feat: MI50 as a Lyra backend (OpenAI-compatible local GPU)

The MI50 box (CT202) runs an OpenAI-compatible llama.cpp server on
10.0.0.44:8080. Wire it in as a third backend:

- llm.complete gains backend="mi50" (OpenAI client pointed at MI50_BASE_URL)
- config: MI50_BASE_URL (default http://10.0.0.44:8080/v1) + MI50_MODEL
- chat.respond labels the model per backend; web _backend_for maps "mi50"
- UI backend selector adds "MI50 — local GPU"

Verified end-to-end: llm.complete(backend="mi50") returns from the live server.
See homelab-inference memory for the box topology.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

lyra/config.py

@@ -14,6 +14,8 @@ load_dotenv()
 class Config:
     local_base_url: str
     local_model: str
+    mi50_base_url: str  # OpenAI-compatible llama.cpp server on the MI50 box
+    mi50_model: str
     openai_api_key: str
     cloud_model: str
     embed_backend: str  # "cloud" (OpenAI) or "local" (Ollama)
@@ -27,6 +29,8 @@ def load() -> Config:
     return Config(
         local_base_url=os.getenv("LOCAL_BASE_URL", "http://localhost:11434"),
         local_model=os.getenv("LOCAL_MODEL", "qwen2.5:7b-instruct"),
+        mi50_base_url=os.getenv("MI50_BASE_URL", "http://10.0.0.44:8080/v1"),
+        mi50_model=os.getenv("MI50_MODEL", "local-gpu"),
         openai_api_key=os.getenv("OPENAI_API_KEY", ""),
         cloud_model=os.getenv("CLOUD_MODEL", "gpt-4o-mini"),
         embed_backend=os.getenv("EMBED_BACKEND", "cloud").lower(),