minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 13 Packages Projects Releases Wiki Activity

[NPU] [Diffusion] Use fused operator to improve Wan model E2E performance. (#24028)

Browse Source 
Co-authored-by: gengjinsong <gengjinsong@huawei.com>
Co-authored-by: sglang-npu-bot <sglangnpu@163.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: gengjinsong <904939979@qq.com>
Co-authored-by: ronnie_zheng <zl19940307@163.com>
This commit is contained in:
gjsheu
2026-05-11 07:17:31 +03:00
committed by GitHub
co-authored by gengjinsong sglang-npu-bot gemini-code-assist[bot] gengjinsong ronnie_zheng
parent 958f35d1e0
commit 36c93fc6fb
3 changed files with 63 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
scripts/ci/npu/npu_ci_install_dependency.sh
+3 -3
View File
@@ -68,10 +68,10 @@ ${UV_PIP_INSTALL} triton-ascend
### Install sgl-kernel-npu
SGLANG_KERNEL_NPU_TAG="2026.03.10.rc1"
SGLANG_KERNEL_NPU_TAG="2026.05.01"
mkdir sgl-kernel-npu
(cd sgl-kernel-npu && wget "${GITHUB_PROXY_URL:=""}https://github.com/sgl-project/sgl-kernel-npu/releases/download/${SGLANG_KERNEL_NPU_TAG}/sgl-kernel-npu-${SGLANG_KERNEL_NPU_TAG}-torch2.8.0-py311-cann8.5.0-${DEVICE_TYPE}-$(arch).zip" \
&& unzip ./sgl-kernel-npu-${SGLANG_KERNEL_NPU_TAG}-torch2.8.0-py311-cann8.5.0-${DEVICE_TYPE}-$(arch).zip \
(cd sgl-kernel-npu && wget "${GITHUB_PROXY_URL:=""}https://github.com/sgl-project/sgl-kernel-npu/releases/download/${SGLANG_KERNEL_NPU_TAG}/sgl-kernel-npu-${SGLANG_KERNEL_NPU_TAG}-torch${PYTORCH_VERSION}-py311-cann8.5.0-${DEVICE_TYPE}-$(arch).zip" \
&& unzip ./sgl-kernel-npu-${SGLANG_KERNEL_NPU_TAG}-torch${PYTORCH_VERSION}-py311-cann8.5.0-${DEVICE_TYPE}-$(arch).zip \
&& ${UV_PIP_INSTALL} ./deep_ep*.whl ./sgl_kernel_npu*.whl \
&& (cd "$(python3 -m pip show deep-ep | grep -E '^Location:' | awk '{print $2}')" && ln -s deep_ep/deep_ep_cpp*.so))