Both the full and light_512_resized versions of the dataset use backslashes in their synthetic data paths. This causes problems on the Linux operating systems. I can write a script to replace backslashes with forward slashes and PR if you like.
As a temporary solution, Linux users can use str.replace().
E.g. image_name = str(row[self.input_key]) from src/data/data_readers.py on line 32 can be replaced with image_name = str(row[self.input_key]).replace("\\", "/") to prevent errors.
Both the
fullandlight_512_resizedversions of the dataset use backslashes in their synthetic data paths. This causes problems on the Linux operating systems. I can write a script to replace backslashes with forward slashes and PR if you like.As a temporary solution, Linux users can use
str.replace().E.g.
image_name = str(row[self.input_key])fromsrc/data/data_readers.pyon line 32 can be replaced withimage_name = str(row[self.input_key]).replace("\\", "/")to prevent errors.