Thank you so much for sharing the implementation. I tried to train the model by the original setting of this implementation but the results are not good. Do you use the default settings to train your model or do any data augmentation/cleaning? Moreover, do you do any post-processing after the inference?
Thank you so much for sharing the implementation. I tried to train the model by the original setting of this implementation but the results are not good. Do you use the default settings to train your model or do any data augmentation/cleaning? Moreover, do you do any post-processing after the inference?