forked from flagos-ai/vllm-plugin-FL
-
Notifications
You must be signed in to change notification settings - Fork 3
123 lines (113 loc) · 3.97 KB
/
Copy path_e2e_test.yml
File metadata and controls
123 lines (113 loc) · 3.97 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
# Copyright 2026 FlagOS Contributors
#
# 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.
# Reusable E2E test workflow — slow, end-to-end model validation.
#
# Runs inference or serving tests that require real model files on disk
# and substantial GPU memory. Each invocation tests all model/case combos
# for a single (task, device) group, avoiding repeated container startup.
# Called by _platform_test.yml for each entry in the E2E test matrix.
name: E2E Test
on:
workflow_call:
inputs:
platform:
description: "Platform name (e.g., cuda, ascend)"
required: true
type: string
task:
description: "Test task category (e.g., inference, serving)"
required: true
type: string
device:
description: "Device type (e.g., a100, 910b)"
required: false
type: string
default: ""
ci_image:
description: "Docker image for the test container"
required: true
type: string
runner_labels:
description: "JSON array of runner labels"
required: true
type: string
container_volumes:
description: "JSON array of volume mounts"
required: false
type: string
default: "[]"
container_options:
description: "Docker container options string"
required: false
type: string
default: ""
timeout:
description: "Timeout for the job in minutes"
required: false
type: number
default: 60
cases:
description: "JSON array of {model,case} dicts to run (from generate_matrix). Empty string means run all cases for the task."
required: false
type: string
default: ""
jobs:
e2e-test:
name: "E2E ${{ inputs.task }} (${{ inputs.platform }}/${{ inputs.device }})"
runs-on: ${{ fromJson(inputs.runner_labels) }}
timeout-minutes: ${{ inputs.timeout }}
container:
image: ${{ inputs.ci_image }}
volumes: ${{ fromJson(inputs.container_volumes) }}
options: ${{ inputs.container_options }}
steps:
- name: Checkout (attempt 1)
id: checkout1
uses: actions/checkout@v4
continue-on-error: true
- name: Checkout (attempt 2)
id: checkout2
if: steps.checkout1.outcome == 'failure'
uses: actions/checkout@v4
continue-on-error: true
- name: Checkout (attempt 3)
if: steps.checkout2.outcome == 'failure'
uses: actions/checkout@v4
- name: Check device availability
run: bash .github/scripts/${{ inputs.platform }}/check.sh
- name: Prepare host models
run: |
model_script=".github/scripts/${{ inputs.platform }}/download_models.sh"
if [ -f "${model_script}" ]; then
bash "${model_script}"
fi
- name: Install project
run: bash .github/scripts/${{ inputs.platform }}/setup.sh
- name: Run E2E test
run: |
python tests/run.py \
--platform ${{ inputs.platform }} \
--device "${{ inputs.device }}" \
--scope e2e \
--task ${{ inputs.task }} \
--cases '${{ inputs.cases }}'
- name: Upload test results
if: always()
uses: actions/upload-artifact@v4
continue-on-error: true
with:
name: e2e-${{ inputs.platform }}-${{ inputs.device }}-${{ inputs.task }}
path: |
test-results-${{ inputs.platform }}.xml
test-results-${{ inputs.platform }}.json