Upload app.py

app.py

@@ -149,8 +149,10 @@ def infer(text, sdp_ratio, noise_scale, noise_scale_w, length_scale, sid, langua
 
 
 def tts_fn(
-    text, speaker, sdp_ratio, noise_scale, noise_scale_w, length_scale,LongSentence
+    text, speaker, sdp_ratio, noise_scale, noise_scale_w, length_scale,LongSentence,correct_language
 ):
+    if correct_language  = "Auto":
+        correct_language = "JP" if is_japanese(text) else "ZH"
     if not LongSentence:
         with torch.no_grad():
             audio = infer(
@@ -160,7 +162,7 @@ def tts_fn(
                 noise_scale_w=noise_scale_w,
                 length_scale=length_scale,
                 sid=speaker,
-                language= "JP" if is_japanese(text) else "ZH",
+                language= correct_language,
             )
             torch.cuda.empty_cache()
             
@@ -184,7 +186,7 @@ def tts_fn(
                     noise_scale_w=noise_scale_w,
                     length_scale=length_scale,
                     sid=speaker,
-                    language= "JP" if is_japanese(text) else "ZH",
+                    language= correct_language,
                 )
             audio_fin.append(audio)
             
@@ -253,10 +255,9 @@ if __name__ == "__main__":
     _ = net_g.eval()
 
     _ = utils.load_checkpoint(args.model, net_g, None, skip_optimizer=True)
-
     speaker_ids = hps.data.spk2id
     speakers = list(speaker_ids.keys())
-    languages = ["ZH", "JP"]
+    languages = ["Auto","ZH", "JP"]
     with gr.Blocks() as app:
     
         gr.Markdown(
@@ -304,6 +305,9 @@ if __name__ == "__main__":
                                     speaker = gr.Dropdown(
                                         choices=speakers, value=name, label="说话人(在这选择说话人将保留输入文本)"
                                     )
+                                    correct_language = gr.Dropdown(
+                                        choices=languages, value=languages[0], label="Language(如果发生报错请手动指定)"
+                                    )
                     btn.click(
                         tts_fn,
                         inputs=[
@@ -314,6 +318,7 @@ if __name__ == "__main__":
                             noise_scale_w,
                             length_scale,
                             LongSentence,
+                            correct_language,
                         ],
                         outputs=[audio_output],
                     )