Update app.py

app.py

@@ -148,34 +148,34 @@ def update_dropdown(new_choices):
     spks = list(model.spk2id.keys())
     new_choices = gr.Dropdown.update(choices=spks) 
     return new_choices
-
-app = gr.Blocks()
-with app:
-    with gr.Tabs():
-        with gr.TabItem(" "):  
+if __name__ == "__main__":
+    app = gr.Blocks()
+    with app:
+        with gr.Tabs():
+            with gr.TabItem(" "):  
         
 
-            gr.Markdown(value=base64.b64decode( b'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').decode())
-
-            with gr.Tabs():
-                with gr.TabItem("单个音频上传"):
-                    vc_input3 = gr.Audio(label="上传音频<280s无BGM无和声的干声", type="filepath", source="upload",value="examples/1.mp3")
-
-                with gr.TabItem("文字转语音（实验性）"):
-                    gr.Markdown(value="文字转语音（TTS）说明：使用edge_tts服务生成音频，并转换为So-VITS模型音色。")
-                    auto_f0 = gr.Checkbox(label="自动f0预测，配合聚类模型f0预测效果更好（仅限转换语音，歌声不要勾选此项会究极跑调）", value=False)
-                    with gr.Row():
-                        text_input = gr.Textbox(label = "在此输入需要转译的文字（建议打开自动f0预测）限定200字以内,建议f0预测器选dio")#, lines=4
-                    with gr.Row():
-                        tts_gender = gr.Radio(label = "说话人性别", choices = ["男","女"], value = "女")
-                        tts_lang = gr.Dropdown(label = "选择语言，Auto为根据输入文字自动识别", choices=SUPPORTED_LANGUAGES, value = "Auto")
-                    with gr.Row():
-                        tts_rate = gr.Slider(label = "TTS语音变速（倍速相对值）", minimum = -1, maximum = 3, value = 0, step = 0.1)
-                        tts_volume = gr.Slider(label = "TTS语音音量（相对值）", minimum = -1, maximum = 1.5, value = 0, step = 0.1)
-                    vc_tts_submit = gr.Button("文本转语音", variant="primary")
-            spks = list(model.spk2id.keys())
+                gr.Markdown(value=base64.b64decode( b'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').decode())
+
+                with gr.Tabs():
+                    with gr.TabItem("单个音频上传"):
+                        vc_input3 = gr.Audio(label="上传音频<280s无BGM无和声的干声", type="filepath", source="upload",value="examples/1.mp3")
+
+                    with gr.TabItem("文字转语音（实验性）"):
+                        gr.Markdown(value="文字转语音（TTS）说明：使用edge_tts服务生成音频，并转换为So-VITS模型音色。")
+                        auto_f0 = gr.Checkbox(label="自动f0预测，配合聚类模型f0预测效果更好（仅限转换语音，歌声不要勾选此项会究极跑调）", value=False)
+                        with gr.Row():
+                            text_input = gr.Textbox(label = "在此输入需要转译的文字（建议打开自动f0预测）限定200字以内,建议f0预测器选dio")#, lines=4
+                        with gr.Row():
+                            tts_gender = gr.Radio(label = "说话人性别", choices = ["男","女"], value = "女")
+                            tts_lang = gr.Dropdown(label = "选择语言，Auto为根据输入文字自动识别", choices=SUPPORTED_LANGUAGES, value = "Auto")
+                        with gr.Row():
+                            tts_rate = gr.Slider(label = "TTS语音变速（倍速相对值）", minimum = -1, maximum = 3, value = 0, step = 0.1)
+                            tts_volume = gr.Slider(label = "TTS语音音量（相对值）", minimum = -1, maximum = 1.5, value = 0, step = 0.1)
+                        vc_tts_submit = gr.Button("文本转语音", variant="primary")
+                spks = list(model.spk2id.keys())
   
-            sid = gr.Dropdown(label="音色", choices=spks, value=base64.b64decode( b'SE9TSElOT19NSU5BTUk=' ).decode())
+                sid = gr.Dropdown(label="音色", choices=spks, value=base64.b64decode( b'SE9TSElOT19NSU5BTUk=' ).decode())
             #sid.change(fn=update_dropdown,inputs=[sid],outputs=[sid])
             #sid.update(interactive=True)
             #with gr.Accordion(label="↓切换模型(音色具有抽奖性质，可切换尝试)", open=False):
@@ -183,21 +183,21 @@ with app:
                 #btnMod = gr.Button("载入模型")
                 #statusa = gr.TextArea()
                 #btnMod.click(loadmodel, inputs=[modelstrs], outputs = [statusa,sid])
-            with gr.Row():
-                 slice_db = gr.Slider(label="切片阈值(较嘈杂时-30，保留呼吸声时-50)",maximum=-30, minimum=-70, step=1, value=-40)                              
-                 vc_transform = gr.Slider(label="变调（整数，可以正负，半音数量，升高八度就是12）",maximum=16, minimum=-16, step=1, value=0)                       
-            f0_predictor = gr.Radio(label=base64.b64decode( b'ZjDpooTmtYvlmago5aaC6YGH5ZOR6Z+z5Y+v5Lul5bCd6K+V5pu05o2iZjAp5Yet5bmy5aOw5bmy5YeA56iL5bqm6YCJ5oup44CC5o6o6I2QZmNwZeWSjHJtdnBl' ).decode(), choices=["pm","dio","harvest","fcpe","rmvpe"], value="fcpe")
-            with gr.Row():    
-                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)#聚
-                output_format = gr.Radio(label=base64.b64decode( b'6Z+z6aKR6L6T5Ye65qC85byPKE1QM+S8muWvvOiHtOaXtumXtOi9tOWkmjI3bXMs6ZyA5ZCI5oiQ6K+36YCJZmxhYyk=' ).decode(), choices=["flac", "mp3"], value = "mp3")#格式
-            vc_submit = gr.Button("音频转换", variant="primary")
-
-            vc_output1 = gr.Textbox(label=base64.b64decode( b'6Z+z6auY5bmz5Z2H5YGP5beu5Y2K6Z+z5pWw6YeP77yM5L2T546w6L2s5o2i6Z+z6aKR55qE6LeR6LCD5oOF5Ya177yI5LiA6Iis5bCP5LqOMC4177yJ' ).decode())
-            vc_output2 = gr.Audio(label="Output Audio")
-
-        vc_submit.click(vc_fn, [output_format,sid, vc_input3, vc_transform,auto_f0,cluster_ratio, slice_db,f0_predictor], [vc_output1, vc_output2])
-        vc_tts_submit.click(tts_fn, [text_input, tts_gender, tts_lang, tts_rate, tts_volume, sid, vc_transform,auto_f0,cluster_ratio, slice_db, f0_predictor], [vc_output1, vc_output2])
-app.launch(share=False)
+                with gr.Row():
+                     slice_db = gr.Slider(label="切片阈值(较嘈杂时-30，保留呼吸声时-50)",maximum=-30, minimum=-70, step=1, value=-40)                              
+                     vc_transform = gr.Slider(label="变调（整数，可以正负，半音数量，升高八度就是12）",maximum=16, minimum=-16, step=1, value=0)                       
+                f0_predictor = gr.Radio(label=base64.b64decode( b'ZjDpooTmtYvlmago5aaC6YGH5ZOR6Z+z5Y+v5Lul5bCd6K+V5pu05o2iZjAp5Yet5bmy5aOw5bmy5YeA56iL5bqm6YCJ5oup44CC5o6o6I2QZmNwZeWSjHJtdnBl' ).decode(), choices=["pm","dio","harvest","fcpe","rmvpe"], value="fcpe")
+                with gr.Row():    
+                    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)#聚
+                    output_format = gr.Radio(label=base64.b64decode( b'6Z+z6aKR6L6T5Ye65qC85byPKE1QM+S8muWvvOiHtOaXtumXtOi9tOWkmjI3bXMs6ZyA5ZCI5oiQ6K+36YCJZmxhYyk=' ).decode(), choices=["flac", "mp3"], value = "mp3")#格式
+                vc_submit = gr.Button("音频转换", variant="primary")
+
+                vc_output1 = gr.Textbox(label=base64.b64decode( b'6Z+z6auY5bmz5Z2H5YGP5beu5Y2K6Z+z5pWw6YeP77yM5L2T546w6L2s5o2i6Z+z6aKR55qE6LeR6LCD5oOF5Ya177yI5LiA6Iis5bCP5LqOMC4177yJ' ).decode())
+                vc_output2 = gr.Audio(label="Output Audio")
+
+            vc_submit.click(vc_fn, [output_format,sid, vc_input3, vc_transform,auto_f0,cluster_ratio, slice_db,f0_predictor], [vc_output1, vc_output2])
+            vc_tts_submit.click(tts_fn, [text_input, tts_gender, tts_lang, tts_rate, tts_volume, sid, vc_transform,auto_f0,cluster_ratio, slice_db, f0_predictor], [vc_output1, vc_output2])
+    app.launch(share=False)