FOTS.PyTorch
FOTS.PyTorch copied to clipboard
synthtext dataset
is it necessary to use synthtext dataset to train recognition part, or is it possible to do that with icdar2015 also?