mirror of
https://github.com/PaddlePaddle/FastDeploy.git
synced 2025-09-26 20:41:53 +08:00
添加XPU CI, test=model (#2701)
* 添加XPU CI, test=model * 添加XPU CI, test=model * 添加XPU CI, test=model * 添加XPU CI, test=model * 添加XPU CI, test=model * 添加XPU CI, test=model * 添加XPU CI, test=model * 添加XPU CI, test=model * 添加XPU CI, test=model
This commit is contained in:
LQX
83
.github/workflows/ci_xpu.yml
vendored
Normal file
83
.github/workflows/ci_xpu.yml
vendored
Normal file
@@ -0,0 +1,83 @@
|
||||
name: CI_XPU
|
||||
|
||||
on:
|
||||
pull_request:
|
||||
branches: [ develop ]
|
||||
workflow_dispatch:
|
||||
|
||||
concurrency:
|
||||
group: ${{ github.event.pull_request.number }}-xpu-ci
|
||||
cancel-in-progress: true
|
||||
|
||||
jobs:
|
||||
build:
|
||||
runs-on: [self-hosted, XPU-P800-8Card]
|
||||
steps:
|
||||
- name: Print current runner name
|
||||
run: |
|
||||
echo "Current runner name: ${{ runner.name }}"
|
||||
# Because the system version is lower than 2.23, the checkout cannot be used.
|
||||
# - name: Checkout code
|
||||
# uses: actions/checkout@v4
|
||||
|
||||
- name: Code Checkout
|
||||
env:
|
||||
docker_image: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/fastdeploy-xpu:2.0.0
|
||||
run: |
|
||||
REPO="https://github.com/${{ github.repository }}.git"
|
||||
FULL_REPO="${{ github.repository }}"
|
||||
REPO_NAME="${FULL_REPO##*/}"
|
||||
# Clean the repository directory before starting
|
||||
docker run --rm --net=host -v $(pwd):/workspace -w /workspace \
|
||||
-e "REPO_NAME=${REPO_NAME}" \
|
||||
${docker_image} /bin/bash -c '
|
||||
if [ -d ${REPO_NAME} ]; then
|
||||
echo "Directory ${REPO_NAME} exists, removing it..."
|
||||
rm -rf ${REPO_NAME}
|
||||
fi
|
||||
'
|
||||
git config --global user.name "FastDeployCI"
|
||||
git config --global user.email "fastdeploy_ci@example.com"
|
||||
git clone ${REPO} ${REPO_NAME}
|
||||
cd FastDeploy
|
||||
if [ "${{ github.event_name }}" = "pull_request" ]; then
|
||||
git fetch origin pull/${{ github.event.pull_request.number }}/head:pr/${{ github.event.pull_request.number }}
|
||||
git merge pr/${{ github.event.pull_request.number }}
|
||||
git log -n 3 --oneline
|
||||
else
|
||||
git checkout ${{ github.sha }}
|
||||
git log -n 3 --oneline
|
||||
fi
|
||||
|
||||
- name: Run CI unittest
|
||||
env:
|
||||
docker_image: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/fastdeploy-xpu:2.0.0
|
||||
run: |
|
||||
runner_name="${{ runner.name }}"
|
||||
last_char="${runner_name: -1}"
|
||||
|
||||
if [[ "$last_char" =~ [0-3] ]]; then
|
||||
gpu_id="$last_char"
|
||||
else
|
||||
gpu_id="0"
|
||||
fi
|
||||
FD_API_PORT=$((9180 + gpu_id * 100))
|
||||
FD_ENGINE_QUEUE_PORT=$((9150 + gpu_id * 100))
|
||||
FD_METRICS_PORT=$((9170 + gpu_id * 100))
|
||||
|
||||
PARENT_DIR=$(dirname "$WORKSPACE")
|
||||
echo "PARENT_DIR:$PARENT_DIR"
|
||||
docker run --rm --net=host --cap-add=SYS_PTRACE --privileged --shm-size=64G \
|
||||
-v $(pwd):/workspace -w /workspace \
|
||||
-v "/ssd3:/ssd3" \
|
||||
-e "MODEL_PATH=/ssd3/model" \
|
||||
-e "http_proxy=$(git config --global --get http.proxy)" \
|
||||
-e "https_proxy=$(git config --global --get https.proxy)" \
|
||||
-e "FD_API_PORT=${FD_API_PORT}" \
|
||||
-e "FD_ENGINE_QUEUE_PORT=${FD_ENGINE_QUEUE_PORT}" \
|
||||
-e "FD_METRICS_PORT=${FD_METRICS_PORT}" \
|
||||
${docker_image} /bin/bash -c "
|
||||
git config --global --add safe.directory /workspace/FastDeploy
|
||||
cd FastDeploy
|
||||
bash scripts/run_ci_xpu.sh
|
||||
"
|
||||
92
scripts/run_ci_xpu.sh
Normal file
92
scripts/run_ci_xpu.sh
Normal file
@@ -0,0 +1,92 @@
|
||||
#!/bin/bash
|
||||
DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
|
||||
echo "$DIR"
|
||||
|
||||
#先kill一遍
|
||||
ps -efww | grep -E 'api_server' | grep -v grep | awk '{print $2}' | xargs kill -9 || true
|
||||
ps -efww | grep -E '8188' | grep -v grep | awk '{print $2}' | xargs kill -9 || true
|
||||
lsof -t -i :8188 | xargs kill -9 || true
|
||||
|
||||
export model_path=${MODEL_PATH}/data/eb45t_4_layer
|
||||
export CLANG_PATH=${MODEL_PATH}/data/xtdk
|
||||
export XVLLM_PATH=${MODEL_PATH}/data/xvllm
|
||||
|
||||
echo "pip requirements"
|
||||
python -m pip install -r requirements.txt
|
||||
echo "uninstall org"
|
||||
python -m pip uninstall paddlepaddle-xpu -y
|
||||
python -m pip uninstall fastdeploy-xpu -y
|
||||
python -m pip install paddlepaddle-xpu -i https://www.paddlepaddle.org.cn/packages/stable/xpu-p800/
|
||||
echo "build whl"
|
||||
bash build.sh || exit 1
|
||||
echo "pip others"
|
||||
python -m pip install openai -U
|
||||
python -m pip uninstall -y triton
|
||||
python -m pip install triton==3.3.0
|
||||
|
||||
unset http_proxy
|
||||
unset https_proxy
|
||||
unset no_proxy
|
||||
|
||||
# 起服务
|
||||
rm -rf log/*
|
||||
rm -f core*
|
||||
# pkill -9 python #流水线不执行这个
|
||||
#清空消息队列
|
||||
ipcrm --all=msg
|
||||
export XPU_VISIBLE_DEVICES="0,1,2,3"
|
||||
python -m fastdeploy.entrypoints.openai.api_server \
|
||||
--model ${model_path} \
|
||||
--port 8188 \
|
||||
--tensor-parallel-size 4 \
|
||||
--num-gpu-blocks-override 16384 \
|
||||
--max-model-len 32768 \
|
||||
--max-num-seqs 128 \
|
||||
--quantization wint4 > server.log 2>&1 &
|
||||
|
||||
sleep 60
|
||||
# 探活
|
||||
TIMEOUT=$((5 * 60))
|
||||
INTERVAL=10 # 检查间隔（秒）
|
||||
ENDPOINT="http://0.0.0.0:8188/health"
|
||||
START_TIME=$(date +%s) # 记录开始时间戳
|
||||
echo "开始服务健康检查，最长等待时间：${TIMEOUT}秒"
|
||||
while true; do
|
||||
# 计算已耗时
|
||||
CURRENT_TIME=$(date +%s)
|
||||
ELAPSED=$((CURRENT_TIME - START_TIME))
|
||||
|
||||
# 超时判断
|
||||
if [ $ELAPSED -ge $TIMEOUT ]; then
|
||||
echo -e "\n服务启动超时：经过 $((TIMEOUT/60)) 分钟服务仍未启动！"
|
||||
cat server.log
|
||||
cat log/workerlog.0
|
||||
exit 1
|
||||
fi
|
||||
|
||||
HTTP_CODE=$(curl -s -o /dev/null -w "%{http_code}" -m 2 "$ENDPOINT" || true)
|
||||
|
||||
if [ "$HTTP_CODE" = "200" ]; then
|
||||
echo -e "\n服务启动成功！耗时 ${ELAPSED} 秒"
|
||||
break
|
||||
else
|
||||
sleep $INTERVAL
|
||||
fi
|
||||
done
|
||||
|
||||
cat server.log
|
||||
|
||||
# 执行服务化推理
|
||||
python test/ci_use/XPU_45T/run_45T.py
|
||||
exit_code=$?
|
||||
echo exit_code is ${exit_code}
|
||||
|
||||
ps -efww | grep -E 'api_server' | grep -v grep | awk '{print $2}' | xargs kill -9 || true
|
||||
ps -efww | grep -E '8188' | grep -v grep | awk '{print $2}' | xargs kill -9 || true
|
||||
lsof -t -i :8188 | xargs kill -9 || true
|
||||
|
||||
if [ ${exit_code} -ne 0 ]; then
|
||||
echo "log/workerlog.0"
|
||||
cat log/workerlog.0
|
||||
exit 1
|
||||
fi
|
||||
32
test/ci_use/XPU_45T/run_45T.py
Normal file
32
test/ci_use/XPU_45T/run_45T.py
Normal file
@@ -0,0 +1,32 @@
|
||||
# Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
|
||||
#
|
||||
# Licensed under the Apache License, Version 2.0 (the "License");
|
||||
# you may not use this file except in compliance with the License.
|
||||
# You may obtain a copy of the License at
|
||||
#
|
||||
# http://www.apache.org/licenses/LICENSE-2.0
|
||||
#
|
||||
# Unless required by applicable law or agreed to in writing, software
|
||||
# distributed under the License is distributed on an "AS IS" BASIS,
|
||||
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||||
# See the License for the specific language governing permissions and
|
||||
# limitations under the License.
|
||||
|
||||
import openai
|
||||
|
||||
ip = "0.0.0.0"
|
||||
service_http_port = "8188" # 服务配置的
|
||||
client = openai.Client(base_url=f"http://{ip}:{service_http_port}/v1", api_key="EMPTY_API_KEY")
|
||||
|
||||
# 非流式对话
|
||||
response = client.chat.completions.create(
|
||||
model="default",
|
||||
messages=[
|
||||
{"role": "user", "content": "你好，你是谁？"},
|
||||
],
|
||||
temperature=1,
|
||||
top_p=0,
|
||||
max_tokens=64,
|
||||
stream=False,
|
||||
)
|
||||
print(response)
|
||||
Reference in New Issue
Block a user