mirror of
https://github.com/fumiama/Retrieval-based-Voice-Conversion-WebUI.git
synced 2026-06-05 01:10:22 +08:00
a8783c6639
* optimzie(train&uvr5): rm sf & simp. AudioPre * fix(audio): too many mallocs * feat(audio): load_audio support stereo * fix(audio): float32 wav saving * fix(train): missing ckpt var
23 lines
705 B
Python
23 lines
705 B
Python
import librosa
|
|
|
|
from rvc.onnx import RVC
|
|
|
|
from infer.lib.audio import save_audio
|
|
|
|
hop_size = 512
|
|
sampling_rate = 40000 # 采样率
|
|
f0_up_key = 0 # 升降调
|
|
sid = 0 # 角色ID
|
|
f0_method = "dio" # F0提取算法
|
|
model_path = "exported_model.onnx" # 模型的完整路径
|
|
vec_path = "vec-256-layer-9.onnx" # 需要onnx的vec模型
|
|
wav_path = "123.wav" # 输入路径或ByteIO实例
|
|
out_path = "out.wav" # 输出路径或ByteIO实例
|
|
|
|
model = RVC(model_path, vec_path=vec_path, hop_len=hop_size, device="cuda")
|
|
|
|
wav, sr = librosa.load(wav_path, sr=sampling_rate)
|
|
|
|
audio = model.infer(wav, sr, sampling_rate, sid, f0_method, f0_up_key)
|
|
|
|
save_audio(out_path, audio, sampling_rate) |