---

name: multimodal-serving description: "Serve and smoke-check align-anything text, multimodal, and omni-modal inference without reading the source checkout." metadata: disco-role: operating disable-model-invocation: true license: Apache 2.0

Multimodal Serving

Use this sub-skill when a task needs to load an align-anything model, choose a text/multimodal/omni Gradio CLI, prepare media/template inputs, or diagnose serving-time model and media failures.

Start here

1. Read references/model-loading-and-cli.md to choose between load_pretrained_models, AnyModel, AnyModelForScore, text_modal_cli, multi_modal_cli, and omni_modal_cli.
2. Read references/media-and-templates.md before constructing image, audio, video, or mixed omni messages.
3. Use scripts/check_model_loading.py for an import-only, dry-run, or real model-loading smoke check.
4. Use scripts/run_cli_template.sh as the bundled serving launcher template instead of relying on source-tree shell snippets.
5. If loading, decoding, templating, or Gradio startup fails, use references/troubleshooting.md.

Routing guidance

Use this sub-skill for:

• loading base or reward models through align-anything's auto-model registry;
• setting device, dtype, cache, trust_remote_code, and optional modality/omni initialization flags;
• launching the text, multimodal image/audio/video, or MiniCPM-O-style omni CLI;
• translating uploaded media into the processor/model inputs expected by align-anything serving code;
• explaining optional dependency and backend failures at serving time.

Prefer another sub-skill when the task is training/alignment, reward-server deployment, evaluation-benchmark orchestration, or repo-level project setup rather than interactive inference.

Boundaries and safety notes

• Treat remote model code as executable code. Enable trust_remote_code only for model repositories you trust.
• The bundled CLI launcher starts the package's Gradio CLIs, which currently request a shareable Gradio link. Use only in a trusted network/session unless you adapt the local launcher to disable sharing.
• Large multimodal and omni models generally need CUDA/NPU-class memory. CPU import checks are useful, but they do not prove full generation throughput or memory fit.
• Keep all local environment names, private install prefixes, and checkout-specific paths out of downstream reports and user-facing instructions.