updated prompt, model

core_logic.py

@@ -1,15 +1,14 @@
 
 # ./core_logic.py -> Token-safe
 
-"""
-The Inference Engine - Where the "Technical Genius" persona lives. It uses the huggingface_hub InferenceClient to run the model without local CPU strain
-"""
-
 import os
 from groq import Groq
 from tools import web_search, parse_file
 
 client = Groq(api_key=os.getenv("GROQ_API_KEY"))
+model = "llama-3.1-70b-versatile" 
+#model = "mixtral-8x7b-32768"
+#model = "llama-3.1-8b-instant"
 
 # Compressed for token efficiency
 SYSTEM_PROMPT = (
@@ -51,7 +50,7 @@ def chat_function(message, history):
 
     try:
         completion = client.chat.completions.create(
-            model="llama-3.1-8b-instant",
+            model=model,
             messages=messages,
             stream=True,
             temperature=0.2,
@@ -65,4 +64,4 @@ def chat_function(message, history):
                 response_text += token
                 yield response_text
     except Exception as e:
-        yield f"TPM/Rate Limit Error: {str(e)}"
+        yield f"Error: {str(e)}"