From b98a577fbef036c3bab5a9ee1fd01edc2af8386c Mon Sep 17 00:00:00 2001 From: Jinyan Yi Date: Thu, 23 Jul 2026 20:25:20 +0800 Subject: [PATCH] Doc/update ascend quickstart (#32205) --- .../ascend-npus/ascend_npu_quick_start.mdx | 6 +++--- .../ascend-npus/best_practice/qwen3_6_35b_a3b.mdx | 9 +++++++++ .../model-tutorials/qwen3_6_35b_a3b.mdx | 15 ++++++++------- 3 files changed, 20 insertions(+), 10 deletions(-) diff --git a/docs_new/docs/hardware-platforms/ascend-npus/ascend_npu_quick_start.mdx b/docs_new/docs/hardware-platforms/ascend-npus/ascend_npu_quick_start.mdx index 63fc7b065..b46c395af 100644 --- a/docs_new/docs/hardware-platforms/ascend-npus/ascend_npu_quick_start.mdx +++ b/docs_new/docs/hardware-platforms/ascend-npus/ascend_npu_quick_start.mdx @@ -40,7 +40,7 @@ at least **30GB** of free space. If you need to download model weights, check th -We publish both **stable releases** and **daily builds**. Choose a stable release tag (e.g., `v0.5.13.post1-cann9.0.0-a3`) +We publish both **stable releases** and **daily builds**. Choose a stable release tag (e.g., `v0.5.14-cann9.0.0-a3`) if you prefer a validated version, or a daily build tag (e.g., `main-cann9.0.0-a3`) if you need the latest development changes. @@ -52,7 +52,7 @@ adding `--volume /path/to/model:/path/to/model` to the `docker run` command belo ```shell Command # Choose one (uncomment the line you want): -export IMAGE=quay.io/ascend/sglang:v0.5.13.post1-cann9.0.0-a3 # Stable release +export IMAGE=quay.io/ascend/sglang:v0.5.14-cann9.0.0-a3 # Stable release # export IMAGE=quay.io/ascend/sglang:main-cann9.0.0-a3 # Daily build docker run -it --rm --privileged --network=host --ipc=host --shm-size=16g \ @@ -77,7 +77,7 @@ docker run -it --rm --privileged --network=host --ipc=host --shm-size=16g \ ```shell Command # Choose one (uncomment the line you want): -export IMAGE=quay.io/ascend/sglang:v0.5.13.post1-cann9.0.0-910b # Stable release +export IMAGE=quay.io/ascend/sglang:v0.5.14-cann9.0.0-910b # Stable release # export IMAGE=quay.io/ascend/sglang:main-cann9.0.0-910b # Daily build docker run -it --rm --privileged --network=host --ipc=host --shm-size=16g \ diff --git a/docs_new/docs/hardware-platforms/ascend-npus/best_practice/qwen3_6_35b_a3b.mdx b/docs_new/docs/hardware-platforms/ascend-npus/best_practice/qwen3_6_35b_a3b.mdx index 07e430428..12f69659d 100644 --- a/docs_new/docs/hardware-platforms/ascend-npus/best_practice/qwen3_6_35b_a3b.mdx +++ b/docs_new/docs/hardware-platforms/ascend-npus/best_practice/qwen3_6_35b_a3b.mdx @@ -8,6 +8,8 @@ metatags: This page focuses on optimal configuration and benchmark results for Qwen3.6-35B-A3B on the Ascend NPU. For environment setup, model weight download, feature configuration, and deployment instructions, etc., see the [Qwen3.6-35B-A3B Model Tutorial](/docs/hardware-platforms/ascend-npus/model-tutorials/qwen3_6_35b_a3b). On A3 each card has 2 dies, so `--tp-size` is twice the card count; see [Ascend NPU Reference](/docs/hardware-platforms/ascend-npus/ascend_npu_reference#hardware) for details. + +Use image **SGLang `>= v0.5.14`** for these NEXTN configurations. Without `--dataset-path`, `bench_serving --dataset-name random` downloads ShareGPT from Hugging Face; in offline environments, pass a local dataset path (for example a ShareGPT JSON file). ### Low Latency @@ -120,6 +122,7 @@ We tested it based on the `RANDOM` dataset. ```bash Command python -m sglang.bench_serving \ --dataset-name random \ + --dataset-path /path/to/dataset \ --backend sglang \ --host 127.0.0.1 \ --port 6688 \ @@ -217,6 +220,7 @@ We tested it based on the `RANDOM` dataset. ```bash Command python -m sglang.bench_serving \ --dataset-name random \ + --dataset-path /path/to/dataset \ --backend sglang \ --host 127.0.0.1 \ --port 6688 \ @@ -316,6 +320,7 @@ We tested it based on the `RANDOM` dataset. ```bash Command python -m sglang.bench_serving \ --dataset-name random \ + --dataset-path /path/to/dataset \ --backend sglang \ --host 127.0.0.1 \ --port 6688 \ @@ -410,6 +415,7 @@ We tested it based on the `RANDOM` dataset. ```bash Command python -m sglang.bench_serving \ --dataset-name random \ + --dataset-path /path/to/dataset \ --backend sglang \ --host 127.0.0.1 \ --port 6688 \ @@ -507,6 +513,7 @@ We tested it based on the `RANDOM` dataset. ```bash Command python -m sglang.bench_serving \ --dataset-name random \ + --dataset-path /path/to/dataset \ --backend sglang \ --host 127.0.0.1 \ --port 6688 \ @@ -604,6 +611,7 @@ We tested it based on the `RANDOM` dataset. ```bash Command python -m sglang.bench_serving \ --dataset-name random \ + --dataset-path /path/to/dataset \ --backend sglang \ --host 127.0.0.1 \ --port 6688 \ @@ -803,6 +811,7 @@ We tested it based on the `RANDOM` dataset. ```bash Command python -m sglang.bench_serving \ --dataset-name random \ + --dataset-path /path/to/dataset \ --backend sglang \ --host 127.0.0.1 \ --port 6688 \ diff --git a/docs_new/docs/hardware-platforms/ascend-npus/model-tutorials/qwen3_6_35b_a3b.mdx b/docs_new/docs/hardware-platforms/ascend-npus/model-tutorials/qwen3_6_35b_a3b.mdx index 61810f5f1..ef85a87dd 100644 --- a/docs_new/docs/hardware-platforms/ascend-npus/model-tutorials/qwen3_6_35b_a3b.mdx +++ b/docs_new/docs/hardware-platforms/ascend-npus/model-tutorials/qwen3_6_35b_a3b.mdx @@ -16,9 +16,10 @@ comprehension, and tool usage. This document demonstrates the deployment of Qwen3.6-35B-A3B on Ascend NPUs using SGLang, including single-node PD mixed mode, feature configuration, and performance optimization. -This document is validated and written based on **SGLang v0.5.13**. The current model (Qwen3.6-35B-A3B) is fully supported -in this version. To use the latest features (e.g., speculative decoding, multimodal), it is recommended to use -v0.5.13 or a later version. +This document is validated and written based on **SGLang v0.5.14**. The current model (Qwen3.6-35B-A3B) is fully supported +in this version. To use the latest features (e.g., speculative decoding / NEXTN, multimodal), use +**v0.5.14 or a later version**. Ascend images based on `v0.5.13(.post1)` can fail when NEXTN is enabled +(`NPUMHATokenToKVPool` missing `data_strides`). ## Supported features @@ -87,7 +88,7 @@ For details, see [Docker image versions](/docs/hardware-platforms/ascend-npus/as ```bash Command -docker pull quay.io/ascend/sglang:v0.5.13.post1-cann9.0.0-a3 +docker pull quay.io/ascend/sglang:v0.5.14-cann9.0.0-a3 docker run -itd --shm-size=16g --name ${NAME} \ --privileged=true --net=host \ @@ -115,14 +116,14 @@ docker run -itd --shm-size=16g --name ${NAME} \ --device=/dev/davinci_manager:/dev/davinci_manager \ --device=/dev/hisi_hdc:/dev/hisi_hdc \ --entrypoint=bash \ -quay.io/ascend/sglang:v0.5.13.post1-cann9.0.0-a3 +quay.io/ascend/sglang:v0.5.14-cann9.0.0-a3 ``` ```bash Command -docker pull quay.io/ascend/sglang:v0.5.13.post1-cann9.0.0-910b +docker pull quay.io/ascend/sglang:v0.5.14-cann9.0.0-910b docker run -itd --shm-size=16g --name ${NAME} \ --privileged=true --net=host \ @@ -142,7 +143,7 @@ docker run -itd --shm-size=16g --name ${NAME} \ --device=/dev/davinci_manager:/dev/davinci_manager \ --device=/dev/hisi_hdc:/dev/hisi_hdc \ --entrypoint=bash \ -quay.io/ascend/sglang:v0.5.13.post1-cann9.0.0-910b +quay.io/ascend/sglang:v0.5.14-cann9.0.0-910b ```