fix: huggingface-cli removed in hub 1.x, use hf download

Dockerfile

@@ -22,8 +22,8 @@ ENV LD_LIBRARY_PATH=/opt/llama
 
 # Bake the model into the image at build time: faster Space restarts,
 # no runtime download. Exact file pin recorded after first build.
-RUN pip install --no-cache-dir "huggingface_hub[cli]>=0.26,<2.0" && \
-    huggingface-cli download unsloth/Qwen3-4B-Instruct-2507-GGUF \
+RUN pip install --no-cache-dir "huggingface_hub>=1.0,<2.0" && \
+    hf download unsloth/Qwen3-4B-Instruct-2507-GGUF \
         --include "*Q4_K_M*.gguf" --local-dir /opt/models && \
     find /opt/models -name "*.gguf" -exec ls -lh {} \;