Update code

app.py

@@ -35,7 +35,7 @@ def inference(wav,instruction):
       text, *_ = nbests[0]
       instruction=instruction.split(" <|")[0]
       # import pdb;pdb.set_trace()
-      text=text.replace(instruction,"").replace("_STOP","").split(".")[-1]
+      text=text.replace(instruction,"").replace("<|audio|>","").replace("_STOP","").split(".")[-1]
       # if lang == "chinese":
       #     wav = text2speechch(text)["wav"]
       #     scipy.io.wavfile.write("out.wav",text2speechch.fs , wav.view(-1).cpu().numpy())