Update app.py

app.py

@@ -26,7 +26,7 @@ MODEL_FILE = os.environ.get(
     "Qwen3.6-27B-NEO-CODE-HERE-2T-OT-Q4_K_M.gguf",  # 16.1 GB, 94.5% of BF16
 )
 N_CTX = int(os.environ.get("N_CTX", "8192"))          # context window
-N_GPU_LAYERS = int(os.environ.get("N_GPU_LAYERS", "99"))  # offload all layers
+N_GPU_LAYERS = int(os.environ.get("N_GPU_LAYERS", "-1"))  # -1 = offload all layers to GPU
 
 # ---------------------------------------------------------------------------
 #  SYSTEM PROMPT — tuned for coding / code review