3a181c8d229d3f45d6457cd329d2336b07b2330b,autokeras/pretrained/voice_generator/voice_generator.py,VoiceGenerator,__init__,#VoiceGenerator#Any#Any#,244

Before Change

        self.hop_length = 0
        self.overwrite = overwrite
        self.device = get_device()
        self.load()

    def load(self):
        self._maybe_download()
        self.sample_rate = Hparams.sample_rate

After Change

class Hparams:
    name = "deepvoice3"

    // Text:
    // [en  jp]
    frontend = "en"

    // Replace words to its pronunciation with fixed probability.
    // e.g.  "hello" to "HH AH0 L OW1"
    // [en  jp]
    // en: Word -> pronunciation using CMUDict
    // jp: Word -> pronounciation usnig MeCab
    // [0 ~ 1.0]: 0 means no replacement happens.
    replace_pronunciation_prob = 0.5

    // Convenient model builder
    // Definitions can be found at deepvoice3_pytorch/builder.py
    // deepvoice3: DeepVoice3 https://arxiv.org/abs/1710.07654
    builder = "deepvoice3"

Instances