The Medical Detection Toolkit contains 2D + 3D implementations of prevalent object detectors such as Mask R-CNN, Retina Net, Retina U-Net, as well as a training and inference framework focused on dealing with medical images.
Apache License 2.0
1.31k
stars
297
forks
source link
Retina U-Net does not work in the toy experiment #87
When I try to use the Retina U-Net in the toy experiment(just changed self.model = 'retina_unet' and the path to the data) the program crashes in line 447 of the retina_unet.py due to the var_seg[:, 0] containing the values [0, 1, 2] while the size of the seg_logits is just [20, 2, 320, 320] and not [20, 3, 320, 320]
When I try to use the Retina U-Net in the toy experiment(just changed self.model = 'retina_unet' and the path to the data) the program crashes in line 447 of the retina_unet.py due to the
var_seg[:, 0]
containing the values[0, 1, 2]
while the size of theseg_logits
is just[20, 2, 320, 320]
and not[20, 3, 320, 320]