When running mfa align . english_us_arpa english_us_arpa . I keep getting a message saying alignment already done, skipping. So I tried to run mfa align --clean instead but it deletes the pretrained english_us_arpa acoustic model and dictionary as well. If I copy a subset of the wav files that I want to align again to a different directory, it generates TextGrids for all the files in the previous directory and still doesn't update the alignment. What is the correct way to update some TextGrids to remove occurrences of spn labels?
When running mfa align . english_us_arpa english_us_arpa . I keep getting a message saying alignment already done, skipping. So I tried to run mfa align --clean instead but it deletes the pretrained english_us_arpa acoustic model and dictionary as well. If I copy a subset of the wav files that I want to align again to a different directory, it generates TextGrids for all the files in the previous directory and still doesn't update the alignment. What is the correct way to update some TextGrids to remove occurrences of spn labels?