Init whisper

app.py

@@ -9,15 +9,29 @@ import torch
 session_token = os.environ.get('SessionToken')      
 # logger.info(f"session_token_: {session_token}")
 
-whisper_model = whisper.load_model("medium")
+whisper_model = whisper.load_model("base")
 
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
-def get_response_from_chatbot(text):
+def get_response_from_chatbot(audio):
     try:
-        print("Testing indentation")
         api = ChatGPT(session_token)
-        resp = api.send_message(text)    
+        audio = whisper.load_audio(audio)
+        audio = whisper.pad_or_trim(audio)
+    
+        mel = whisper.log_mel_spectrogram(audio).to(whisper_model.device)
+    
+        _, probs = whisper_model.detect_language(mel)
+        translate_options = whisper.DecodingOptions(task="translate", fp16 = False)
+    
+        translation = whisper.decode(whisper_model, mel, translate_options)
+    
+        print("Language Spoken: " + transcription.language)
+      
+        print("Translated: " + translation.text)
+    
+        
+        resp = api.send_message(translation.text)    
         api.refresh_auth() 
         api.reset_conversation() 
         response = resp['message']
@@ -137,7 +151,7 @@ with gr.Blocks(title='Talk to chatGPT') as demo:
             chatbot = gr.Chatbot(elem_id="chat_bot", visible=False).style(color_map=("green", "blue"))
             chatbot1 = gr.Chatbot(elem_id="chat_bot1").style(color_map=("green", "blue"))
         with gr.Row(elem_id="prompt_row"):
-            prompt_input = gr.Textbox(lines=2, label="Input text",show_label=True)
+            prompt_input_audio = audio_input_r = gr.Audio(label = 'Record Audio Input',source="microphone",type="filepath")
             chat_history = gr.Textbox(lines=4, label="prompt", visible=False)
             submit_btn = gr.Button(value = "submit",elem_id="submit-btn").style(
                     margin=True,
@@ -145,7 +159,7 @@ with gr.Blocks(title='Talk to chatGPT') as demo:
                     width=100
                 )
             submit_btn.click(fn=chat, 
-                             inputs=[prompt_input, chat_history], 
+                             inputs=[prompt_input_audio, chat_history], 
                              outputs=[chatbot, chat_history],
                             )