knowledge: sync upstream main and duplex source routes

adapters/vllm_omni/manifest.yaml

@@ -172,9 +172,9 @@ modules:
     import_check: "from vllm_omni.benchmarks.serve import main; from vllm_omni.entrypoints.cli.benchmark.serve import OmniBenchmarkServingSubcommand; print('Benchmarks imports OK')"
     wave: 2
   fullduplex:
-    local_paths: [vllm_omni/experimental/fullduplex/]
-    test_paths: [tests/benchmarks/duplex/, tests/e2e/features/fullduplex/]
-    import_check: "from vllm_omni.experimental.fullduplex.client import RealtimeDuplexClient; print('Full-duplex client imports OK')"
+    local_paths: [vllm_omni/clients/, vllm_omni/experimental/fullduplex/]
+    test_paths: [tests/clients/, tests/benchmarks/duplex/, tests/e2e/features/fullduplex/]
+    import_check: "from vllm_omni.clients.duplex import DuplexClient; print('Full-duplex client imports OK')"
     wave: 2
 
 review_routes:
@@ -187,6 +187,9 @@ review_routes:
 - prefix: vllm_omni/core/
   owner: scheduler
   doc: repos/vllm-omni/components/scheduler/rules.md
+- prefix: vllm_omni/clients/
+  owner: serving
+  doc: repos/vllm-omni/components/serving/rules.md
 - prefix: vllm_omni/entrypoints/
   owner: serving
   doc: repos/vllm-omni/components/serving/rules.md

adapters/vllm_omni/release_baseline.yaml

@@ -2,9 +2,9 @@ schema_version: 1
 
 upstream:
   repository: vllm-project/vllm-omni
-  previous_audited_sha: 1e74807c356be5cba4239faae7c38763fe1a44be
-  audited_ref: dev/vllm-align
-  audited_sha: 1015b1ced073195ed19f0610264dfdf8a8150e6c
+  previous_audited_sha: 1015b1ced073195ed19f0610264dfdf8a8150e6c
+  audited_ref: main
+  audited_sha: 44d3ae100afa8411770b8b4d9442318bc5e897b3
 
 # These fingerprints are generated from sorted registry/deploy inventories.
 # They make the baseline exact without duplicating hundreds of entries here.
@@ -13,8 +13,8 @@ inventories:
     count: 89
     sha256: cd6b56fe44e88ca35e4d2c961a00cf7a257695ca193e8279d6562eb72ca58fc4
   diffusion:
-    count: 65
-    sha256: 6107384ad499e668c1ed3690d740b7be816c900c504f244dce557ef51a257f2a
+    count: 66
+    sha256: caa4b884df4985b9d828cd54a3da3cf4728dabf655f345df1c361b4143ef94a3
   pipelines:
     count: 59
     sha256: 4f012fb016221c2115520b16577bb69be03e84987122729bea696c14e5bca42f
@@ -69,6 +69,7 @@ path_owners:
   scheduler:
   - vllm_omni/core/
   serving:
+  - vllm_omni/clients/
   - vllm_omni/engine/
   - vllm_omni/entrypoints/
   testing:

doc/architecture/KNOWLEDGE.md

@@ -41,10 +41,8 @@ extensions merely because they are absent from the common source.
   (per-file upstream created/updated dates, captured before the submodule's git
   history was removed — used for page frontmatter).
 - **Code-mirror pin:** the `knowledge/repos/vllm-omni/components/` source maps
-  are verified against vllm-omni `v0.29.0-alignment @
-  1015b1ced073195ed19f0610264dfdf8a8150e6c`. Here
-  `v0.29.0-alignment` is a snapshot label for `dev/vllm-align`, not an
-  upstream tag or main revision. The canonical machine baseline is
+  are verified against vllm-omni `main @
+  44d3ae100afa8411770b8b4d9442318bc5e897b3`. The canonical machine baseline is
   `adapters/vllm_omni/release_baseline.yaml`.
 - **AFD code-mirror pin:** the `knowledge/repos/afd-plugin/components/` owner
   maps and architecture summaries are verified against

knowledge/repos/vllm-omni/benchmark/guides/_index.md

@@ -1,7 +1,7 @@
 ---
 title: "vLLM-Omni 性能指南"
 created: 2026-07-10
-updated: 2026-09-04
+updated: 2026-09-06
 type: index
 tags: [vllm-omni, benchmark]
 sources: ["PR #6522", "PR #6634", vllm_omni/benchmarks/data_modules/omniinteract_dataset.py, vllm_omni/benchmarks/omniinteract.py, vllm_omni/benchmarks/duplex/]

knowledge/repos/vllm-omni/benchmark/guides/omniinteract-realtime.md

@@ -1,10 +1,10 @@
 ---
 title: "OmniInteract Realtime Benchmark"
 created: 2026-09-04
-updated: 2026-09-05
+updated: 2026-09-06
 type: guide
 tags: [vllm-omni, benchmark]
-sources: ["PR #6522", "PR #6696", "PR #6818", vllm_omni/benchmarks/metrics/metrics.py, vllm_omni/experimental/fullduplex/openai/protocol.py, vllm_omni/experimental/fullduplex/openai/chat_fallback.py, tests/benchmarks/metrics/test_metrics.py, tests/benchmarks/patch/test_patch.py, docs/cli/bench/serve.md, vllm_omni/benchmarks/data_modules/omniinteract_dataset.py, vllm_omni/benchmarks/omniinteract.py, vllm_omni/benchmarks/patch/patch.py, vllm_omni/entrypoints/cli/benchmark/cli_args.py]
+sources: ["PR #6522", "PR #6696", "PR #6818", vllm_omni/benchmarks/metrics/metrics.py, vllm_omni/entrypoints/duplex/protocol.py, vllm_omni/entrypoints/duplex/chat_fallback.py, tests/benchmarks/metrics/test_metrics.py, tests/benchmarks/patch/test_patch.py, docs/cli/bench/serve.md, vllm_omni/benchmarks/data_modules/omniinteract_dataset.py, vllm_omni/benchmarks/omniinteract.py, vllm_omni/benchmarks/patch/patch.py, vllm_omni/entrypoints/cli/benchmark/cli_args.py]
 confidence: high
 ---
 

knowledge/repos/vllm-omni/components/configuration/_index.md

@@ -1,7 +1,7 @@
 ---
 title: "vLLM-Omni Configuration"
 created: 2026-07-16
-updated: 2026-09-05
+updated: 2026-09-06
 type: index
 tags: [vllm-omni, components, config]
 sources: ["claude-workflow-starter-private@296ea45", vllm_omni/config/]
@@ -15,7 +15,7 @@ sources: ["claude-workflow-starter-private@296ea45", vllm_omni/config/]
   `tests/diffusion/test_diffusion_config_propagation.py`，以及各公开入口附近的配置测试
 - 部署配置：`vllm_omni/deploy/*.yaml`，以及 `pipeline_registry.py`、
   `endpoint_policy.py`、`server_settings.py`、`yaml_util.py`、`composable_parallel/`
-- 源码校验：以上路径在 `v0.29.0-alignment @ 1015b1ce`（`dev/vllm-align` 快照）验证存在；机器基线见
+- 源码校验：以上路径在 `main @ 44d3ae10`（upstream `main` 快照）验证存在；机器基线见
   `adapters/vllm_omni/release_baseline.yaml`
 
 ## 什么时候查这里

knowledge/repos/vllm-omni/components/diffusion/_index.md

@@ -1,7 +1,7 @@
 ---
 title: "Diffusion"
 created: 2026-07-10
-updated: 2026-09-05
+updated: 2026-09-06
 type: index
 tags: [vllm-omni, components, diffusion]
 sources: ["PR #5543", vllm_omni/platforms/cuda/platform.py, vllm_omni/diffusion/attention/backends/flashinfer_attn.py]
@@ -10,7 +10,7 @@ sources: ["PR #5543", vllm_omni/platforms/cuda/platform.py, vllm_omni/diffusion/
 # Diffusion
 
 - 源码入口：`vllm_omni/diffusion/` 全树，含 17 个子模块：attention、cache、diffusion_kv、distributed、executor、hooks、layers、lora、model_loader、models、offloader、postprocess、profiler、quantization、sched、utils、worker
-- 源码校验：以上子模块均已在 `v0.29.0-alignment @ 1015b1ce`（`dev/vllm-align` 快照）验证存在；MiniMax-H3 的 VAE eager
+- 源码校验：以上子模块均已在 `main @ 44d3ae10`（upstream `main` 快照）验证存在；MiniMax-H3 的 VAE eager
   ops 仍由模型 owner 维护，其他 shared/模型机制按各自规则审查
 - 主要职责：多个 diffusion 模型共用的 pipeline、执行循环、scheduler 接入和运行机制
 

knowledge/repos/vllm-omni/components/distributed/_index.md

@@ -1,7 +1,7 @@
 ---
 title: "Distributed（跨 stage 通信与数据搬运）"
 created: 2026-07-16
-updated: 2026-09-05
+updated: 2026-09-06
 type: index
 tags: [vllm-omni, components, distributed]
 sources: [vllm_omni/distributed/omni_connectors/, vllm_omni/distributed/omni_coordinator/, docs/design/feature/disaggregated_inference.md]
@@ -14,7 +14,7 @@ sources: [vllm_omni/distributed/omni_connectors/, vllm_omni/distributed/omni_coo
   `vllm_omni/distributed/omni_coordinator/`（协调器与 load balancer）
 - 知识面另覆盖跨 stage ZMQ 路由/端口分配（`vllm_omni/engine/stage_engine_startup.py::OmniMasterServer`）
   ——组件划分服务知识归属，与 manifest 运行时粒度不同
-- 源码校验：以上路径与下列锚点均已在 `v0.29.0-alignment @ 1015b1ce`（`dev/vllm-align` 快照）验证存在：
+- 源码校验：以上路径与下列锚点均已在 `main @ 44d3ae10`（upstream `main` 快照）验证存在：
   `OmniConnectorBase`（connectors/base.py:12）、`OmniKVTransferManager`
   （kv_transfer_manager.py:348）、`LoadBalancer` 三实现（load_balancer.py:39/64/74/102）、
   `OmniMasterServer._allocate_route_locked`（stage_engine_startup.py:262）

knowledge/repos/vllm-omni/components/model-executor/_index.md

@@ -1,7 +1,7 @@
 ---
 title: "Model Executor"
 created: 2026-07-10
-updated: 2026-09-05
+updated: 2026-09-06
 type: index
 tags: [vllm-omni, components, model-executor]
 sources: []
@@ -13,7 +13,7 @@ sources: []
   `vllm_omni/model_extras/`（模型专有请求/prompt 转换）、`vllm_omni/worker/`（gpu_*_worker、
   gpu_*_model_runner、mixins）、`vllm_omni/inputs/`（runner 输入预处理）和设备平台层
   `vllm_omni/platforms/<cuda|musa|npu|rocm|xpu>/`
-- 源码校验：以上路径均已在 `v0.29.0-alignment @ 1015b1ce`（`dev/vllm-align` 快照）验证存在；stage 配置已经迁移到
+- 源码校验：以上路径均已在 `main @ 44d3ae10`（upstream `main` 快照）验证存在；stage 配置已经迁移到
   `vllm_omni/deploy/`，NPU/XPU 等平台可继续拥有自己的 worker 覆盖
 - 测试入口：共享 runner 行为看 `tests/worker/`，具体模型 consumer 看 `tests/model_executor/`
 - 主要职责：AR/LLM stage、stage 配置、并行与设备启动、runner 到模型的输入预处理合同和跨阶段数据桥接

knowledge/repos/vllm-omni/components/scheduler/_index.md

@@ -1,7 +1,7 @@
 ---
 title: "Scheduler（AR/生成请求调度）"
 created: 2026-07-16
-updated: 2026-09-05
+updated: 2026-09-06
 type: index
 tags: [vllm-omni, components, scheduler]
 sources: [vllm_omni/core/sched/omni_ar_scheduler.py, vllm_omni/core/sched/omni_generation_scheduler.py, vllm_omni/core/sched/omni_scheduler_mixin.py, vllm_omni/core/sched/output.py, vllm_omni/core/prefix_cache.py, docs/design/module/ar_runtime.md, docs/design/module/archive/ar_module.md]
@@ -11,7 +11,7 @@ sources: [vllm_omni/core/sched/omni_ar_scheduler.py, vllm_omni/core/sched/omni_g
 
 - 源码入口：`vllm_omni/core/sched/`（`omni_ar_scheduler.py`、`omni_generation_scheduler.py`、
   `omni_scheduler_mixin.py`、`omni_scheduling_coordinator.py`）和 `vllm_omni/core/prefix_cache.py`
-- 源码校验：以上路径与下列类均已在 `v0.29.0-alignment @ 1015b1ce`（`dev/vllm-align` 快照）验证存在：
+- 源码校验：以上路径与下列类均已在 `main @ 44d3ae10`（upstream `main` 快照）验证存在：
   `OmniARScheduler`（omni_ar_scheduler.py:77）、`OmniARAsyncScheduler`（:1033）、
   `KVCacheTransferData`（distributed/omni_connectors/kv_transfer_manager.py:146）、
   `OmniGenerationScheduler`（omni_generation_scheduler.py:35）、

knowledge/repos/vllm-omni/components/serving/_index.md

@@ -1,16 +1,17 @@
 ---
 title: "Serving"
 created: 2026-07-10
-updated: 2026-09-05
+updated: 2026-09-06
 type: index
 tags: [vllm-omni, components, serving]
 sources: []
 ---
 
 # Serving
 
+- Duplex 公共客户端：`vllm_omni/clients/`；HTTP/WS 实现位于 `vllm_omni/entrypoints/duplex/`，模型 adapter 留在各模型的 `duplex/` 子目录。
 - 主要源码入口：`vllm_omni/entrypoints/`（cli、openai、openpi 及 omni/async_omni 入口）和 `vllm_omni/engine/`（orchestrator、stage engine core、stage pool/runtime、output processor）
-- 源码校验：以上路径均已在 `v0.29.0-alignment @ 1015b1ce`（`dev/vllm-align` 快照）验证存在
+- 源码校验：以上路径均已在 `main @ 44d3ae10`（upstream `main` 快照）验证存在
 - 主要职责：用户入口、请求解析、在线服务和 engine 边界
 
 ## 什么时候查这里

knowledge/repos/vllm-omni/components/serving/rules-session-lifecycle.md

@@ -1,10 +1,10 @@
 ---
 title: "Serving session lifecycle 规则"
 created: 2026-09-05
-updated: 2026-09-05
+updated: 2026-09-06
 type: rule
 tags: [vllm-omni, components, serving]
-sources: ["PR #6537", "PR #6354", vllm_omni/config/stage_config.py, vllm_omni/entrypoints/openai/api_server.py, vllm_omni/experimental/fullduplex/core/adapter.py, vllm_omni/experimental/fullduplex/core/runtime.py, vllm_omni/experimental/fullduplex/mage_vl/adapter.py, vllm_omni/experimental/fullduplex/mage_vl/serving/backend.py, vllm_omni/experimental/fullduplex/mage_vl/serving/server.py, tests/e2e/features/fullduplex/test_mage_vl_adapter.py, tests/e2e/features/fullduplex/test_mage_vl_serving.py, tests/e2e/online_serving/nemotron_voicechat_realtime_duplex.py]
+sources: ["PR #6537", "PR #6354", vllm_omni/config/stage_config.py, vllm_omni/entrypoints/openai/api_server.py, vllm_omni/engine/duplex/adapter.py, vllm_omni/engine/duplex/runtime.py, vllm_omni/experimental/fullduplex/mage_vl/adapter.py, vllm_omni/experimental/fullduplex/mage_vl/serving/backend.py, vllm_omni/experimental/fullduplex/mage_vl/serving/server.py, tests/e2e/features/fullduplex/test_mage_vl_adapter.py, tests/e2e/features/fullduplex/test_mage_vl_serving.py, tests/e2e/online_serving/nemotron_voicechat_realtime_duplex.py]
 confidence: high
 ---
 

knowledge/repos/vllm-omni/models/_index.md

@@ -1,7 +1,7 @@
 ---
 title: "vLLM-Omni 模型 owner"
 created: 2026-07-10
-updated: 2026-09-02
+updated: 2026-09-06
 type: index
 tags: [vllm-omni, models]
 sources: []

knowledge/repos/vllm-omni/models/catalog.md

@@ -1,17 +1,16 @@
 ---
 title: "模型代码入口与 registry 快照"
 created: 2026-07-16
-updated: 2026-09-05
+updated: 2026-09-06
 type: guide
 tags: [vllm-omni, models]
-sources: [vllm_omni/model_executor/models/registry.py, vllm_omni/diffusion/registry.py, vllm_omni/config/pipeline_registry.py, vllm_omni/deploy/, "PR #5885", "PR #6354", "PR #6727"]
+sources: [vllm_omni/model_executor/models/registry.py, vllm_omni/diffusion/registry.py, vllm_omni/config/pipeline_registry.py, vllm_omni/deploy/, "PR #5885", "PR #6354", "PR #6727", "PR #5918"]
 ---
 
 # 模型代码入口与 registry 快照
 
 本页提供模型描述到代码目录的自动定位入口，不维护逐模型 class 映射。下方计数是
-`v0.29.0-alignment @ 1015b1ce`（2026-09-05）快照，数字会漂移，不能凭它断言“不支持”。
-`v0.29.0-alignment` 是 `dev/vllm-align` 的快照标签，不是上游 tag 或 main 修订。
+`main @ 44d3ae10`（2026-09-06 校验）快照，数字会漂移，不能凭它断言“不支持”。
 
 ## Direct 模型代码入口
 
@@ -36,13 +35,15 @@ adapter。已有专属知识 owner 可从 [models index](_index.md) 按名称进
 
 | 注册点 | 位置 | 计数 |
 |---|---|---|
-| AR/omni 架构 | `model_executor/models/registry.py` `_OMNI_MODELS` | 89 个架构名 / 32 个模型族目录 |
-| Diffusion pipeline | `diffusion/registry.py` `_DIFFUSION_MODELS` | 65 条 pipeline / 40 个模型族目录 |
+| AR/omni 架构 | `model_executor/models/registry.py` `_OMNI_MODELS` | 89 个架构名 / 33 个模型族目录 |
+| Diffusion pipeline | `diffusion/registry.py` `_DIFFUSION_MODELS` | 66 条 pipeline / 41 个模型族目录 |
 | Pipeline（model_type） | `config/pipeline_registry.py` `OMNI_PIPELINES` | 59 个 key |
 | Deploy YAML | `vllm_omni/deploy/*.yaml` | 89 份 |
 
-`1e74807c → 1015b1ce` 的对齐分支审计未改变四个 registry/deploy inventory 的
-条目或指纹；inventory 不变不代表注册逻辑或已有 deploy 文件内容没有变化。
+`1015b1ce → 44d3ae10` 将审计目标从 alignment 分支切回 upstream main。
+新增 `Magi2Pipeline`，实现位于 `vllm_omni/diffusion/models/magi2/pipeline_magi2.py`；
+AR、`OMNI_PIPELINES` 与 deploy YAML inventory 不变。MAGI-2 暂走共享 Diffusion owner，
+注册存在不等于所有硬件、online serving 或性能路径均已验证。^[PR #5918]
 
 PR #5885 adds `MiniMaxH3TextEncoder`, the opt-in `minimax_h3_disaggregated` pipeline key,
 and regular/Turbo deployment profiles. It preserves bare H3's fused single-stage fallback;
@@ -101,37 +102,37 @@ SANA-Video 2B 新增 native `SanaVideoPipeline`（T2V）和
 `OMNI_PIPELINES` key 或 deploy YAML。模型专有的 checkpoint、VAE 和 adapter 边界见
 [SANA Video](sana-video/_index.md)。^[PR #5508]
 
-## AR/omni 模型族（31）
+## AR/omni 模型族（33）
 
-aura_omni、audex、bagel、cosyvoice3、covo_audio、dots_tts、dynin_omni、fish_speech、
-gepard、glm_image、glm_tts、higgs_audio_v2、higgs_audio_v3、hunyuan_image3、indextts2、
-mammoth_moda2、mimo_audio、ming_flash_omni、ming_tts、minicpmo_4_5、minimax_music3、moss_tts、
-moss_tts_nano、nemotron_voicechat、omnivoice、personaplex、qwen2_5_omni、qwen3_omni、
-qwen3_tts、step_audio2、voxcpm2、voxtral_tts
+audex、aura_omni、bagel、cosyvoice3、covo_audio、dots_tts、dynin_omni、fish_speech、gepard、glm_image、
+glm_tts、higgs_audio_v2、higgs_audio_v3、hunyuan_image3、indextts2、mammoth_moda2、mimo_audio、
+ming_flash_omni、ming_tts、minicpmo_4_5、minimax_h3、minimax_music3、moss_tts、moss_tts_nano、
+nemotron_voicechat、omnivoice、personaplex、qwen2_5_omni、qwen3_omni、qwen3_tts、step_audio2、voxcpm2、
+voxtral_tts
 
-## Diffusion 模型族（40）
+## Diffusion 模型族（41）
 
-bagel、boogu_image、cosmos3、diffusers_adapter（通用 diffusers 桥）、dreamzero、ernie_image、
-flux、flux2、flux2_klein、glm_image、gr00t、helios、
-hidream_image、hidream_o1_image、hunyuan_image3、hunyuan_video、internvla_a1、krea2、lance、
-lingbot_video、lingbot_world、longcat_image、longcat_video、ltx2、ming_flash_omni、
-minimax_h3、nextstep_1_1、omnigen2、omnivoice、ovis_image、pi0、qwen_image、sana_video、sana_wm、sd3、
-sdxl、sensenova_u1、stable_audio、wan2_2、z_image
+bagel、boogu_image、cosmos3、diffusers_adapter、dreamzero、ernie_image、flux、flux2、flux2_klein、
+glm_image、gr00t、helios、hidream_image、hidream_o1_image、hunyuan_image3、hunyuan_video、internvla_a1、
+krea2、lance、lingbot_video、lingbot_world、longcat_image、longcat_video、ltx2、magi2、ming_flash_omni、
+minimax_h3、nextstep_1_1、omnigen2、omnivoice、ovis_image、pi0、qwen_image、sana_video、sana_wm、sd3、sdxl、
+sensenova_u1、stable_audio、wan2_2、z_image
 
-## OMNI_PIPELINES key（56）
+## OMNI_PIPELINES key（59）
 
-Gr00tN1d7（注意:唯一 CamelCase key）、audex_s2s、audex_thinker_only、audex_tta、
-audex_tts、aura_omni、bagel、bagel_single_stage、bagel_think、cosyvoice3、covo_audio、
-dots_tts、dreamzero、dynin_omni、fish_qwen3_omni、gepard、glm_image、glm_tts、
-higgs_audio_v2、higgs_multimodal_qwen3、hunyuan_image3_ar、hunyuan_image3_dit、
-hunyuan_image_3_moe、hunyuan_video_15、indextts2、indextts2_5、lance、mimo_audio、
-ming_flash_omni、ming_flash_omni_image、ming_flash_omni_thinker_only、
-ming_flash_omni_tts、ming_tts、ming_tts_moe、mammoth_moda2、mammoth_moda2_ar、minicpmo_4_5、minimax_music3、
+Gr00tN1d7、audex_s2s、audex_thinker_only、audex_tta、audex_tts、aura_omni、bagel、bagel_single_stage、
+bagel_think、cosyvoice3、covo_audio、dots_tts、dreamzero、dynin_omni、fish_qwen3_omni、gepard、glm_image、
+glm_tts、higgs_audio_v2、higgs_multimodal_qwen3、hunyuan_image3_ar、hunyuan_image3_dit、
+hunyuan_image_3_moe、hunyuan_video_15、indextts2、indextts2_5、lance、mammoth_moda2、mammoth_moda2_ar、
+mimo_audio、ming_flash_omni、ming_flash_omni_image、ming_flash_omni_thinker_only、
+ming_flash_omni_tts、ming_tts、ming_tts_moe、minicpmo_4_5、minimax_h3_disaggregated、minimax_music3、
 moss_tts_delay、moss_tts_local、moss_tts_nano、moss_tts_realtime、nemotron_labs_audex、
 nemotron_labs_voicechat、nemotron_voicechat、omnivoice、personaplex、pi0、qwen2_5_omni、
-qwen2_5_omni_thinker_only、qwen3_omni_moe（resolver）、qwen3_omni_moe_thinker_only、qwen3_tts、step_audio_2、
+qwen2_5_omni_thinker_only、qwen3_omni_moe、qwen3_omni_moe_thinker_only、qwen3_tts、step_audio_2、
 step_audio_2_asr、voxcpm2、voxtral_tts、wan2_2_ti2v
 
+`diffusers_adapter` 是通用 diffusers 桥；`Gr00tN1d7` 保留其大小写，`qwen3_omni_moe` 是 resolver key。
+
 注意：单 stage diffusion 模型**多数不在** `OMNI_PIPELINES`（引擎为它们生成
 默认 diffusion stage 配置,见 [Config 组件](../components/configuration/architecture.md)）;
 但存在例外——omnivoice、Gr00tN1d7、lance、dreamzero 等单 stage
@@ -141,8 +142,8 @@ step_audio_2_asr、voxcpm2、voxtral_tts、wan2_2_ti2v
 
 ```bash
 python tools/audit_vllm_omni_release.py \
-  --from 1e74807c \
-  --to 1015b1ce \
+  --from 1015b1ce \
+  --to 44d3ae10 \
   --repo <vllm-omni-checkout> \
   --mode report-only
 ```