Hi, thank you for releasing the code and for your great work.
I have a small question regarding the validation procedure. I noticed that during training, the validation phase still appears to use the ground-truth text for feature reordering.
Could you please clarify whether this is the intended setting? More specifically, should validation use GT text for reordering, or should it follow the same setting as inference, where GT text is not available?
Thank you very much for your time and clarification.
Hi, thank you for releasing the code and for your great work.
I have a small question regarding the validation procedure. I noticed that during training, the validation phase still appears to use the ground-truth text for feature reordering.
Could you please clarify whether this is the intended setting? More specifically, should validation use GT text for reordering, or should it follow the same setting as inference, where GT text is not available?
Thank you very much for your time and clarification.