Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Lin__Joey
Innovator
231 Views

issues on using post_training_optimization_toolkit with faster_rcnn_resnet50_coco

Jump to solution

Dear someone,

I am testing the openVino 2020.2, released just few days ago, on its post training optimization tool functions.

I followed the README.md (see attachment) in the sample config folder, and  picked up the faster_rcnn_resent50_coo.yaml 

for accuracy check test. I updated the sample config file with the path of downloaded val2017, and annotation file instances_val2017.json.

Finally, I ran the command, 

python main.py -c  faster_rcnn_resnet50_coco.yaml

The error shows I must specify model name, launchers, datasets (see attached JPEG). It seems like that I have messed up with the path, but whatever I tried

, no improvement has been made.  Could some one help figure out what went wrong?

The error:

~\post_training_optimization_toolkit>python main.py -c faster_rcnn_resnet50_coco.yaml

......

......

raise ConfigError('Each model must specify {}'.format(', '.join(required_model_entries)))
libs.open_model_zoo.tools.accuracy_checker.accuracy_checker.config.config_validator.ConfigError: Each model must specify name, launchers, datasets

Thanks in advance,

Joey Lin

 

 

 

 

 

0 Kudos
1 Solution
SIRIGIRI_V_Intel
Employee
231 Views

Can you confirm that you have followed the installation steps for Post-Training Optimization Toolkit.

Please follow the documentation and provide the config file with parameters required.

Regards,

Ram prasad

View solution in original post

2 Replies
SIRIGIRI_V_Intel
Employee
232 Views

Can you confirm that you have followed the installation steps for Post-Training Optimization Toolkit.

Please follow the documentation and provide the config file with parameters required.

Regards,

Ram prasad

View solution in original post

Lin__Joey
Innovator
231 Views

Hi, Ram:

Thank you for the documentation link. I followed the steps and successfully optimized faster_rcnn_resnet50_coco model.

regards,

Joey

 

 

 

Reply