fixed audio

app.py

@@ -30,7 +30,7 @@ def check_hallucination(assertion,citation):
 
 
 
-def process_speech(audio_input,input_language):
+def process_speech(input_language, audio_input):
     """
     processing sound using seamless_m4t
     """
@@ -38,14 +38,6 @@ def process_speech(audio_input,input_language):
         return "no audio or audio did not save yet \nplease try again ! "
     print(f"audio : {audio_input}")
     print(f"audio type : {type(audio_input)}")
-    
-    try : 
-        audio_name = f"{np.random.randint(0, 100)}.wav"
-        sr, data = audio_input
-        write(audio_name, sr, data.astype(np.int16))
-        audio_input = audio_name
-    except : 
-        pass
     out = seamless_client.predict(
         "S2TT",
         "file",
@@ -432,11 +424,11 @@ with gr.Blocks(theme='ParityError/Anime') as iface :
         image_output = gr.Markdown(label="output text")
         image_button = gr.Button("process image")
     with gr.Tab("speech to text"):
-        with gr.Row():
-            input_language = gr.Dropdown(languages, label="select the language",value="English",interactive=True)
+        input_language = gr.Dropdown(languages, label="select the language",value="English",interactive=True)
         audio_input = gr.Audio(label="speak",type="filepath",sources="microphone")
         audio_output = gr.Markdown(label="output text")
         audio_button = gr.Button("process audio")
+        audio_button.click(process_speech, inputs=[input_language,audio_input], outputs=audio_output)
     with gr.Tab("hallucination check"):
         assertion = gr.Textbox(label="assertion")
         citation =  gr.Textbox(label="citation text")
@@ -445,7 +437,6 @@ with gr.Blocks(theme='ParityError/Anime') as iface :
         gr.Examples([["i am drunk","sarah is pregnant"]],inputs=[assertion,citation])
     text_button.click(process_and_query, inputs=text_input, outputs=text_output)
     image_button.click(process_image, inputs=image_input, outputs=image_output)
-    audio_button.click(process_speech, inputs=[audio_input,input_language], outputs=audio_output)
     audio_button.click(check_hallucination,inputs=[assertion,citation],outputs=hullucination_output)