fumiama/Retrieval-based-Voice-Conversion-WebUI
1
0
Fork 0
mirror of https://github.com/fumiama/Retrieval-based-Voice-Conversion-WebUI.git synced 2026-06-10 13:00:36 +08:00
Code Activity

fix: MacOS 纯 CPU 推理时 Segmentation fault: 11

Browse Source 
see: facebookresearch/faiss#2317 facebookresearch#2410
This commit is contained in:
源文雨
2023-04-10 18:28:39 +08:00
parent 6c7c1d933f
commit ff1a711cad
5 changed files with 39 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
vc_infer_pipeline.py
View File

@@ -72,6 +72,7 @@ class VC(object):
"output_layer": 9, # layer 9
}
t0 = ttime()
print("vc npy start time:", t0)
with torch.no_grad():
logits = model.extract_features(**inputs)
feats = model.final_proj(logits[0])
@@ -79,13 +80,14 @@ class VC(object):
if(isinstance(index,type(None))==False and isinstance(big_npy,type(None))==False and index_rate!=0):
npy = feats[0].cpu().numpy()
if(self.is_half==True):npy=npy.astype("float32")
D, I = index.search(npy, 1)
_, I = index.search(npy, 1)
npy=big_npy[I.squeeze()]
if(self.is_half==True):npy=npy.astype("float16")
feats = torch.from_numpy(npy).unsqueeze(0).to(self.device)*index_rate + (1-index_rate)*feats
feats = F.interpolate(feats.permute(0, 2, 1), scale_factor=2).permute(0, 2, 1)
t1 = ttime()
print("vc infer start time:", t1)
p_len = audio0.shape[0]//self.window
if(feats.shape[1]<p_len):
p_len=feats.shape[1]
@@ -99,8 +101,9 @@ class VC(object):
else:
audio1 = (net_g.infer(feats, p_len, sid)[0][0, 0] * 32768).data.cpu().float().numpy().astype(np.int16)
del feats,p_len,padding_mask
torch.cuda.empty_cache()
if torch.cuda.is_available(): torch.cuda.empty_cache()
t2 = ttime()
print("vc infer end time:", t2)
times[0] += (t1 - t0)
times[2] += (t2 - t1)
return audio1
@@ -125,6 +128,7 @@ class VC(object):
audio_opt=[]
t=None
t1=ttime()
print("f0 start time:", t1)
audio_pad = np.pad(audio, (self.t_pad, self.t_pad), mode='reflect')
p_len=audio_pad.shape[0]//self.window
inp_f0=None
@@ -146,6 +150,7 @@ class VC(object):
pitch = torch.tensor(pitch,device=self.device).unsqueeze(0).long()
pitchf = torch.tensor(pitchf,device=self.device).unsqueeze(0).float()
t2=ttime()
print("f0 end time:", t2)
times[1] += (t2 - t1)
for t in opt_ts:
t=t//self.window*self.window
@@ -160,5 +165,5 @@ class VC(object):
audio_opt.append(self.vc(model,net_g,sid,audio_pad[t:],None,None,times,index,big_npy,index_rate)[self.t_pad_tgt:-self.t_pad_tgt])
audio_opt=np.concatenate(audio_opt)
del pitch,pitchf,sid
torch.cuda.empty_cache()
if torch.cuda.is_available(): torch.cuda.empty_cache()
return audio_opt