tryolabs / luminoth

Deep Learning toolkit for Computer Vision.
https://tryolabs.com
BSD 3-Clause "New" or "Revised" License
2.4k stars 401 forks source link

Issue with "--only-classes" option for lumi data set transform #290

Open huangka97 opened 4 years ago

huangka97 commented 4 years ago

The "--only-classes" option is not filtering out the classes I desire. Currently I have the following:

lumi dataset transform --type pascal --data-dir datasets/pascal/2007/VOCdevkit/VOC2007/ --output-dir datasets/pascal/tf/2007/ --split train --only-classes=car

However, all classes are being read according to what is being outputted to my screen:

I0814 18:33:38.137515 140095359665984 object_detection_writer.py:95] Saved 2501 records to "datasets/pascal/tf/2007/train.tfrecords"
I0814 18:33:38.137785 140095359665984 transform.py:65] Composition per class (train):
I0814 18:33:38.137905 140095359665984 transform.py:68]  person: 2705
I0814 18:33:38.137994 140095359665984 transform.py:68]  car: 826
I0814 18:33:38.138073 140095359665984 transform.py:68]  chair: 726
I0814 18:33:38.138149 140095359665984 transform.py:68]  bottle: 338
I0814 18:33:38.138223 140095359665984 transform.py:68]  pottedplant: 305
I0814 18:33:38.138297 140095359665984 transform.py:68]  bird: 294
I0814 18:33:38.138370 140095359665984 transform.py:68]  dog: 271
I0814 18:33:38.138443 140095359665984 transform.py:68]  sofa: 218
I0814 18:33:38.138517 140095359665984 transform.py:68]  boat: 208
I0814 18:33:38.138589 140095359665984 transform.py:68]  horse: 207
I0814 18:33:38.138692 140095359665984 transform.py:68]  bicycle: 202
I0814 18:33:38.138776 140095359665984 transform.py:68]  motorbike: 193
I0814 18:33:38.138851 140095359665984 transform.py:68]  cat: 191
I0814 18:33:38.138925 140095359665984 transform.py:68]  tvmonitor: 191
I0814 18:33:38.138999 140095359665984 transform.py:68]  sheep: 191
I0814 18:33:38.139070 140095359665984 transform.py:68]  cow: 185
I0814 18:33:38.139151 140095359665984 transform.py:68]  train: 158
I0814 18:33:38.139223 140095359665984 transform.py:68]  aeroplane: 156
I0814 18:33:38.139295 140095359665984 transform.py:68]  diningtable: 148
I0814 18:33:38.139367 140095359665984 transform.py:68]  bus: 131