Tool/software:
Hi experts,
I am seeing using the ARC fault model on model composer. After training the model, I compile the model but saw a failure. Below is the error log:
The current compilation run has failed.
No log file was produced
/opt/tinyml/.pyenv/versions/py310/lib/python3.10/site-packages/edgeai_torchmodelopt/xmodelopt/quantization/v2/qconfig_types.py:48: UserWarning: could not find _get_default_qconfig_mapping_with_default_qconfig in torch.ao.quantization.qconfig_mapping
warnings.warn("could not find _get_default_qconfig_mapping_with_default_qconfig in torch.ao.quantization.qconfig_mapping")
Traceback (most recent call last):
File "/opt/tinyml/code/tinyml-modelmaker/tinyml_modelmaker/utils/config_dict.py", line 77, in __getattr__
return self[key]
KeyError: 'model_export_path_quantization'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/opt/tinyml/code/tinyml-mlbackend/./run.py", line 138, in <module>
main(vars(args))
File "/opt/tinyml/code/tinyml-mlbackend/./run.py", line 129, in main
run_modelmaker.main(config)
File "/opt/tinyml/code/tinyml-modelmaker/scripts/run_tinyml_modelmaker.py", line 97, in main
model_runner.run()
File "/opt/tinyml/code/tinyml-modelmaker/tinyml_modelmaker/ai_modules/timeseries/runner.py", line 208, in run
self.model_compilation.run()
File "/opt/tinyml/code/tinyml-modelmaker/tinyml_modelmaker/ai_modules/common/compilation/tinyml_benchmark.py", line 117, in run
model_file = self.params.training.model_export_path_quantization if self.params.training.quantization != TinyMLQuantizationVersion.NO_QUANTIZATION else self.params.training.model_export_path
File "/opt/tinyml/code/tinyml-modelmaker/tinyml_modelmaker/utils/config_dict.py", line 79, in __getattr__
raise AttributeError(key)
AttributeError: model_export_path_quantization
May I know how to solve this?
Regards,
Hang.
