Update app.py

app.py

@@ -3,18 +3,18 @@ import whisper
 from gtts import gTTS
 import gradio as gr
 from groq import Groq
-from googletrans import Translator
+from transformers import pipeline
 
 GROQ_API_KEY = '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'
 
-# Load Whisper model for transcription
-model = whisper.load_model("base")
+# Load Whisper model for transcription (use a multilingual model to support Urdu)
+model = whisper.load_model("large")  # Use "large" or "multilingual" for better Urdu support
 
 # Set up Groq API client
 client = Groq(api_key=GROQ_API_KEY)
 
-# Translator for English to Urdu translation
-translator = Translator()
+# Load the translation model
+translator = pipeline("translation_en_to_ur", model="Helsinki-NLP/opus-mt-en-ur")
 
 # Function to get the LLM response from Groq
 def get_llm_response(user_input):
@@ -26,32 +26,32 @@ def get_llm_response(user_input):
 
 # Function to convert text to speech using gTTS
 def text_to_speech(text, output_audio="output_audio.mp3"):
-    tts = gTTS(text, lang='ur')
+    tts = gTTS(text, lang='ur')  # Use 'ur' for Urdu
     tts.save(output_audio)
     return output_audio
 
 # Main chatbot function to handle audio input and output
 def chatbot(audio):
-    # Step 1: Transcribe the audio using Whisper
-    result = model.transcribe(audio)
+    # Step 1: Transcribe the Urdu audio using Whisper
+    result = model.transcribe(audio, language="ur")  # Specify Urdu language
     user_text = result["text"]
 
-    # Step 2: Get LLM response from Groq
+    # Step 2: Get LLM response from Groq (in English)
     response_text = get_llm_response(user_text)
 
-    # Step 3: Translate response to Urdu
-    translated_text = translator.translate(response_text, src='en', dest='ur').text
+    # Step 3: Translate the response from English to Urdu
+    translated_response = translator(response_text)[0]['translation_text']
 
-    # Step 4: Convert the translated text to speech
-    output_audio = text_to_speech(translated_text)
+    # Step 4: Convert the translated response text to Urdu speech
+    output_audio = text_to_speech(translated_response)
 
-    return translated_text, output_audio
+    return translated_response, output_audio
 
-# Gradio interface for real-time interaction
+# Gradio interface for real-time interaction with live microphone input
 iface = gr.Interface(
     fn=chatbot,
-    inputs=gr.Audio(type="filepath"),  # Input from mic or file
-    outputs=[gr.Textbox(), gr.Audio(type="filepath")],  # Output: response text and audio
+    inputs=gr.Audio(type="filepath"),  
+    outputs=[gr.Textbox(), gr.Audio(type="filepath")],  # Output: Urdu text and audio
     live=True
 )