Update app.py

app.py

@@ -34,7 +34,7 @@ if 'processor' not in locals():
 #text = st.text_area('entre com algum texto')
 #st.title("Wav a ser transcrito ")
 
-wav_up = st.file_uploader("Upload",type=['wav'])
+wav_up = st.file_uploader("Upload",type=['wav','ogg','mp3'])
 if wav_up is not None:
       file_details = {"FileName":wav_up.name,"FileType":wav_up.type}
       st.write(file_details)
@@ -54,7 +54,8 @@ if st.button('Processa'):
     #        audio,
     #        language = None
     #    )
-        input_features = processor(wav_up.getbuffer() , return_tensors="pt").input_features 
+        processado=np.frombuffer(wav_up.getbuffer(), np.int16).flatten().astype(np.float32) / 32768.0
+        input_features = processor(processado , return_tensors="pt").input_features 
         forced_decoder_ids = processor.get_decoder_prompt_ids(language = None, task = "transcribe")
       
         predicted_ids = model.generate(input_features, forced_decoder_ids = forced_decoder_ids)