Update project_model.py

project_model.py

@@ -1,10 +1,3 @@
-'-*- coding: utf-8 -*-'
-"""project_model.ipynb
-Automatically generated by Colab.
-Original file is located at
-    https://colab.research.google.com/drive/1oopkA5yIlfizFuhXOPmTK7MUNh3Qasa3
-"""
-
 # project_module.py
 
 # Import libraries for ML, CV, NLP, audio, and TTS
@@ -197,7 +190,7 @@ def process_inputs(
     if audio_path:
         # Process audio to text
         audio_text = whisper_pipe(audio_path)["text"]
-        question += ' You are a helpful visual assistant designed for visually impaired users that assists users by answering the following question. If unsure, say "I am not certain."' + audio_text.strip()
+        question += ' ' + audio_text.strip()
 
     # Add user's new question to the history
     session.add_question(question)
@@ -215,9 +208,12 @@ def process_inputs(
     
     # Process the input through Gemma
     gemma_output = gemma_pipe(
-        images=session.current_image,
-        text=question + " This is the shared visual context: " + session.visual_context
-        )
+    images=session.current_image,
+    text=[
+        {"role": "system", "content": 'You are a helpful visual assistant for visually impaired users. If unsure, say "I am not certain."'},
+        {"role": "user", "content": "Question: " + question + " This is the shared visual context: " + session.visual_context}
+    ]
+    )
 
     # Handle the output from Gemma model safely
     if isinstance(gemma_output, list) and len(gemma_output) > 0: