Spaces:

LeeSangHoon

/

HierSpeech_TTS

like 320

Running on T4

I update an audio loader for multi-channel audio.

Our model only supports a single channel audio so we only use a first channel of input prompt audio.

        audio, sample_rate = torchaudio.load(prompt)

        # support only single channel
        if audio.shape[0] != 1:
            audio = audio[:1,:]