Rm inference from hf hub

vlm.py

@@ -8,7 +8,7 @@ from translation import translate_query
 
 # Initialise once
 HF_TOKEN = os.getenv("HF_TOKEN")
-client = InferenceClient(provider="auto", api_key=HF_TOKEN)
+# client = InferenceClient(provider="auto", api_key=HF_TOKEN)
 
 logger = logging.getLogger("vlm-agent")
 logging.basicConfig(level=logging.INFO, format="%(asctime)s — %(name)s — %(levelname)s — %(message)s", force=True) # Change INFO to DEBUG for full-ctx JSON loader