Update app.py

app.py

@@ -6,6 +6,7 @@ from transformers import GPT2Tokenizer, GPT2LMHeadModel
 from itertools import chain
 
 import os 
+import librosa 
 
 import tempfile
 from typing import Optional
@@ -57,7 +58,11 @@ def stt_record(audio_record_buffer):
     acoustic_model = Model(hf_hub_download(repo_id = REPO_ID, filename = "english/model.tflite"))
     scorer_path =  hf_hub_download(repo_id = REPO_ID, filename = "english/huge-vocabulary.scorer")
 
-    _, y = reformat_freq(*audio_record_buffer)  
+    if type(audio_record_buffer)!=tuple:
+        y, sr = librosa.load(audio_record_buffer)
+        _, y = reformat_freq(sr,y)  
+    else:
+        _, y = reformat_freq(*audio_record_buffer)  
     scorer = True # use scorer
     if scorer:
         acoustic_model.enableExternalScorer(scorer_path)
@@ -288,7 +293,7 @@ history =   {"character": "None", "message_history" : [] }
 
 interface_full = gr.Interface(fn=greet_stt_to_tts,
                         inputs=[gr.Dropdown(personality_choices),
-                                gr.Audio(source="microphone", type="numpy", label="Record Audio") ,
+                                gr.Audio(source="microphone", type="filepath", label="Record Audio") ,
                                 "state"], 
                         outputs=["html","state",gr.Audio(type="filepath")],      
                         css=css, title="Chat with Your Voice", description=description,article=article ,