Spaces:
Running
Running
Commit
·
fafb1c4
1
Parent(s):
95c435a
Update app.py
Browse files
app.py
CHANGED
@@ -41,15 +41,15 @@ def create_vc_fn(model, sid):
|
|
41 |
soundfile.write(raw_path, audio, 16000, format="wav")
|
42 |
raw_path.seek(0)
|
43 |
out_audio, out_sr = model.infer(sid, vc_transform, raw_path,
|
44 |
-
auto_predict_f0=auto_f0, F0_mean_pooling=True
|
45 |
)
|
46 |
return "Success", (44100, out_audio.cpu().numpy())
|
47 |
if input_audio is None:
|
48 |
return "You need to upload an audio", None
|
49 |
sampling_rate, audio = input_audio
|
50 |
duration = audio.shape[0] / sampling_rate
|
51 |
-
if duration >
|
52 |
-
return "Please upload an audio file that is less than
|
53 |
audio = (audio / np.iinfo(audio.dtype).max).astype(np.float32)
|
54 |
if len(audio.shape) > 1:
|
55 |
audio = librosa.to_mono(audio.transpose(1, 0))
|
@@ -59,7 +59,7 @@ def create_vc_fn(model, sid):
|
|
59 |
soundfile.write(raw_path, audio, 16000, format="wav")
|
60 |
raw_path.seek(0)
|
61 |
out_audio, out_sr = model.infer(sid, vc_transform, raw_path,
|
62 |
-
auto_predict_f0=auto_f0, F0_mean_pooling=True
|
63 |
)
|
64 |
return "Success", (44100, out_audio.cpu().numpy())
|
65 |
return vc_fn
|
@@ -114,7 +114,7 @@ if __name__ == '__main__':
|
|
114 |
)
|
115 |
with gr.Row():
|
116 |
with gr.Column():
|
117 |
-
vc_input = gr.Audio(label="Input audio"+' (less than
|
118 |
vc_transform = gr.Number(label="vc_transform", value=0)
|
119 |
auto_f0 = gr.Checkbox(label="auto_f0", value=False)
|
120 |
tts_mode = gr.Checkbox(label="tts (use edge-tts as input)", value=False)
|
|
|
41 |
soundfile.write(raw_path, audio, 16000, format="wav")
|
42 |
raw_path.seek(0)
|
43 |
out_audio, out_sr = model.infer(sid, vc_transform, raw_path,
|
44 |
+
auto_predict_f0=auto_f0, F0_mean_pooling=True,
|
45 |
)
|
46 |
return "Success", (44100, out_audio.cpu().numpy())
|
47 |
if input_audio is None:
|
48 |
return "You need to upload an audio", None
|
49 |
sampling_rate, audio = input_audio
|
50 |
duration = audio.shape[0] / sampling_rate
|
51 |
+
if duration > 120 and limitation:
|
52 |
+
return "Please upload an audio file that is less than 120 seconds. If you need to generate a longer audio file, please use Colab.", None
|
53 |
audio = (audio / np.iinfo(audio.dtype).max).astype(np.float32)
|
54 |
if len(audio.shape) > 1:
|
55 |
audio = librosa.to_mono(audio.transpose(1, 0))
|
|
|
59 |
soundfile.write(raw_path, audio, 16000, format="wav")
|
60 |
raw_path.seek(0)
|
61 |
out_audio, out_sr = model.infer(sid, vc_transform, raw_path,
|
62 |
+
auto_predict_f0=auto_f0, F0_mean_pooling=True,
|
63 |
)
|
64 |
return "Success", (44100, out_audio.cpu().numpy())
|
65 |
return vc_fn
|
|
|
114 |
)
|
115 |
with gr.Row():
|
116 |
with gr.Column():
|
117 |
+
vc_input = gr.Audio(label="Input audio"+' (less than 120 seconds)' if limitation else '')
|
118 |
vc_transform = gr.Number(label="vc_transform", value=0)
|
119 |
auto_f0 = gr.Checkbox(label="auto_f0", value=False)
|
120 |
tts_mode = gr.Checkbox(label="tts (use edge-tts as input)", value=False)
|