know more about the notation for input shapes.

You may also refer the inception-resnet-v2-tf model description.

Regards,

Jaivin

[FPART1]

Hi Jaivin,

Thanks for the reply, actually there was a issue in my environment preventing Model Optimizer to run properly. I believe this happened due to Keras installation which corrupted my TensorFlow backend.

I've set up na fresh environment so the ilegal instruction is no longer happeing.

Thanks,

Felipe

[lyumax]

if it's not difficult, can you explain in more detail how you got rid of this problem?
You reinstalled OpenVINO or what?

Thank you for answer.

[JAIVIN_J_Intel]

Hi Felipe,

Thank you for sharing the information with the OpenVINO Community!

Regards,

Jaivin