Open iamanigeeit opened 1 year ago
Hi @iamanigeeit, thank you so much for your installation steps, it helps a lot! but I haven't set it up, so I may need to bother you with some questions. After I use these
two lines to install mfa
conda config --add channels conda-forge
conda install montreal-forced-aligner
I can't run mfa thirdparty download
directly, it will have "thirdparty command not exist" error. May I know the version of kaldi pynini mfa
you installed?
@Rongjiehuang I got this repo to work, but i had to correct some things. Hope it helps someone else.
sudo apt install gfortran libopenblas-base
. These are required but not specified.environment.yaml
to remove duplicates in scipy and numpy, and remove version requirements on scipy and numba (old vresions cause conflicts with numpy).environment.yaml
withpip uninstall nvidia_cublas_cu11
(or whatever version you have).modules/GenerSpeech/config/generspeech.yaml
, changeemotion_encoder_path
tocheckpoints/Emotion_encoder.pt
sys.path
, either by moving GenerSpeech.py to the GenerSpeech dir or adding these lines at the top of GenerSpeech (otherwise Python can't find the imports)mfa thirdparty download
utils.hparams.py
, lines 29 and 32 should removehelp='location of the data corpus'
becuase it's misleading. Line 41 needs to includeremove=False
.data_gen_utils
line 299 if there is a word missing frommfa_dict.txt
, because the TextGrid will skip the phones of the missing word. Actually, some common words are not in the dictionary like "her" (HH_ER1) and "processing" (P_R_AA1_S_EH0_S_IH0_NG). You have add them to the dictionary yourself. The correct way is to runmfa validate
and append tomfa_dict.txt
first (see this script),praatio
as the standard TextGrid parser.