A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 16, 23:15 UTC
[ci] Deprecate original CI template (#5624)
Signed-off-by: kevin <kevin@anyscale.com>
Kevin H. Luu committed
Jun 18, 2024 at 07:26 UTC
13db4369d9ab3158a01192d60c744c6523961824
2 files changed
+1
-102
.buildkite/test-pipeline.yaml
+1
-1
@@ -1,6 +1,6 @@
1
# In this file, you can add more tests to run either by adding a new step or
2
# adding a new command to an existing step. See different options here for examples.
3
-# This script will be feed into Jinja template in `test-template.j2` to generate
3
+# This script will be feed into Jinja template in `test-template-aws.j2` to generate
4
# the final pipeline yaml file.
5
6
steps:
.buildkite/test-template.j2
deleted
-101
@@ -1,101 +0,0 @@
1
-{% set docker_image = "us-central1-docker.pkg.dev/vllm-405802/vllm-ci-test-repo/vllm-test:$BUILDKITE_COMMIT" %}
2
-{% set default_num_gpu = 1 %}
3
-{% set default_working_dir = "/vllm-workspace/tests" %}
4
-
5
-steps:
6
- - label: ":docker: build image"
7
- commands:
8
- - "docker build --build-arg max_jobs=16 --tag {{ docker_image }} --target test --progress plain ."
9
- - "docker push {{ docker_image }}"
10
- env:
11
- DOCKER_BUILDKIT: "1"
12
- retry:
13
- automatic:
14
- - exit_status: -1 # Agent was lost
15
- limit: 5
16
- - exit_status: -10 # Agent was lost
17
- limit: 5
18
- - wait
19
-
20
- - group: "AMD Tests"
21
- depends_on: ~
22
- steps:
23
- {% for step in steps %}
24
- {% if step.mirror_hardwares and "amd" in step.mirror_hardwares %}
25
- - label: "AMD: {{ step.label }}"
26
- agents:
27
- queue: amd
28
- command: bash .buildkite/run-amd-test.sh "cd {{ (step.working_dir or default_working_dir) | safe }} ; {{ step.command or (step.commands | join(" ; ")) | safe }}"
29
- env:
30
- DOCKER_BUILDKIT: "1"
31
- soft_fail: true
32
- {% endif %}
33
- {% endfor %}
34
-
35
- - label: "Neuron Test"
36
- depends_on: ~
37
- agents:
38
- queue: neuron
39
- command: bash .buildkite/run-neuron-test.sh
40
- soft_fail: false
41
-
42
- - label: "Intel Test"
43
- depends_on: ~
44
- agents:
45
- queue: intel
46
- command: bash .buildkite/run-cpu-test.sh
47
-
48
- - label: "XPU Test"
49
- agents:
50
- queue: intel
51
- command: bash .buildkite/run-xpu-test.sh
52
-
53
- {% for step in steps %}
54
- - label: "{{ step.label }}"
55
- agents:
56
- queue: kubernetes
57
- soft_fail: {{ step.soft_fail or false }}
58
- {% if step.parallelism %}
59
- parallelism: {{ step.parallelism }}
60
- {% endif %}
61
- retry:
62
- automatic:
63
- - exit_status: -1 # Agent was lost
64
- limit: 5
65
- - exit_status: -10 # Agent was lost
66
- limit: 5
67
- plugins:
68
- - kubernetes:
69
- podSpec:
70
- {% if step.num_gpus %}
71
- priorityClassName: gpu-priority-cls-{{ step.num_gpus }}
72
- {% endif %}
73
- volumes:
74
- - name: dshm
75
- emptyDir:
76
- medium: Memory
77
- containers:
78
- - image: "{{ docker_image }}"
79
- command: ["bash"]
80
- args:
81
- - '-c'
82
- - "'cd {{ (step.working_dir or default_working_dir) | safe }} && {{ step.command or (step.commands | join(' && ')) | safe }}'"
83
- {% if not step.no_gpu %}
84
- resources:
85
- requests:
86
- nvidia.com/gpu: "{{ step.num_gpus or default_num_gpu }}"
87
- limits:
88
- nvidia.com/gpu: "{{ step.num_gpus or default_num_gpu }}"
89
- {% endif %}
90
- env:
91
- - name: VLLM_USAGE_SOURCE
92
- value: ci-test
93
- - name: HF_TOKEN
94
- valueFrom:
95
- secretKeyRef:
96
- name: hf-token-secret
97
- key: token
98
- volumeMounts:
99
- - mountPath: /dev/shm
100
- name: dshm
101
- {% endfor %}