{ "sample_rate": 24000, "hop": 256, "n_fft": 1024, "n_mels": 80, "f0_min": 50.0, "f0_max": 600.0, "speakers": [ "neutral", "conv", "happy", "sad", "anger", "fear", "surprise", "disgust" ], "vocab_size": 70, "channels": 256, "heads": 4, "feedforward": 1024, "kernel": 7, "encoder_layers": 4, "prosody_layers": 2, "acoustic_layers": 6, "dropout": 0.1, "predictor_channels": 128, "synth_fft": 512, "synth_hop": 128, "noise_fft": 256, "noise_hop": 64, "cep_causal": 80, "cep_anti": 16, "noise_cep": 32, "harmonic_limit": 0.47, "max_tokens": 512, "max_frames": 3000, "mel_mean": -5.0, "mel_scale": 5.0, "energy_mean": -4.0, "energy_scale": 2.0, "pitch_reference_hz": 200.0 }