init

app.py

@@ -39,6 +39,7 @@ def get_text(text, hps):
     print(text_norm.shape)
     return text_norm
 
+convert_cnt = [0]
 
 hps_ms = utils.get_hparams_from_file("configs/nyarumul.json")
 net_g_ms = SynthesizerTrn(
@@ -62,7 +63,7 @@ def vc_fn(sid, input_audio,vc_transform):
     # print(audio.shape,sampling_rate)
     duration = audio.shape[0] / sampling_rate
     if duration > 45:
-        return "Error: Audio is too long", None
+        return "请上传小于45s的音频，需要转换长音频请使用colab", None
     audio = (audio / np.iinfo(audio.dtype).max).astype(np.float32)
     if len(audio.shape) > 1:
         audio = librosa.to_mono(audio.transpose(1, 0))
@@ -85,7 +86,8 @@ def vc_fn(sid, input_audio,vc_transform):
         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,
                                noise_scale_w=0.1, length_scale=1)[0][0, 0].data.float().numpy()
-
+    convert_cnt[0] += 1
+    print(convert_cnt[0])
     return "Success", (hps_ms.data.sampling_rate, audio)