accuracy_checker ERROR: 2

[Aurjay]

I get these errors for certain tasks in my project. Can someone help find the reason for this error and possible solutions to try?

[Wan_Intel]

Hi Aurjay,

Thanks for reaching out to us.

Could you please share the following information with us for further investigation?

·      Which version of the OpenVINO™ toolkit are you using?

·      Could you please share your source model and custom model with us?

·      Could you please share the config file you used with the Accuracy Checker?

·      Could you share the command you use to run Accuracy Checker?

Regards,

Wan

[Aurjay]

The version of openvino is 2021

The version of accuracy checker is 0.8.7

config file format :

models:

    - name: resnet_50

      launchers:

        - framework: dlsdk

          device: CPU

          model: "model .xml file path"

          weights: "model .bin file path"

          adapter: ssd

          batch: 8

      datasets:

          - name: Task_1

            data_source: "data img folder path"

            annotation_conversion:

              converter: mscoco_detection

              annotation_file: "annotations.json path"

              dataset_meta_file: "datasetmeta.json path"

            preprocessing:

            - type: resize

              dst_width: 544

              dst_height: 320

            postprocessing:

            - type: resize_prediction_boxes

            metrics:

            - type: map

            - type: coco_precision

Accuracy check command:

accuracy_check -c "path to config.yml" -m "path to model" --log_file "path to log.txt"

Btw. The logs getting saved in log.txt is getting saved for each dataset individually. Is there a way to get a average value because for example individual map values are not helpful but the average of map values for all the dataset is helpful to analyze the model performance.

[Wan_Intel]

Hi Aurjay,

Thanks for your information.

For your information, you can refer to Metrics in Deep Learning accuracy validation framework to compare predicted data with annotated data and perform quality measurement. On another note, you may refer to resnet-50-tf accuracy check yml file and resnet-50-pytorch accuracy check yml file to see how to prepare yml file for resnet model. On the other hand, could you please share your log.txt file, resnet_50 model and dataset with us for further investigation?

Regards,

Wan

[Wan_Intel]

Hi Aurjay,

Thanks for your question.

This thread will no longer be monitored since we have provided suggestions. 

If you need any additional information from Intel, please submit a new question.

Best regards,

Wan