cookbook-add-model

1---2title: __MODEL_DISPLAY__3description: "Deploy __MODEL_DISPLAY__ with SGLang Diffusion."4metatags:5    description: "Serve __MODEL_DISPLAY__ for __PRIMARY_CAPABILITY__ with SGLang Diffusion."6---7 8import { DiffusionModelTags } from '/src/snippets/diffusion/model-tags.jsx';9import { Deployment } from '/src/snippets/_deployment.jsx';10import { config } from '/src/snippets/configs/__HF_ORG__/__MODEL_SLUG__.jsx';11 12<DiffusionModelTags tags={["__MODALITY__", "__PRIMARY_TASK__", "__DIFFERENTIATOR__", "__IMPORTANT_BOUNDARY__"]} />13 14## 1. Quick start15 16Install with `uv pip install "sglang[diffusion]" --prerelease=allow`, then choose17a verified recipe below.18 19<Deployment config={config} />20 21## 2. Model capabilities22 23[__MODEL_DISPLAY__](__MODEL_URL__) is __CONCRETE_MODEL_DESCRIPTION__. Its strongest use case is __MEASURED_OR_DOCUMENTED_STRENGTH__.24 25Choose it when __SELECTION_GUIDANCE__. The main tradeoff is __CAPABILITY_OR_DEPLOYMENT_BOUNDARY__.26 27| Checkpoint or mode | Best fit | Important boundary |28| --- | --- | --- |29| `__MODEL_ID__` | __USE_CASE__ | __LIMIT__ |30 31## 3. Deployment details32 33Document topology-dependent defaults, platform-specific installation details,34and only the hardware recipes that need explanation beyond the builder.35 36## 4. Request examples37 38__REQUEST_EXAMPLES__39 40## 5. Feature details41 42Document quality contracts, installation requirements, and benchmark scope for43the Server and Request choices exposed in the builder. Keep `torch.compile` and44other experimental controls here until they have a broadly compatible recipe.45