import io import os os.system("chmod 777 ffmpeg") import torch import gradio as gr import librosa import numpy as np import soundfile import logging from fairseq import checkpoint_utils from my_utils import load_audio from vc_infer_pipeline import VC import traceback from config import Config from infer_pack.models import ( SynthesizerTrnMs256NSFsid, SynthesizerTrnMs256NSFsid_nono, SynthesizerTrnMs768NSFsid, SynthesizerTrnMs768NSFsid_nono, ) from i18n import I18nAuto logging.getLogger("numba").setLevel(logging.WARNING) logging.getLogger("markdown_it").setLevel(logging.WARNING) logging.getLogger("urllib3").setLevel(logging.WARNING) logging.getLogger("matplotlib").setLevel(logging.WARNING) i18n = I18nAuto() i18n.print() config = Config() models, _, _ = checkpoint_utils.load_model_ensemble_and_task( ["hubert_base.pt"], suffix="", ) hubert_model = models[0] hubert_model = hubert_model.to(config.device) hubert_model = hubert_model.float() hubert_model.eval() global n_spk, tgt_sr, net_g, vc, cpt, version person = "weights/simple-guitar-crepe-guolv_e1000.pth" print("loading %s" % person) cpt = torch.load(person, map_location="cpu") tgt_sr = cpt["config"][-1] cpt["config"][-3] = cpt["weight"]["emb_g.weight"].shape[0] # n_spk net_g = SynthesizerTrnMs768NSFsid(*cpt["config"], is_half=False) del net_g.enc_q print(net_g.load_state_dict(cpt["weight"], strict=False)) net_g.eval().to(config.device) net_g = net_g.float() vc = VC(tgt_sr, config) n_spk = cpt["config"][-3] version="v2" default_audio=load_audio("logs/mute/1_16k_wavs/mute.wav",16000) def vc_single( # sid=0, input_audio_path,#待选取 f0_up_key,#待选取 f0_method, file_index="logs/added_IVF2225_Flat_nprobe_1_simple-guitar-crepe-guolv_v2.index",#写死 index_rate=1,#写死1 filter_radius=3,#不需要,随便写,3 resample_sr=0,#写死0不需要 rms_mix_rate=1,#写死1不需要 protect=0.5,#写死0.5不需要 ): global tgt_sr, net_g, vc, hubert_model, version if input_audio_path is None: return "You need to upload an audio", None f0_up_key = int(f0_up_key) try: audio = input_audio_path[1] / 32768.0 if len(audio.shape) == 2: audio = np.mean(audio, -1) audio = librosa.resample(audio, orig_sr=input_audio_path[0], target_sr=16000) audio_max = np.abs(audio).max() / 0.95 if audio_max > 1: audio /= audio_max times = [0, 0, 0] audio_opt = vc.pipeline( model=hubert_model, net_g=net_g, sid=0, audio=audio, input_audio_path="123", times=times, f0_up_key=f0_up_key, f0_method=f0_method, file_index=file_index, index_rate=index_rate, if_f0=1, filter_radius=filter_radius, tgt_sr=tgt_sr, resample_sr=resample_sr, rms_mix_rate=rms_mix_rate, version="v2", protect=protect, f0_file=None, ) if resample_sr >= 16000 and tgt_sr != resample_sr: tgt_sr = resample_sr index_info = ( "Using index:%s." % file_index if os.path.exists(file_index) else "Index not used." ) return "Success.\n %s\nTime:\n npy:%ss, f0:%ss, infer:%ss" % ( index_info, times[0], times[1], times[2], ), (tgt_sr, audio_opt) except: info = traceback.format_exc() print(info) return "报错了!信息如下:%s"%info, (16000, default_audio) app = gr.Blocks() with app: with gr.Tabs(): with gr.TabItem("人声转吉他极简在线demo"): gr.Markdown( value=""" 变调越高吉他音越细,越低越沉闷 """ ) vc_input = gr.Audio(label="上传音频") with gr.Column(): with gr.Row(): vc_transform = gr.Slider( minimum=-12, maximum=12, label="变调(半音数量,升八度12降八度-12)", value=0, step=1, interactive=True, ) f0method = gr.Radio( label=i18n( "选择音高提取算法:语音推荐dio歌声推荐pm" ), choices=["pm", "dio"], value="dio", interactive=True, ) with gr.Row(): but = gr.Button(i18n("转换"), variant="primary") vc_output1 = gr.Textbox(label=i18n("输出信息")) vc_output2 = gr.Audio(label=i18n("输出音频(右下角三个点,点了可以下载)")) but.click( vc_single, [ vc_input, vc_transform, f0method ], [vc_output1, vc_output2], ) app.launch(server_name="0.0.0.0",quiet=True)