Fix: remove hardcoded ENV MODEL_ID=qwen that was silently overriding SPACE_ROLE on every Space

Rewrite: load model directly with Llama class, drop buggy llama_cpp.server subprocess/proxy entirely

Switch base image to Alpine — the pinned wheel is musl-linked, not glibc

Pin llama-cpp-python to fix build timeout (Job timeout / BUILD_ERROR)

Update Dockerfile

Update Dockerfile

Update Dockerfile

Create Dockerfile