Update app.py

app.py

@@ -168,7 +168,7 @@ def infer_batch(ref_audio, ref_text, gen_text_batches, exp_name, remove_silence,
         # Prepare the text
         # if len(ref_text[-1].encode('utf-8')) == 1:
              #ref_text = ref_text + " "
-         #    gen_text = " " + gen_text
+        gen_text = " " + gen_text
         text_list = [ref_text + gen_text]
         final_text_list = convert_char_to_pinyin(text_list)
 
@@ -186,10 +186,10 @@ def infer_batch(ref_audio, ref_text, gen_text_batches, exp_name, remove_silence,
         ref_text_len = len(ref_text.encode('utf-8')) + 3 * len(re.findall(zh_pause_punc, ref_text))
         gen_text_len = len(gen_text.encode('utf-8')) + 3 * len(re.findall(zh_pause_punc, gen_text))
         if len(ref_text.encode('utf-8')) >= 3:
-            duration = min(5000, max(300, (ref_audio_len + int(ref_audio_len / ref_text_len * gen_text_len / speed))))
+            duration = min(2000, max(300, (ref_audio_len + int(ref_audio_len / ref_text_len * gen_text_len / speed))))
             print(f"Duration: {duration} seconds")
         else:
-            duration = min(5000, max(300, int(133 * gen_text_len / (speed * 10))))
+            duration = min(2000, max(300, int(150 * gen_text_len / (speed * 10))))
             print(f"Duration: {duration} seconds")
 
         # inference