Week 3

• Exporting CSV from sqlite3 cli

  .headers on
  .mode csv
  .output data.csv
  SELECT * from labelled_frames;
  .quit

Questions for Vannary on localrunner.sh for CCTV model training.

1. What steps come after generating labels from Image_label notebook?

    video_id,fname,labels,video_group,condition_code_list,OCR_distance,distance,continuous_defect
    "1212022-112949 AM-COREY BREWER","1212022-112949 AM-COREY BREWER_000000",ND,01202023,"",0.0,,
    "1212022-112949 AM-COREY BREWER","1212022-112949 AM-COREY BREWER_000030",ND,01202023,"",0.0,,
    "1212022-112949 AM-COREY BREWER","1212022-112949 AM-COREY BREWER_000060",ND,01202023,"",,,
2. We ran the following command to run the training script:

    ./localrunner.sh -s 06_multicat_clean_sewerml_v3 -r 123 -b 2 -e 1 -i train00 -m test_model -t train00
   • Where should we put the blurred images? (Is it <current_working_directory>/train/)
3. Is this the format that training script expects as input? How the output of image-label is converted into that format?

Task list:

CCTV:

• Change the 10th point in the CCTV workflow - Replace 'distance labels' with correct data description.