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

[NPU] chore: bump to CANN 8.3.RC1 and Pytorch 2.8.0 (#13647)

Browse Source
This commit is contained in:
Even Zhou
2025-11-21 17:07:08 +08:00
committed by GitHub
parent a244c0309c
commit 589d9ad55b
10 changed files with 66 additions and 83 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/workflows/pr-test-npu.yml
+4 -4
View File
@@ -47,7 +47,7 @@ jobs:
if: needs.check-changes.outputs.main_package == 'true' if: needs.check-changes.outputs.main_package == 'true'
runs-on: linux-arm64-npu-1 runs-on: linux-arm64-npu-1
container: container:
image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.2.rc1-910b-ubuntu22.04-py3.11 image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc1-910b-ubuntu22.04-py3.11
steps: steps:
- name: Checkout code - name: Checkout code
uses: actions/checkout@v4 uses: actions/checkout@v4
@@ -90,7 +90,7 @@ jobs:
matrix: matrix:
part: [0, 1, 2] part: [0, 1, 2]
container: container:
image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.2.rc1-910b-ubuntu22.04-py3.11 image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc1-910b-ubuntu22.04-py3.11
steps: steps:
- name: Checkout code - name: Checkout code
uses: actions/checkout@v4 uses: actions/checkout@v4
@@ -129,7 +129,7 @@ jobs:
if: needs.check-changes.outputs.main_package == 'true' if: needs.check-changes.outputs.main_package == 'true'
runs-on: linux-arm64-npu-4 runs-on: linux-arm64-npu-4
container: container:
image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.2.rc1-910b-ubuntu22.04-py3.11 image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc1-910b-ubuntu22.04-py3.11
steps: steps:
- name: Checkout code - name: Checkout code
uses: actions/checkout@v4 uses: actions/checkout@v4
@@ -172,7 +172,7 @@ jobs:
matrix: matrix:
part: [0, 1] part: [0, 1]
container: container:
image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.2.rc1-a3-ubuntu22.04-py3.11 image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc1-a3-ubuntu22.04-py3.11
steps: steps:
- name: Checkout code - name: Checkout code
uses: actions/checkout@v4 uses: actions/checkout@v4
.github/workflows/release-docker-npu-nightly.yml
+2 -2
View File
@@ -19,7 +19,7 @@ jobs:
runs-on: ubuntu-22.04-arm runs-on: ubuntu-22.04-arm
strategy: strategy:
matrix: matrix:
cann_version: ["8.2.rc1"] cann_version: ["8.3.rc1"]
device_type: ["910b", "a3"] device_type: ["910b", "a3"]
steps: steps:
- name: Checkout repository - name: Checkout repository
@@ -73,6 +73,6 @@ jobs:
push: ${{ github.repository == 'sgl-project/sglang' && github.event_name != 'pull_request' }} push: ${{ github.repository == 'sgl-project/sglang' && github.event_name != 'pull_request' }}
provenance: false provenance: false
build-args: | build-args: |
SGLANG_KERNEL_NPU_TAG=20251110 SGLANG_KERNEL_NPU_TAG=20251120
CANN_VERSION=${{ matrix.cann_version }} CANN_VERSION=${{ matrix.cann_version }}
DEVICE_TYPE=${{ matrix.device_type }} DEVICE_TYPE=${{ matrix.device_type }}
.github/workflows/release-docker-npu.yml
+2 -2
View File
@@ -17,7 +17,7 @@ jobs:
runs-on: ubuntu-22.04-arm runs-on: ubuntu-22.04-arm
strategy: strategy:
matrix: matrix:
cann_version: ["8.2.rc1"] cann_version: ["8.3.rc1"]
device_type: ["910b", "a3"] device_type: ["910b", "a3"]
steps: steps:
- name: Checkout repository - name: Checkout repository
@@ -70,6 +70,6 @@ jobs:
push: ${{ github.repository == 'sgl-project/sglang' && github.event_name != 'pull_request' }} push: ${{ github.repository == 'sgl-project/sglang' && github.event_name != 'pull_request' }}
provenance: false provenance: false
build-args: | build-args: |
SGLANG_KERNEL_NPU_TAG=20251110 SGLANG_KERNEL_NPU_TAG=20251120
CANN_VERSION=${{ matrix.cann_version }} CANN_VERSION=${{ matrix.cann_version }}
DEVICE_TYPE=${{ matrix.device_type }} DEVICE_TYPE=${{ matrix.device_type }}
docker/npu.Dockerfile
+26 -25
View File
@@ -1,4 +1,4 @@
ARG CANN_VERSION=8.2.rc1 ARG CANN_VERSION=8.3.rc1
ARG DEVICE_TYPE=a3 ARG DEVICE_TYPE=a3
ARG OS=ubuntu22.04 ARG OS=ubuntu22.04
ARG PYTHON_VERSION=py3.11 ARG PYTHON_VERSION=py3.11
@@ -6,20 +6,22 @@ ARG PYTHON_VERSION=py3.11
FROM quay.io/ascend/cann:$CANN_VERSION-$DEVICE_TYPE-$OS-$PYTHON_VERSION FROM quay.io/ascend/cann:$CANN_VERSION-$DEVICE_TYPE-$OS-$PYTHON_VERSION
# Update pip & apt sources # Update pip & apt sources
ARG DEVICE_TYPE
ARG PIP_INDEX_URL="https://pypi.org/simple/" ARG PIP_INDEX_URL="https://pypi.org/simple/"
ARG APTMIRROR="" ARG APTMIRROR=""
ARG MEMFABRIC_URL=https://sglang-ascend.obs.cn-east-3.myhuaweicloud.com/sglang/mf_adapter-1.0.0-cp311-cp311-linux_aarch64.whl ARG PYTORCH_VERSION="2.8.0"
ARG PYTORCH_VERSION=2.6.0 ARG TORCHVISION_VERSION="0.23.0"
ARG TORCHVISION_VERSION=0.21.0 ARG PTA_VERSION="v7.2.0-pytorch${PYTORCH_VERSION}"
ARG PTA_URL="https://sglang-ascend.obs.cn-east-3.myhuaweicloud.com/ops/torch_npu-2.6.0.post2%2Bgit95d6260-cp311-cp311-linux_aarch64.whl" ARG PTA_NAME="torch_npu-${PYTORCH_VERSION}-cp311-cp311-manylinux_2_28_aarch64.whl"
ARG VLLM_TAG=v0.8.5 ARG PTA_URL="https://gitcode.com/Ascend/pytorch/releases/download/${PTA_VERSION}/${PTA_NAME}"
ARG TRITON_ASCEND_URL="https://sglang-ascend.obs.cn-east-3.myhuaweicloud.com/sglang/triton_ascend-3.2.0%2Bgitb0ea0850-cp311-cp311-linux_aarch64.whl" ARG TRITON_ASCEND_URL="https://sglang-ascend.obs.cn-east-3.myhuaweicloud.com/sglang/triton_ascend-3.2.0%2Bgitb0ea0850-cp311-cp311-linux_aarch64.whl"
ARG BISHENG_URL="https://sglang-ascend.obs.cn-east-3.myhuaweicloud.com/sglang/Ascend-BiSheng-toolkit_aarch64.run" ARG BISHENG_URL="https://sglang-ascend.obs.cn-east-3.myhuaweicloud.com/sglang/Ascend-BiSheng-toolkit_aarch64.run"
ARG SGLANG_TAG=main ARG SGLANG_TAG=main
ARG ASCEND_CANN_PATH=/usr/local/Ascend/ascend-toolkit ARG ASCEND_CANN_PATH=/usr/local/Ascend/ascend-toolkit
ARG SGLANG_KERNEL_NPU_TAG=main ARG SGLANG_KERNEL_NPU_TAG=main
ARG PIP_INSTALL="python3 -m pip install --no-cache-dir"
ARG DEVICE_TYPE
WORKDIR /workspace WORKDIR /workspace
# Define environments # Define environments
@@ -54,45 +56,44 @@ ENV LANG=en_US.UTF-8
ENV LANGUAGE=en_US:en ENV LANGUAGE=en_US:en
ENV LC_ALL=en_US.UTF-8 ENV LC_ALL=en_US.UTF-8
# Install dependencies
# TODO: install from pypi released memfabric
RUN pip install $MEMFABRIC_URL --no-cache-dir
# Install vLLM ### Install MemFabric
RUN git clone --depth 1 https://github.com/vllm-project/vllm.git --branch $VLLM_TAG && \ RUN ${PIP_INSTALL} mf-adapter==1.0.0
(cd vllm && VLLM_TARGET_DEVICE="empty" pip install -v . --no-cache-dir) && rm -rf vllm ### Install SGLang Model Gateway
RUN ${PIP_INSTALL} sglang-router
### Install PyTorch and PTA
RUN (${PIP_INSTALL} torch==${PYTORCH_VERSION} torchvision==${TORCHVISION_VERSION} --index-url https://download.pytorch.org/whl/cpu) && \
(wget -O "${PTA_NAME}" "${PTA_URL}" && ${PIP_INSTALL} "./${PTA_NAME}" && rm "./${PTA_NAME}")
# TODO: install from pypi released triton-ascend # TODO: install from pypi released triton-ascend
RUN pip install torch==$PYTORCH_VERSION torchvision==$TORCHVISION_VERSION --index-url https://download.pytorch.org/whl/cpu --no-cache-dir \ RUN ${PIP_INSTALL} attrs==24.2.0 numpy==1.26.4 scipy==1.13.1 decorator==5.1.1 psutil==6.0.0 pytest==8.3.2 pytest-xdist==3.6.1 pyyaml pybind11 && \
&& wget ${PTA_URL} && pip install "./torch_npu-2.6.0.post2+git95d6260-cp311-cp311-linux_aarch64.whl" --no-cache-dir \ ${PIP_INSTALL} ${TRITON_ASCEND_URL}
&& python3 -m pip install --no-cache-dir attrs==24.2.0 numpy==1.26.4 scipy==1.13.1 decorator==5.1.1 psutil==6.0.0 pytest==8.3.2 pytest-xdist==3.6.1 pyyaml pybind11 \
&& pip install ${TRITON_ASCEND_URL} --no-cache-dir
# Install SGLang # Install SGLang
RUN git clone https://github.com/sgl-project/sglang --branch $SGLANG_TAG && \ RUN git clone https://github.com/sgl-project/sglang --branch $SGLANG_TAG && \
(cd sglang/python && rm -rf pyproject.toml && mv pyproject_other.toml pyproject.toml && pip install -v .[srt_npu] --no-cache-dir) && \ (cd sglang/python && rm -rf pyproject.toml && mv pyproject_other.toml pyproject.toml && ${PIP_INSTALL} -v .[srt_npu]) && \
rm -rf sglang rm -rf sglang
# Install SGLang Model Gateway
RUN pip install sglang-router --no-cache-dir
# Install Deep-ep # Install Deep-ep
# pin wheel to 0.45.1 ref: https://github.com/pypa/wheel/issues/662 # pin wheel to 0.45.1 ref: https://github.com/pypa/wheel/issues/662
RUN pip install wheel==0.45.1 && git clone --branch $SGLANG_KERNEL_NPU_TAG https://github.com/sgl-project/sgl-kernel-npu.git \ RUN ${PIP_INSTALL} wheel==0.45.1 && git clone --branch $SGLANG_KERNEL_NPU_TAG https://github.com/sgl-project/sgl-kernel-npu.git \
&& export LD_LIBRARY_PATH=${ASCEND_CANN_PATH}/latest/runtime/lib64/stub:$LD_LIBRARY_PATH && \ && export LD_LIBRARY_PATH=${ASCEND_CANN_PATH}/latest/runtime/lib64/stub:$LD_LIBRARY_PATH && \
source ${ASCEND_CANN_PATH}/set_env.sh && \ source ${ASCEND_CANN_PATH}/set_env.sh && \
cd sgl-kernel-npu && \ cd sgl-kernel-npu && \
bash build.sh \ bash build.sh \
&& pip install output/deep_ep*.whl output/sgl_kernel_npu*.whl --no-cache-dir \ && ${PIP_INSTALL} output/deep_ep*.whl output/sgl_kernel_npu*.whl \
&& cd .. && rm -rf sgl-kernel-npu \ && cd .. && rm -rf sgl-kernel-npu \
&& cd "$(pip show deep-ep | awk '/^Location:/ {print $2}')" && ln -s deep_ep/deep_ep_cpp*.so && cd "$(python3 -m pip show deep-ep | awk '/^Location:/ {print $2}')" && ln -s deep_ep/deep_ep_cpp*.so
# Install CustomOps # Install CustomOps
RUN wget https://sglang-ascend.obs.cn-east-3.myhuaweicloud.com/ops/CANN-custom_ops-8.2.0.0-$DEVICE_TYPE-linux.aarch64.run && \ RUN wget https://sglang-ascend.obs.cn-east-3.myhuaweicloud.com/ops/CANN-custom_ops-8.2.0.0-$DEVICE_TYPE-linux.aarch64.run && \
chmod a+x ./CANN-custom_ops-8.2.0.0-$DEVICE_TYPE-linux.aarch64.run && \ chmod a+x ./CANN-custom_ops-8.2.0.0-$DEVICE_TYPE-linux.aarch64.run && \
./CANN-custom_ops-8.2.0.0-$DEVICE_TYPE-linux.aarch64.run --quiet --install-path=/usr/local/Ascend/ascend-toolkit/latest/opp && \ ./CANN-custom_ops-8.2.0.0-$DEVICE_TYPE-linux.aarch64.run --quiet --install-path=/usr/local/Ascend/ascend-toolkit/latest/opp && \
wget https://sglang-ascend.obs.cn-east-3.myhuaweicloud.com/ops/custom_ops-1.0.$DEVICE_TYPE-cp311-cp311-linux_aarch64.whl && \ wget https://sglang-ascend.obs.cn-east-3.myhuaweicloud.com/ops/custom_ops-1.0.$DEVICE_TYPE-cp311-cp311-linux_aarch64.whl && \
pip install ./custom_ops-1.0.$DEVICE_TYPE-cp311-cp311-linux_aarch64.whl ${PIP_INSTALL} ./custom_ops-1.0.$DEVICE_TYPE-cp311-cp311-linux_aarch64.whl
# Install Bisheng # Install Bisheng
RUN wget ${BISHENG_URL} && chmod a+x Ascend-BiSheng-toolkit_aarch64.run && ./Ascend-BiSheng-toolkit_aarch64.run --install && rm Ascend-BiSheng-toolkit_aarch64.run RUN wget ${BISHENG_URL} && chmod a+x Ascend-BiSheng-toolkit_aarch64.run && ./Ascend-BiSheng-toolkit_aarch64.run --install && rm Ascend-BiSheng-toolkit_aarch64.run
docs/platforms/ascend_npu.md
+7 -24
View File
@@ -48,41 +48,23 @@ conda activate sglang_npu
#### MemFabric Adaptor #### MemFabric Adaptor
_TODO: MemFabric is still a working project yet open sourced til August/September, 2025. We will release it as prebuilt wheel package for now._ _TODO: MemFabric is still a working project yet open sourced til end of year 2025. We will release it as prebuilt wheel package for now._
_Notice: Prebuilt wheel package is based on `aarch64`, please leave an issue [here at sglang](https://github.com/sgl-project/sglang/issues) to let us know the requests for `amd64` build._
MemFabric Adaptor is a drop-in replacement of Mooncake Transfer Engine that enables KV cache transfer on Ascend NPU clusters. MemFabric Adaptor is a drop-in replacement of Mooncake Transfer Engine that enables KV cache transfer on Ascend NPU clusters.
```shell ```shell
MF_WHL_NAME="mf_adapter-1.0.0-cp311-cp311-linux_aarch64.whl" pip install mf-adapter==1.0.0
MEMFABRIC_URL="https://sglang-ascend.obs.cn-east-3.myhuaweicloud.com/sglang/${MF_WHL_NAME}"
wget -O "${MF_WHL_NAME}" "${MEMFABRIC_URL}" && pip install "./${MF_WHL_NAME}"
``` ```
#### Pytorch and Pytorch Framework Adaptor on Ascend #### Pytorch and Pytorch Framework Adaptor on Ascend
Only `torch==2.6.0` is supported currently due to NPUgraph and Triton-on-Ascend's limitation, however a more generalized version will be release by the end of September, 2025.
```shell ```shell
PYTORCH_VERSION=2.6.0 PYTORCH_VERSION="2.8.0"
TORCHVISION_VERSION=0.21.0 TORCHVISION_VERSION="0.23.0"
pip install torch==$PYTORCH_VERSION torchvision==$TORCHVISION_VERSION --index-url https://download.pytorch.org/whl/cpu pip install torch==$PYTORCH_VERSION torchvision==$TORCHVISION_VERSION --index-url https://download.pytorch.org/whl/cpu
PTA_VERSION="v7.1.0.1-pytorch2.6.0" PTA_VERSION="2.8.0"
PTA_NAME="torch_npu-2.6.0.post1-cp311-cp311-manylinux_2_28_aarch64.whl" pip install torch-npu==$PTA_VERSION
PTA_URL="https://gitee.com/ascend/pytorch/releases/download/${PTA_VERSION}/${PTA_WHL_NAME}"
wget -O "${PTA_NAME}" "${PTA_URL}" && pip install "./${PTA_NAME}"
```
#### vLLM
vLLM is still a major prerequisite on Ascend NPU. Because of `torch==2.6.0` limitation, only vLLM v0.8.5 is supported.
```shell
VLLM_TAG=v0.8.5
git clone --depth 1 https://github.com/vllm-project/vllm.git --branch $VLLM_TAG
(cd vllm && VLLM_TARGET_DEVICE="empty" pip install -v -e .)
``` ```
#### Triton on Ascend #### Triton on Ascend
@@ -103,6 +85,7 @@ git clone -b v0.5.5.post3 https://github.com/sgl-project/sglang.git
cd sglang cd sglang
pip install --upgrade pip pip install --upgrade pip
rm -vf python/pyproject.toml && mv python/pyproject_other.toml python/pyproject.toml
pip install -e python[srt_npu] pip install -e python[srt_npu]
``` ```
python/pyproject_other.toml
+1
View File
@@ -27,6 +27,7 @@ runtime_common = [
"datasets", "datasets",
"einops", "einops",
"fastapi", "fastapi",
"gguf",
"hf_transfer", "hf_transfer",
"huggingface_hub", "huggingface_hub",
"interegular", "interegular",
python/sglang/check_env.py
+7 -1
View File
@@ -311,9 +311,15 @@ class HIPEnv(BaseEnv):
class NPUEnv(BaseEnv): class NPUEnv(BaseEnv):
"""Environment checker for Ascend NPU""" """Environment checker for Ascend NPU"""
EXTRA_PACKAGE_LIST = [
"torch_npu",
"sgl-kernel-npu",
"deep_ep",
]
def __init__(self): def __init__(self):
super().__init__() super().__init__()
self.package_list = ["torch_npu", "sgl-kernel-npu"] + self.package_list self.package_list.extend(NPUEnv.EXTRA_PACKAGE_LIST)
def get_info(self): def get_info(self):
cuda_info = {"NPU available": torch.npu.is_available()} cuda_info = {"NPU available": torch.npu.is_available()}
python/sglang/srt/layers/attention/ascend_backend.py
+1 -1
View File
@@ -376,7 +376,7 @@ class AscendAttnBackend(AttentionBackend):
num_key_value_heads=layer.tp_k_head_num, num_key_value_heads=layer.tp_k_head_num,
input_layout="BSND", # todo, TND not supports q_heads!=k_heads input_layout="BSND", # todo, TND not supports q_heads!=k_heads
atten_mask=self.fia_mask.unsqueeze(0), atten_mask=self.fia_mask.unsqueeze(0),
sparse_mode=3, sparse_mode=3 if q_len != 1 else 0,
scale=layer.scaling, scale=layer.scaling,
next_tokens=0, next_tokens=0,
)[0] )[0]
python/sglang/test/test_utils.py
+2 -2
View File
@@ -1245,8 +1245,8 @@ def run_bench_offline_throughput(model, other_args):
try: try:
stdout, stderr = process.communicate() stdout, stderr = process.communicate()
output = stdout.decode() output = stdout.decode(errors="backslashreplace")
error = stderr.decode() error = stderr.decode(errors="backslashreplace")
print(f"Output: {output}", flush=True) print(f"Output: {output}", flush=True)
print(f"Error: {error}", flush=True) print(f"Error: {error}", flush=True)
scripts/ci/npu_ci_install_dependency.sh
+14 -22
View File
@@ -1,7 +1,7 @@
#!/bin/bash #!/bin/bash
set -euo pipefail set -euo pipefail
PIP_INSTALL="pip install --no-cache-dir" PIP_INSTALL="python3 -m pip install --no-cache-dir"
DEVICE_TYPE=$1 DEVICE_TYPE=$1
@@ -19,29 +19,23 @@ apt update -y && apt install -y \
ccache \ ccache \
ca-certificates ca-certificates
update-ca-certificates update-ca-certificates
python3 -m ${PIP_INSTALL} --upgrade pip ${PIP_INSTALL} --upgrade pip
# Pin wheel to 0.45.1, REF: https://github.com/pypa/wheel/issues/662
${PIP_INSTALL} wheel==0.45.1
### Download MemFabricV2 ### Install MemFabric
MF_WHL_NAME="mf_adapter-1.0.0-cp311-cp311-linux_aarch64.whl" ${PIP_INSTALL} mf-adapter==1.0.0
MEMFABRIC_URL="https://sglang-ascend.obs.cn-east-3.myhuaweicloud.com/sglang/${MF_WHL_NAME}"
wget -O "${MF_WHL_NAME}" "${MEMFABRIC_URL}" && ${PIP_INSTALL} "./${MF_WHL_NAME}"
### Install vLLM
VLLM_TAG=v0.8.5
git clone --depth 1 https://github.com/vllm-project/vllm.git --branch $VLLM_TAG
(cd vllm && VLLM_TARGET_DEVICE="empty" ${PIP_INSTALL} -v -e .)
### Install PyTorch and PTA ### Install PyTorch and PTA
PYTORCH_VERSION=2.6.0 PYTORCH_VERSION="2.8.0"
TORCHVISION_VERSION=0.21.0 TORCHVISION_VERSION="0.23.0"
${PIP_INSTALL} torch==$PYTORCH_VERSION torchvision==$TORCHVISION_VERSION --index-url https://download.pytorch.org/whl/cpu ${PIP_INSTALL} torch==${PYTORCH_VERSION} torchvision==${TORCHVISION_VERSION} --index-url https://download.pytorch.org/whl/cpu
PTA_VERSION="v7.1.0.1-pytorch2.6.0" PTA_VERSION="v7.2.0-pytorch${PYTORCH_VERSION}"
PTA_NAME="torch_npu-2.6.0.post2+git95d6260-cp311-cp311-linux_aarch64.whl" PTA_NAME="torch_npu-${PYTORCH_VERSION}-cp311-cp311-manylinux_2_28_aarch64.whl"
PTA_URL="https://sglang-ascend.obs.cn-east-3.myhuaweicloud.com/ops/torch_npu-2.6.0.post2%2Bgit95d6260-cp311-cp311-linux_aarch64.whl" PTA_URL="https://gitcode.com/Ascend/pytorch/releases/download/${PTA_VERSION}/${PTA_NAME}"
wget -O "${PTA_NAME}" "${PTA_URL}" && ${PIP_INSTALL} "./${PTA_NAME}" wget -O "${PTA_NAME}" "${PTA_URL}" && ${PIP_INSTALL} "./${PTA_NAME}"
@@ -59,11 +53,9 @@ wget -O "${BISHENG_NAME}" "${BISHENG_URL}" && chmod a+x "${BISHENG_NAME}" && "./
### Install sgl-kernel-npu ### Install sgl-kernel-npu
SGL_KERNEL_NPU_TAG="20251110" SGL_KERNEL_NPU_TAG="20251120"
git clone --depth 1 https://github.com/sgl-project/sgl-kernel-npu.git --branch ${SGL_KERNEL_NPU_TAG} git clone --depth 1 https://github.com/sgl-project/sgl-kernel-npu.git --branch ${SGL_KERNEL_NPU_TAG}
# pin wheel to 0.45.1 ref: https://github.com/pypa/wheel/issues/662 (cd sgl-kernel-npu && bash ./build.sh && ${PIP_INSTALL} output/deep_ep*.whl output/sgl_kernel_npu*.whl && cd "$(python3 -m pip show deep-ep | grep -E '^Location:' | awk '{print $2}')" && ln -s deep_ep/deep_ep_cpp*.so)
pip install wheel==0.45.1
(cd sgl-kernel-npu && bash ./build.sh && pip install output/deep_ep*.whl output/sgl_kernel_npu*.whl && cd "$(pip show deep-ep | grep -E '^Location:' | awk '{print $2}')" && ln -s deep_ep/deep_ep_cpp*.so)
### Install CustomOps (TODO: to be removed once merged into sgl-kernel-npu) ### Install CustomOps (TODO: to be removed once merged into sgl-kernel-npu)