init

app.py

@@ -56,7 +56,7 @@ hubert = torch.hub.load("bshall/hubert:main", "hubert_soft")
 
 _ = utils.load_checkpoint("nyarumodel.pth", net_g_ms, None)
 
-def vc_fn(sid, input_audio,vc_transform):
+def vc_fn(sid,random1, input_audio,vc_transform):
     if input_audio is None:
         return "You need to upload an audio", None
     sampling_rate, audio = input_audio
@@ -84,7 +84,7 @@ def vc_fn(sid, input_audio,vc_transform):
     with torch.no_grad():
         x_tst = stn_tst.unsqueeze(0)
         x_tst_lengths = torch.LongTensor([stn_tst.size(0)])
-        audio = net_g_ms.infer(x_tst, x_tst_lengths, pitch=pitch,sid=sid, noise_scale=0.4,
+        audio = net_g_ms.infer(x_tst, x_tst_lengths, pitch=pitch,sid=sid, noise_scale=float(random1),
                                noise_scale_w=0.1, length_scale=1)[0][0, 0].data.float().numpy()
     convert_cnt[0] += 1
     print(convert_cnt[0])
@@ -109,9 +109,10 @@ with app:
             sid = gr.Dropdown(label="音色",choices=['猫雷'], value="猫雷")
             vc_input3 = gr.Audio(label="上传音频（长度小于45秒）")
             vc_transform = gr.Number(label="变调（整数，可以正负，半音数量，升高八度就是12）",value=0)
+            random1 = gr.Number(label="随机化程度，似乎会影响音质，建议保持默认",value=0.4)
             vc_submit = gr.Button("转换", variant="primary")
             vc_output1 = gr.Textbox(label="Output Message")
             vc_output2 = gr.Audio(label="Output Audio")
-        vc_submit.click(vc_fn, [sid,  vc_input3, vc_transform], [vc_output1, vc_output2])
+        vc_submit.click(vc_fn, [sid,random1,  vc_input3, vc_transform], [vc_output1, vc_output2])
 
     app.launch()

nyarumodel.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a14d75537894f6726e5b89e62a679986a0ef8bee52827f656fc57c333f668887
-size 256011087
+oid sha256:c489836c6519f7a3d1b98673a241a29a516834a642ca31c2c1bd9235c237b5c3
+size 256010117