Update app.py

app.py

@@ -3,7 +3,7 @@ import subprocess
 import time
 import requests
 import gradio as gr
-from huggingface_hub import hf_hub_download
+from huggingface_hub import snapshot_download
 
 MODEL_REPO = "bartowski/Qwen_Qwen3.5-9B-GGUF"
 MODEL_FILE = "Qwen3.5-9B-Q4_K_M.gguf"
@@ -12,12 +12,28 @@ LLAMA_DIR = "llama.cpp"
 SERVER_PORT = "8000"
 
 
+def install_build_tools():
+    subprocess.run(
+        ["apt-get", "update"],
+        stdout=subprocess.DEVNULL,
+        stderr=subprocess.DEVNULL
+    )
+
+    subprocess.run(
+        ["apt-get", "install", "-y", "build-essential", "cmake", "git"],
+        stdout=subprocess.DEVNULL,
+        stderr=subprocess.DEVNULL
+    )
+
+
 def download_model():
-    return hf_hub_download(
+    path = snapshot_download(
         repo_id=MODEL_REPO,
-        filename=MODEL_FILE
+        allow_patterns=[MODEL_FILE]
     )
 
+    return os.path.join(path, MODEL_FILE)
+
 
 def setup_llama():
     if not os.path.exists(LLAMA_DIR):
@@ -33,7 +49,7 @@ def setup_llama():
     )
 
 
-def start_server(model_path):
+def start_llama_server(model_path):
     subprocess.Popen(
         [
             "./server",
@@ -48,6 +64,7 @@ def start_server(model_path):
 
 def wait_for_server():
     url = f"http://localhost:{SERVER_PORT}/health"
+
     for _ in range(60):
         try:
             requests.get(url, timeout=1)
@@ -72,11 +89,13 @@ def chat(prompt):
     return data["choices"][0]["message"]["content"]
 
 
+install_build_tools()
+
 model_path = download_model()
 
 setup_llama()
 
-start_server(model_path)
+start_llama_server(model_path)
 
 wait_for_server()