Update app.py

app.py

@@ -1,11 +1,35 @@
+import json
+import spaces
+import subprocess
 import gradio as gr
+from llama_cpp import Llama
+from llama_cpp_agent.chat_history.messages import Roles
+from llama_cpp_agent.chat_history import BasicChatHistory
+from llama_cpp_agent.providers import LlamaCppPythonProvider
+from llama_cpp_agent import LlamaCppAgent, MessagesFormatterType
 from huggingface_hub import InferenceClient
+from huggingface_hub import hf_hub_download
 
 """
 For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
 """
+SYSTEM = "You are a helpful math assistant. You should always provide your answer in Chinese."
+
 client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+hf_hub_download(
+    repo_id="bartowski/Qwen2.5-Math-7B-Instruct-GGUF",
+    filename="Qwen2.5-Math-7B-Instruct-Q8_0.gguf",
+    local_dir="./models"
+)
+
+llm = Llama(
+    model_path="models/Qwen2.5-Math-7B-Instruct-Q8_0.gguf",
+    flash_attn=True,
+    n_ctx=8192,
+    n_batch=1024
+)
 
+provider = LlamaCppPythonProvider(llm)
 
 def respond(
     message,