stage-6: model inference server

scripts/serve_model.py — FastAPI in the CUDA container, loads base Qwen3.5-4B
+ a psyc adapter once and serves POST /infer. Lets the cockpit (no torch in
its venv) put a real fine-tuned model behind a Worker Mesh bot over HTTP.
Dockerfile.train gains a fastapi + uvicorn layer.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Dockerfile.train

@@ -27,6 +27,9 @@ ENV PYTHONUNBUFFERED=1 \
 RUN pip install --upgrade pip && \
     pip install unsloth unsloth_zoo trl datasets
 
+# fastapi + uvicorn power scripts/serve_model.py (the inference server).
+RUN pip install fastapi uvicorn
+
 WORKDIR /workspace
 
 # Scripts are mounted at run time (-v $(pwd)/scripts:/scripts), never baked in.