| import soundfile | |
| from infer_pack.onnx_inference import OnnxRVC | |
| hop_size = 512 | |
| sampling_rate = 40000 # Sampling rate | |
| f0_up_key = 0 # rising and falling tones | |
| sid = 0 # Role ID | |
| f0_method = "dio" # F0 extraction algorithm | |
| model_path = "ShirohaRVC.onnx" # full path to model | |
| vec_name = "vec-256-layer-9" # The internal automatic completion is f"pretrained/{vec_name}.onnx" requires the onnx vec model | |
| wav_path = "123.wav" # Input path or ByteIO instance | |
| out_path = "out.wav" # Output path or ByteIO instance | |
| model = OnnxRVC( | |
| model_path, vec_path=vec_name, sr=sampling_rate, hop_size=hop_size, device="cuda" | |
| ) | |
| audio = model.inference(wav_path, sid, f0_method=f0_method, f0_up_key=f0_up_key) | |
| soundfile.write(out_path, audio, sampling_rate) | |