You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
in the dataset.py source file , after if not exist_positive: code line , why do you restrict the xind, yind to be 0-12 but not 0-25 or 0-51 ?
```
xind = np.clip(xind, 0, self.train_output_sizes[i] - 1)
yind = np.clip(yind, 0, self.train_output_sizes[i] - 1)
# self.train_output_sizes[i] - 1 =12 all the time here why ?????????
# This will mitigate errors generated when the location computed by this is more the grid cell location.
# e.g. For 52x52 grid cells possible values of xind and yind are in range [0-51] including both.
# But sometimes the coomputation makes it 52 and then it will try to find that location in label array
# which is not present and throws error during training.
The text was updated successfully, but these errors were encountered:
in the dataset.py source file , after
if not exist_positive:
code line , why do you restrict the xind, yind to be 0-12 but not 0-25 or 0-51 ?The text was updated successfully, but these errors were encountered: