diff --git a/synthesizer/hparams.py b/synthesizer/hparams.py index f7d38f0..ee4e3ed 100644 --- a/synthesizer/hparams.py +++ b/synthesizer/hparams.py @@ -33,16 +33,16 @@ def parse(self, string): preemphasize = True, ### Tacotron Text-to-Speech (TTS) - tts_embed_dims = 512, # Embedding dimension for the graphemes/phoneme inputs + tts_embed_dims = 256, # Embedding dimension for the graphemes/phoneme inputs tts_encoder_dims = 256, tts_decoder_dims = 128, - tts_postnet_dims = 512, + tts_postnet_dims = 256, tts_encoder_K = 5, - tts_lstm_dims = 1024, + tts_lstm_dims = 512, tts_postnet_K = 5, tts_num_highways = 4, tts_dropout = 0.5, - tts_cleaner_names = ["english_cleaners"], + tts_cleaner_names = ["transliteration_cleaners"], tts_stop_threshold = -3.4, # Value below which audio generation ends. # For example, for a range of [-4, 4], this # will terminate the sequence at the first diff --git a/synthesizer/utils/symbols.py b/synthesizer/utils/symbols.py index 132d3a6..d32d497 100644 --- a/synthesizer/utils/symbols.py +++ b/synthesizer/utils/symbols.py @@ -8,8 +8,7 @@ _pad = "_" _eos = "~" -_characters = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz!\'\"(),-.:;? " - +_characters = "'-aábcdeéfghiíjklmnñoópqrstuúüvwxyz " # Prepend "@" to ARPAbet symbols to ensure uniqueness (some are the same as uppercase letters): #_arpabet = ["@' + s for s in cmudict.valid_symbols]