Error when training new model

[Amy Cox]

Hello,  

 

We are currently trying to create an acoustic model using our text grids. We have been receiving an error message during the mono training section saying  

 

Traceback (most recent call last): 

  File "/opt/miniconda3/envs/aligner/bin/mfa", line 8, in <module> 

    sys.exit(main()) 

  File "/opt/miniconda3/envs/aligner/lib/python3.8/site-packages/montreal_forced_aligner/command_line/mfa.py", line 383, in main 

    run_train_corpus(args, unknown) 

  File "/opt/miniconda3/envs/aligner/lib/python3.8/site-packages/montreal_forced_aligner/command_line/train_and_align.py", line 158, in run_train_corpus 

    align_corpus(args, unknown_args) 

  File "/opt/miniconda3/envs/aligner/lib/python3.8/site-packages/montreal_forced_aligner/command_line/train_and_align.py", line 111, in align_corpus 

    a.train() 

  File "/opt/miniconda3/envs/aligner/lib/python3.8/site-packages/montreal_forced_aligner/aligner/trainable.py", line 72, in train 

    trainer.init_training(identifier, self.temp_directory, self.corpus, self.dictionary, previous) 

  File "/opt/miniconda3/envs/aligner/lib/python3.8/site-packages/montreal_forced_aligner/trainers/monophone.py", line 65, in init_training 

    self._setup_for_init(identifier, temporary_directory, corpus, dictionary) 

  File "/opt/miniconda3/envs/aligner/lib/python3.8/site-packages/montreal_forced_aligner/trainers/base.py", line 175, in _setup_for_init 

    self.data_directory = corpus.subset_directory(self.subset, self.feature_config) 

  File "/opt/miniconda3/envs/aligner/lib/python3.8/site-packages/montreal_forced_aligner/corpus/align_corpus.py", line 748, in subset_directory 

    self.create_subset(subset, feature_config) 

  File "/opt/miniconda3/envs/aligner/lib/python3.8/site-packages/montreal_forced_aligner/corpus/align_corpus.py", line 823, in create_subset 

    subset_utts = set(random.sample(larger_subset, subset)) 

  File "/opt/miniconda3/envs/aligner/lib/python3.8/random.py", line 363, in sample 

    raise ValueError("Sample larger than population or is negative") 

ValueError: Sample larger than population or is negative 

 

Currently, we are running version 2.0.0a24 of the Montreal Forced Aligner on macOS version 11.6.2. 

 

Any help would be appreciated,  

 

Amy C