change to mistral v0.3

app.py

@@ -1,16 +1,16 @@
 import gradio as gr
 from llama_cpp import Llama
 
-model = "lmstudio-community/gemma-2-2b-it-GGUF"
+model = "MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF"
 llm = Llama.from_pretrained(
     repo_id=model,
-    filename="gemma-2-2b-it-Q8_0.gguf",
+    filename="Mistral-7B-Instruct-v0.3.Q4_K_M.gguf",
     verbose=True,
     use_mmap=False,
     use_mlock=True,
     n_threads=2,
     n_threads_batch=2,
-    n_ctx=8000,
+    n_ctx=32000,
 )
 
 
@@ -22,8 +22,7 @@ def respond(
     temperature,
     top_p,
 ):
-    # messages = [{"role": "system", "content": system_message}]
-    messages = []
+    messages = [{"role": "system", "content": system_message}]
 
     for val in history:
         if val[0]: