Update app.py

app.py

@@ -22,8 +22,8 @@ def vc_fn(sid, input_audio, vc_transform, auto_f0,cluster_ratio, noise_scale):
     sampling_rate, audio = input_audio
     # print(audio.shape,sampling_rate)
     duration = audio.shape[0] / sampling_rate
-    if duration > 100:
-        return "请上传小于100s的音频，需要转换长音频请本地进行转换", None
+    if duration > 90:
+        return "请上传小于90s的音频，需要转换长音频请本地进行转换", None
     audio = (audio / np.iinfo(audio.dtype).max).astype(np.float32)
     if len(audio.shape) > 1:
         audio = librosa.to_mono(audio.transpose(1, 0))
@@ -33,12 +33,8 @@ def vc_fn(sid, input_audio, vc_transform, auto_f0,cluster_ratio, noise_scale):
     out_wav_path = "temp.wav"
     soundfile.write(out_wav_path, audio, 16000, format="wav")
     print( cluster_ratio, auto_f0, noise_scale)
-    out_audio, out_sr = model.infer(sid, vc_transform, out_wav_path,
-                                   cluster_infer_ratio=cluster_ratio,
-                                   auto_predict_f0=auto_f0,
-                                   noice_scale=noise_scale
-                                   )
-    return "转换成功", (44100, out_audio.numpy())
+    _audio = model.slice_inference(out_wav_path, sid, vc_transform, slice_db, cluster_ratio, auto_f0, noise_scale)
+    return "转换成功", (44100, _audio)
 
 
 app = gr.Blocks()
@@ -62,7 +58,7 @@ with app:
                 """)
             spks = list(model.spk2id.keys())
             sid = gr.Dropdown(label="音色", choices=["XT4.0"], value="XT4.0")
-            vc_input3 = gr.Audio(label="上传音频（长度建议小于100秒）")
+            vc_input3 = gr.Audio(label="上传音频（长度建议小于90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)
             cluster_ratio = gr.Number(label="聚类模型混合比例，0-1之间，默认为0不启用聚类，能提升音色相似度，但会导致咬字下降（如果使用建议0.5左右）", value=0)
             auto_f0 = gr.Checkbox(label="自动f0预测，配合聚类模型f0预测效果更好,会导致变调功能失效（仅限转换语音，歌声不要勾选此项会究极跑调）", value=False)