Words Mismatch during preparing LJSpeech dataset

[kin0303]

Hi, I got this warning when prepare LJSpeech Dataset:

2023-03-07 11:27:29,886 WARNING [words_mismatch.py:88] words count mismatch on 100.0% of the lines (1/1)
...
2023-03-07 11:27:29,964 WARNING [words_mismatch.py:88] words count mismatch on 100.0% of the lines (1/1)
2023-03-07 11:27:29,964 WARNING [words_mismatch.py:88] words count mismatch on 100.0% of the lines (1/1)
2023-03-07 11:27:29,965 WARNING [words_mismatch.py:88] words count mismatch on 100.0% of the lines (1/1)
2023-03-07 11:27:29,966 WARNING [words_mismatch.py:88] words count mismatch on 100.0% of the lines (1/1)
100%|█████████████████████████████████████| 200/200 [00:00<00:00, 2466.56it/s]
2023-03-07 11:27:29,999 INFO [tokenizer.py:190] unique phonemes: {'t', 's', 'v', 'ʒ', 'l', 'ɐ', ':', 'o', '?', '"', 'ʊ', 'ʌ', 'ʔ', 'r', 'ɹ', ';', '̩', 'ɚ', '!', 'j', 'ɪ', 'ŋ', 'b', ',', 'f', 'z', 'k', 'ɾ', 'ə', 'ʃ', 'ᵻ', 'ː', 'θ', 'x', 'u', 'e', 'm', 'ð', 'a', 'ɑ', 'i', 'p', 'ɛ', '.', 'w', 'ɜ', 'ɡ', 'h', 'd', 'æ', '_', 'ɔ', 'n'}

How to fix that? If I do the training, have I got the problem later? And what is the minimum vram used for training with LJSpeech?

[RasterXomotron]

I get these warnings too. A bit concerning.

[lifeiteng]

@RasterXomotron @kin0303 more info in #5

I don't think this issue will have a big impact, I appreciate that If you guys can fix them.