minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

docs: update SANA-WM cookbook serve examples (#27626)

Browse Source
This commit is contained in:
Mick
2026-06-09 11:44:31 +08:00
committed by GitHub
parent 0f8673851c
commit f6d53d6d16
1 changed files with 8 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs_new/cookbook/diffusion/SANA-WM/SANA-WM.mdx
+8 -8
View File
@@ -44,9 +44,9 @@ For more details, see the [SANA-WM paper (arXiv)](https://arxiv.org/abs/2605.151
## 2. Installation
SGLang-diffusion offers multiple installation methods depending on your hardware platform. Please refer to the [official SGLang-diffusion installation guide](https://github.com/sgl-project/sglang/blob/main/python/sglang/multimodal_gen/docs/install.md).
SGLang-diffusion offers multiple installation methods depending on your hardware platform. Please refer to the [SGLang Diffusion installation guide](../../../docs/sglang-diffusion/installation).
SANA-WM adds the `SanaWMTransformer3DModel` + GDN kernels, the `SanaWMTwoStagePipeline` (dense bidirectional + chunk-causal streaming), and the `SanaWMRealtimePipeline` with the `/v1/realtime_video` WebSocket router. The diffusion server CLI is invoked as `python -m sglang.multimodal_gen.runtime.entrypoints.cli.main`.
SANA-WM adds the `SanaWMTransformer3DModel` + GDN kernels, the `SanaWMTwoStagePipeline` (dense bidirectional + chunk-causal streaming), and the `SanaWMRealtimePipeline` with the `/v1/realtime_video` WebSocket router. Use `sglang serve` to launch the diffusion server.
## 3. Model Setup
@@ -87,7 +87,7 @@ The **bidirectional** checkpoint generates the whole clip in **one shot** (full
Launch with the two-stage pipeline and **no** `--streaming` flag (dense is the default — `streaming` defaults to `False`):
```bash Command
python -m sglang.multimodal_gen.runtime.entrypoints.cli.main serve \
sglang serve \
--model-path Efficient-Large-Model/SANA-WM_bidirectional \
--pipeline-class-name SanaWMTwoStagePipeline \
--host 127.0.0.1 --port 30000
@@ -122,7 +122,7 @@ The **streaming** checkpoint generates a **full camera-controlled clip in one re
Launch with the two-stage pipeline + the streaming flags:
```bash Command
python -m sglang.multimodal_gen.runtime.entrypoints.cli.main serve \
sglang serve \
--model-path Efficient-Large-Model/SANA-WM_streaming \
--pipeline-class-name SanaWMTwoStagePipeline \
--streaming --refiner-chunked \
@@ -167,7 +167,7 @@ The response is a `VideoResponse`; fetch the rendered MP4 via the returned refer
Launch with the realtime pipeline **pinned** — the checkpoint defaults to `SanaWMTwoStagePipeline`, so realtime must be selected explicitly (see §3). The `/v1/realtime_video` router is always mounted and becomes functional once the realtime config is active, because `SanaWMRealtimeConfig` has a registered realtime adapter (`SanaWMRealtimeAdapter`).
```bash Command
python -m sglang.multimodal_gen.runtime.entrypoints.cli.main serve \
sglang serve \
--model-path Efficient-Large-Model/SANA-WM_streaming \
--pipeline-class-name SanaWMRealtimePipeline \
--host 127.0.0.1 --port 30000
@@ -177,20 +177,20 @@ Common launch variants:
```bash Command
# recommended multi-GPU realtime profile
python -m sglang.multimodal_gen.runtime.entrypoints.cli.main serve \
sglang serve \
--model-path Efficient-Large-Model/SANA-WM_streaming \
--pipeline-class-name SanaWMRealtimePipeline \
--num-gpus 8 --sp-degree 8 \
--host 127.0.0.1 --port 30000
# single GPU
python -m sglang.multimodal_gen.runtime.entrypoints.cli.main serve \
sglang serve \
--model-path Efficient-Large-Model/SANA-WM_streaming \
--pipeline-class-name SanaWMRealtimePipeline \
--num-gpus 1 --host 127.0.0.1 --port 30000
# offload DiT + text encoder to CPU (tight VRAM)
python -m sglang.multimodal_gen.runtime.entrypoints.cli.main serve \
sglang serve \
--model-path Efficient-Large-Model/SANA-WM_streaming \
--pipeline-class-name SanaWMRealtimePipeline \
--host 127.0.0.1 --port 30000 \