fix: update smoke test command in modal_vision_gguf-L4.py for clarity

feat: add new GGUF vision model server implementation and update smoke test configuration

feat: add GGUF vision model server and smoke test for endpoint

feat: update app name to include GPU support and enable GPU snapshotting

feat: update GPU type to L4 and pin vLLM version to 0.22.0 in modal_vision.py

feat: add modal_vision.py to serve Rune Goblin vision model on Modal