Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Highlighted
23 Views

Cannot Run 8 bit quantized model

After lots of struggle getting the accuracy checker to work on a resnet model with my data, I managed to run the Post-Training Optimization Command-line Tool.  It finished and created .xml, .bin, and .mapping files.  I set up a YAML file to run this model through the accuracy checker.  When I do, I get a "Can't find layer" error, detailed below.  Any tips on how to deal with this?

 

IE version: 2.1.37988
Loaded CPU plugin version: 2.1.37988
Input info:
        Layer name: input
        precision: FP32
        shape [1, 3, 424, 512]

Output info
        Layer name: output
        precision: FP32
        shape [1, 2]

16:57:56 accuracy_checker ERROR: Can't find layer 507/fq_input_1 among layer 505/fq_input_0 outputs
Traceback (most recent call last):
  File "C:\Program Files (x86)\IntelSWTools\openvino\deployment_tools\open_model_zoo\tools\accuracy_checker\accuracy_checker\main.py", line 227, in main
    evaluator = evaluator_class.from_configs(config_entry)
  File "C:\Program Files (x86)\IntelSWTools\openvino\deployment_tools\open_model_zoo\tools\accuracy_checker\accuracy_checker\evaluators\model_evaluator.py", line 72, in from_configs
    launcher = create_launcher(launcher_config)
  File "C:\Program Files (x86)\IntelSWTools\openvino\deployment_tools\open_model_zoo\tools\accuracy_checker\accuracy_checker\launcher\launcher.py", line 198, in create_launcher
    return Launcher.provide(config_framework, launcher_config, delayed_model_loading=delayed_model_loading)
  File "C:\Program Files (x86)\IntelSWTools\openvino\deployment_tools\open_model_zoo\tools\accuracy_checker\accuracy_checker\dependency.py", line 67, in provide
    return root_provider(*args, **kwargs)
  File "C:\Program Files (x86)\IntelSWTools\openvino\deployment_tools\open_model_zoo\tools\accuracy_checker\accuracy_checker\launcher\dlsdk_launcher.py", line 237, in __init__
    self.load_network(log=True)
  File "C:\Program Files (x86)\IntelSWTools\openvino\deployment_tools\open_model_zoo\tools\accuracy_checker\accuracy_checker\launcher\dlsdk_launcher.py", line 673, in load_network
    self.exec_network = self.plugin.load(network=self.network, num_requests=self.num_requests)
  File "ie_api.pyx", line 1205, in openvino.inference_engine.ie_api.IEPlugin.load
  File "ie_api.pyx", line 1215, in openvino.inference_engine.ie_api.IEPlugin.load
RuntimeError: Can't find layer 507/fq_input_1 among layer 505/fq_input_0 outputs

 

0 Kudos
3 Replies
Highlighted
Employee
23 Views

Hi Joshua,

Please share the model, yml file and other required files so that we could look at the the issue from our end.

Regards,

Jaivin

0 Kudos
Highlighted
23 Views

Find those files attached.  Let me know if you need anything else.  Thanks!

0 Kudos
Highlighted
Employee
23 Views

Hi Joshua,

Apologies for the late response.

Can you please attach the .bin files of the models to try it from our end?

Also, Have you tried this model with any samples/demos?

If not, can you try loading the model to IE using benchmark application?

Regards,

Jaivin

0 Kudos