Custom datasets about wsolevaluation

[coallaoh]

Thank you for your interest in our work.
To add a dataset, prepare the metadata mimicking the file structures and file formats in e.g. https://github.com/clovaai/wsolevaluation/tree/master/metadata/CUB.

You then need to adapt the code to this new dataset by changing e.g.

• wsolevaluation/config.py

  Line 32 in 45ac3a1

  _DATASET_NAMES = ('CUB', 'ILSVRC', 'OpenImages')

• wsolevaluation/config.py

  Line 122 in 45ac3a1

  def check_dependency(args):

• wsolevaluation/evaluation.py

  Line 468 in 6692538

  evaluator = {"OpenImages": MaskEvaluator,

• wsolevaluation/inference.py

  Line 63 in 8ca3849

  self.evaluator = {"OpenImages": MaskEvaluator,

• wsolevaluation/main.py

  Line 67 in 8ca3849

  _NUM_CLASSES_MAPPING = {

and set the data root (

wsolevaluation/config.py

Line 136 in 6692538

parser.add_argument('--data_root', metavar='/PATH/TO/DATASET',

) to your local dataset location.

[santoshreddy254]

Very quick response Thank you for that. I will update you if I can do that

[santoshreddy254]

Hi,
I am facing issues in how to load VOC format data, you have bounding boxes in a txt file for each image but for me, I have in XML file which is parsed and got the box but I was not able to incorporate in this code. Any help or suggestion is appreciated.

[coallaoh]

Sorry, we don't have enough manpower to provide guidance on turning XML files into txt files :(

[coallaoh]

Closing the issue, assuming the question was answered :)
Please re-open the issue as necessary.