423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2026-08-17 13:35:44 +08:00
Code Issues Packages Projects Releases Wiki Activity

[docker] upgrade NPU images to CANN 9.1 and PyTorch 2.10 (#10729)

Browse Source
This commit is contained in:
xvxuopop
2026-08-10 11:20:41 +08:00
committed by GitHub
parent 63a89710c7
commit 0bbe481e6e
10 changed files with 297 additions and 384 deletions
Download Patch File Download Diff File Expand all files Collapse all files

45
.github/workflows/docker_npu.yml vendored
View File

@@ -17,16 +17,16 @@ jobs:
include:
- device: "npu-a2"
os: "ubuntu"
base_image: "quay.io/ascend/cann:9.0.0-910b-ubuntu22.04-py3.11"
base_image: "quay.io/ascend/cann:9.1.0-910b-ubuntu22.04-py3.12"
- device: "npu-a3"
os: "ubuntu"
base_image: "quay.io/ascend/cann:9.0.0-a3-ubuntu22.04-py3.11"
base_image: "quay.io/ascend/cann:9.1.0-a3-ubuntu22.04-py3.12"
- device: "npu-a2"
os: "openeuler"
base_image: "quay.io/ascend/cann:9.0.0-910b-openeuler24.03-py3.11"
base_image: "quay.io/ascend/cann:9.1.0-910b-openeuler24.03-py3.12"
- device: "npu-a3"
os: "openeuler"
base_image: "quay.io/ascend/cann:9.0.0-a3-openeuler24.03-py3.11"
base_image: "quay.io/ascend/cann:9.1.0-a3-openeuler24.03-py3.12"
runs-on: ubuntu-latest
@@ -48,7 +48,7 @@ jobs:
- name: Checkout
uses: actions/checkout@v6
- name: Get llamafactory version
- name: Get LlamaFactory version
id: version
run: |
if [ "${{ github.event_name }}" = "release" ]; then
@@ -61,29 +61,46 @@ jobs:
id: npu_tag
env:
BASE_IMAGE: ${{ matrix.base_image }}
DEVICE: ${{ matrix.device }}
MATRIX_DEVICE: ${{ matrix.device }}
MATRIX_OS: ${{ matrix.os }}
LLAMAFACTORY_VERSION: ${{ steps.version.outputs.tag }}
run: |
base_image_tag="${BASE_IMAGE##*:}"
cann_version="${base_image_tag%%-*}"
torch_npu_version="$(sed -nE 's/^torch[-_]npu==([0-9]+(\.[0-9]+)*).*/\1/p' requirements/npu.txt)"
accelerator="${DEVICE#npu-}"
accelerator="${accelerator^^}"
operating_system="$(grep -oE '(ubuntu|openeuler)' <<< "${base_image_tag}" | head -n 1)"
torch_npu_version="$(sed -nE 's/^torch[-_]npu==([0-9]+(\.[0-9]+)*(\.post[0-9]+)?).*/\1/p' requirements/npu.txt)"
soc="$(grep -oE '(910b|a3)' <<< "${base_image_tag}" | head -n 1)"
operating_system="$(grep -oE '(ubuntu|openeuler)[0-9]+(\.[0-9]+)*' <<< "${base_image_tag}" | head -n 1)"
python_version="$(grep -oE 'py[0-9]+\.[0-9]+' <<< "${base_image_tag}" | head -n 1)"
if [[ -z "${cann_version}" || -z "${torch_npu_version}" || -z "${operating_system}" || -z "${python_version}" ]]; then
if [[ -z "${cann_version}" || -z "${torch_npu_version}" || -z "${soc}" || -z "${operating_system}" || -z "${python_version}" ]]; then
echo "Failed to derive the NPU image tag from ${BASE_IMAGE} and requirements/npu.txt" >&2
exit 1
fi
if [[ "${operating_system}" != "${MATRIX_OS}" ]]; then
echo "Operating system ${operating_system} derived from ${BASE_IMAGE} does not match matrix OS ${MATRIX_OS}" >&2
if [[ "${operating_system}" != "${MATRIX_OS}"* ]]; then
echo "Operating system ${operating_system} derived from ${BASE_IMAGE} does not match matrix OS family ${MATRIX_OS}" >&2
exit 1
fi
echo "tag=${LLAMAFACTORY_VERSION}-cann${cann_version}-torch_npu${torch_npu_version}-${accelerator}-${operating_system}-${python_version}" >> "$GITHUB_OUTPUT"
case "${MATRIX_DEVICE}" in
npu-a2) expected_soc="910b" ;;
npu-a3) expected_soc="a3" ;;
*)
echo "Unsupported NPU device ${MATRIX_DEVICE}" >&2
exit 1
;;
esac
if [[ "${soc}" != "${expected_soc}" ]]; then
echo "SoC ${soc} derived from ${BASE_IMAGE} does not match matrix device ${MATRIX_DEVICE}" >&2
exit 1
fi
if [[ "${LLAMAFACTORY_VERSION}" == "latest" ]]; then
echo "tag=latest-${soc}-${MATRIX_OS}" >> "$GITHUB_OUTPUT"
else
echo "tag=${LLAMAFACTORY_VERSION}-cann${cann_version}-torch_npu${torch_npu_version}-${soc}-${operating_system}-${python_version}" >> "$GITHUB_OUTPUT"
fi
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v3

6
.github/workflows/tests_npu.yml vendored
View File

@@ -25,11 +25,11 @@ jobs:
fail-fast: false
matrix:
python:
- "3.11"
- "3.12"
os:
- "linux-aarch64-a2-4"
pytorch_npu:
- "2.7.1"
- "2.10.0"
runs-on: ${{ matrix.os }}
@@ -38,7 +38,7 @@ jobs:
cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
container:
image: ascendai/cann:9.0.0-910b-ubuntu22.04-py3.11
image: ascendai/cann:9.1.0-910b-ubuntu22.04-py3.12
env:
HF_ENDPOINT: https://hf-mirror.com
HF_TOKEN: ${{ secrets.HF_TOKEN }}