The OCR approach is rephrased as Segmentation Transformer: https://arxiv.org/abs/1909.11065. This is an official implementation of semantic segmentation for HRNet. https://arxiv.org/abs/1908.07919
Other
3.09k
stars
682
forks
source link
Training my own data(6 classes) as cityscapes-format, but got eval [0. 0.00423967 0.35140248 0. 0. 0. ] #263
I made my own data as Cityscapes Dataset format, but my dataset only has 6 classes labeled 0-5.
I change the "self.label_mapping" in lib/datasets/cityscapes.py:
self.label_mapping = {0: 0, 1: 1, 2: 2, 3: 3, 4: 4, 5: 5}
But when I was training, the eval on val is:
[0. 0.00423967 0.35140248 0. 0. 0. ]
Could you tell me how to fix this problem? Was there anything I forgot to modify?
I made my own data as Cityscapes Dataset format, but my dataset only has 6 classes labeled 0-5. I change the "self.label_mapping" in lib/datasets/cityscapes.py: self.label_mapping = {0: 0, 1: 1, 2: 2, 3: 3, 4: 4, 5: 5} But when I was training, the eval on val is: [0. 0.00423967 0.35140248 0. 0. 0. ]
Could you tell me how to fix this problem? Was there anything I forgot to modify?