Closed mmmmmore closed 3 years ago
Maybe some fonts do not support some characters.
Since the performance is still good enough, I haven't dealt with them yet. I will correct this part in the future.
For now, you can just use the dataset as it is. The proportion should be very low.
Some characters seem to be unsupported by the fonts files and the 10 pixel rule doesn't seem to apply.
example file:
english/sub_101/labels/444.json
Some unknown characters are rendered on the top left corner ( under Toun ) in the following image.
May I know if there's a good way to solve it ?
Thanks !