Handling of OOVs

[enhuiz]

I have encountered some OOVs when building the alignment for the reference audio. It's probably due to the inconsistent lexicon used by the ASR and aligner, and it will trigger the following assertion:

GenerSpeech/data_gen/tts/data_gen_utils.py

Line 298 in 2979dd2

assert tg_len == ph_len, (tg_len, ph_len, tg_align, ph_list, tg_fn)

Do you know if there is any workaround for this? Many thanks.

[Rongjiehuang]

Hi, it's likely due to the mismatching of phoneme dict, please load our provided dictionary from OneDrive.