Deploy MiniCPM5-1B and MiniCPM5-2B — backend router

You're being asked to deploy / serve / chat-with a MiniCPM5-1B or MiniCPM5-2B checkpoint. Your job is to pick exactly one backend skill below based on the user's hardware, format, and goal, then invoke that skill rather than improvising.

1. Required input from the user

Before picking a backend, you MUST know:

Available checkpoints on Hugging Face

If the user has a local copy, accept any directory path that contains config.json and model.safetensors (or the equivalent GGUF / MLX layout).

2. Decision matrix — pick exactly one

If the user has not specified any of the above and asks "how do I run this?":

• CUDA box, want fastest server: pick minicpm5-deploy-vllm.
• Ascend NPU, want an OpenAI-compatible server: pick minicpm5-deploy-vllm-ascend.
• CUDA box, want minimal Python: pick minicpm5-deploy-transformers.
• Apple Silicon laptop: pick minicpm5-deploy-ollama (easiest) or minicpm5-deploy-mlx (fastest).
• Android phone, or an on-device app (Android / iOS): pick minicpm5-deploy-litert.
• CPU only / Windows / low-VRAM: pick minicpm5-deploy-llama-cpp (Q4_K_M).

3. Invocation contract

Once you've picked a backend skill, invoke that skill with MODEL_PATH set. Do NOT inline the backend's commands here — each backend has its own pitfalls (mandatory flags, env-var pins, install order) that the dedicated skill handles. The user explicitly does NOT want you to "improvise" — read the picked sub-skill in full first.

4. Sanity check after deploy

Whichever backend you pick, after launch run this universal sanity check:

# Replace localhost:PORT with the backend's actual port (default in each sub-skill)
curl http://localhost:PORT/v1/chat/completions \
    -H "Content-Type: application/json" \
    -d '{
        "model": "MiniCPM5-2B",
        "messages": [{"role":"user","content":"1+1=?"}],
        "temperature": 1.0, "top_p": 0.95, "max_tokens": 64,
        "chat_template_kwargs": {"enable_thinking": true}
    }'

Expected: HTTP 200 with choices[0].message.content containing "2".

minicpm5-deploy-litert also serves this endpoint: after litert-lm import … minicpm5-2b, litert-lm serve listens on port 9379 (guide); use "model": "minicpm5-2b", add "reasoning_effort": "none" for a direct answer, and cap with max_completion_tokens.

5. Known cross-backend pitfalls

These are common to multiple backends — surface to the user up front:

• Think vs no-think: MiniCPM5-2B uses Think mode with temperature=1.0, top_p=0.95. MiniCPM5-1B also supports No-think with enable_thinking=false + temperature=0.7, top_p=0.95.
• 128 K context: max_position_embeddings=131072, rope_theta=5e6, no rope-scaling. Pass --max-model-len 131072 (vLLM) / --context-length 131072 (SGLang) / -c 131072 (llama.cpp) to use the full window. Lower if VRAM is tight.
• Untied lm_head: tie_word_embeddings=false. Tools that assume the Llama tied default (e.g. mlx_lm.convert < 0.31) will silently drop lm_head → output collapses to random tokens. The MLX skill bakes in the fix.

6. Don't reinvent: link to the cookbook

Each sub-skill is paired with a one-page cookbook in docs/deployment/. The skill is the machine-readable shortcut; the cookbook is the human-readable reference. Both are kept in sync.