Skip to content

Commit 6e1237e

Browse files
ltoniazziLorenzo Toniazzi
ltoniazzi
authored and
Lorenzo Toniazzi
committed
Add lora test workflow
1 parent fb487bb commit 6e1237e

File tree

1 file changed

+166
-0
lines changed

1 file changed

+166
-0
lines changed

‎.github/workflows/server-convert-and-infer.yml

Lines changed: 166 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,166 @@
1+
# Server build and tests conversion from safetensors and inference
2+
name: Server conversion and inference
3+
4+
on:
5+
workflow_dispatch: # allows manual triggering
6+
inputs:
7+
sha:
8+
description: 'Commit SHA1 to build'
9+
required: false
10+
type: string
11+
slow_tests:
12+
description: 'Run slow tests'
13+
required: true
14+
type: boolean
15+
push:
16+
branches:
17+
- master
18+
paths: ['.github/workflows/server.yml', '**/CMakeLists.txt', '**/Makefile', '**/*.h', '**/*.hpp', '**/*.c', '**/*.cpp', '**/*.cu', '**/*.swift', '**/*.m', 'examples/server/**.*']
19+
pull_request:
20+
types: [opened, synchronize, reopened]
21+
paths: ['.github/workflows/server.yml', '**/CMakeLists.txt', '**/Makefile', '**/*.h', '**/*.hpp', '**/*.c', '**/*.cpp', '**/*.cu', '**/*.swift', '**/*.m', 'examples/server/**.*']
22+
23+
concurrency:
24+
group: ${{ github.workflow }}-${{ github.ref }}-${{ github.head_ref || github.run_id }}
25+
cancel-in-progress: true
26+
27+
jobs:
28+
server:
29+
runs-on: ubuntu-latest
30+
31+
strategy:
32+
matrix:
33+
sanitizer: [ADDRESS, UNDEFINED] # THREAD is broken
34+
build_type: [RelWithDebInfo]
35+
include:
36+
- build_type: Release
37+
sanitizer: ""
38+
fail-fast: false # While -DLLAMA_SANITIZE_THREAD=ON is broken
39+
40+
steps:
41+
- name: Dependencies
42+
id: depends
43+
run: |
44+
sudo apt-get update
45+
sudo apt-get -y install \
46+
build-essential \
47+
xxd \
48+
git \
49+
cmake \
50+
curl \
51+
wget \
52+
language-pack-en \
53+
libcurl4-openssl-dev
54+
55+
- name: Clone
56+
id: checkout
57+
uses: actions/checkout@v4
58+
with:
59+
fetch-depth: 0
60+
ref: ${{ github.event.inputs.sha || github.event.pull_request.head.sha || github.sha || github.head_ref || github.ref_name }}
61+
62+
- name: Python setup
63+
id: setup_python
64+
uses: actions/setup-python@v5
65+
with:
66+
python-version: '3.11'
67+
68+
- name: Tests dependencies
69+
id: test_dependencies
70+
run: |
71+
pip install -r examples/server/tests/requirements.txt
72+
73+
- name: Verify server deps
74+
id: verify_server_deps
75+
run: |
76+
git config --global --add safe.directory $(realpath .)
77+
cd examples/server
78+
git ls-files --others --modified
79+
git status
80+
./deps.sh
81+
git status
82+
not_ignored_files="$(git ls-files --others --modified)"
83+
echo "Modified files: ${not_ignored_files}"
84+
if [ -n "${not_ignored_files}" ]; then
85+
echo "Repository is dirty or server deps are not built as expected"
86+
echo "${not_ignored_files}"
87+
exit 1
88+
fi
89+
90+
- name: Build (no OpenMP)
91+
id: cmake_build_no_openmp
92+
if: ${{ matrix.sanitizer == 'THREAD' }}
93+
run: |
94+
cmake -B build \
95+
-DGGML_NATIVE=OFF \
96+
-DLLAMA_BUILD_SERVER=ON \
97+
-DLLAMA_CURL=ON \
98+
-DCMAKE_BUILD_TYPE=${{ matrix.build_type }} \
99+
-DLLAMA_SANITIZE_${{ matrix.sanitizer }}=ON \
100+
-DGGML_OPENMP=OFF ;
101+
cmake --build build --config ${{ matrix.build_type }} -j $(nproc) --target llama-server
102+
103+
- name: Build
104+
id: cmake_build
105+
if: ${{ matrix.sanitizer != 'THREAD' }}
106+
run: |
107+
cmake -B build \
108+
-DGGML_NATIVE=OFF \
109+
-DLLAMA_BUILD_SERVER=ON \
110+
-DLLAMA_CURL=ON \
111+
-DCMAKE_BUILD_TYPE=${{ matrix.build_type }} \
112+
-DLLAMA_SANITIZE_${{ matrix.sanitizer }}=ON ;
113+
cmake --build build --config ${{ matrix.build_type }} -j $(nproc) --target llama-server
114+
115+
- name: Lora convert and inference tests
116+
id: test_lora_conversion_inference
117+
if: ${{ (github.event.schedule || github.event.inputs.slow_tests == 'true') && matrix.build_type == 'Release' }}
118+
run: ./tests/test-lora-conversion-inference.sh
119+
120+
121+
server-windows:
122+
runs-on: windows-2019
123+
124+
steps:
125+
- name: Clone
126+
id: checkout
127+
uses: actions/checkout@v4
128+
with:
129+
fetch-depth: 0
130+
ref: ${{ github.event.inputs.sha || github.event.pull_request.head.sha || github.sha || github.head_ref || github.ref_name }}
131+
132+
- name: libCURL
133+
id: get_libcurl
134+
env:
135+
CURL_VERSION: 8.6.0_6
136+
run: |
137+
curl.exe -o $env:RUNNER_TEMP/curl.zip -L "https://curl.se/windows/dl-${env:CURL_VERSION}/curl-${env:CURL_VERSION}-win64-mingw.zip"
138+
mkdir $env:RUNNER_TEMP/libcurl
139+
tar.exe -xvf $env:RUNNER_TEMP/curl.zip --strip-components=1 -C $env:RUNNER_TEMP/libcurl
140+
141+
- name: Build
142+
id: cmake_build
143+
run: |
144+
cmake -B build -DLLAMA_CURL=ON -DCURL_LIBRARY="$env:RUNNER_TEMP/libcurl/lib/libcurl.dll.a" -DCURL_INCLUDE_DIR="$env:RUNNER_TEMP/libcurl/include"
145+
cmake --build build --config Release -j ${env:NUMBER_OF_PROCESSORS} --target llama-server
146+
147+
- name: Python setup
148+
id: setup_python
149+
uses: actions/setup-python@v5
150+
with:
151+
python-version: '3.11'
152+
153+
- name: Tests dependencies
154+
id: test_dependencies
155+
run: |
156+
pip install -r examples/server/tests/requirements.txt
157+
158+
- name: Copy Libcurl
159+
id: prepare_libcurl
160+
run: |
161+
cp $env:RUNNER_TEMP/libcurl/bin/libcurl-x64.dll ./build/bin/Release/libcurl-x64.dll
162+
163+
- name: Lora convert and inference tests
164+
id: test_lora_conversion_inference
165+
if: ${{ (github.event.schedule || github.event.inputs.slow_tests == 'true') && matrix.build_type == 'Release' }}
166+
run: ./tests/test-lora-conversion-inference.sh

0 commit comments

Comments
 (0)