Add llama-cpp-python optional dep

requirements.txt

@@ -1,5 +1,8 @@
-fastapi==0.115.0
-uvicorn[standard]==0.34.0
-python-multipart==0.0.20
-biplist==1.0.3
-
+fastapi>=0.104.0
+uvicorn[standard]>=0.24.0
+python-multipart>=0.0.6
+biplist>=1.0.3
+# llama-cpp-python enables GGUF inference endpoints.
+# Uncomment to enable /api/inference/* endpoints.
+# On HF Spaces with GPU, use: llama-cpp-python[cuda]
+# llama-cpp-python>=0.2.0