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 17, 03:00 UTC
[CI] Add test case to run examples scripts (#3638)
Simon Mo committed
Mar 28, 2024 at 14:36 UTC
a4075cba4de42470565737f668c8cf7512fbe935
2 files changed
+19
-3
.buildkite/test-pipeline.yaml
+10
@@ -36,6 +36,16 @@ steps:
36
- label: Entrypoints Test
37
command: pytest -v -s entrypoints
38
39
+- label: Examples Test
40
+ working_dir: "/vllm-workspace/examples"
41
+ commands:
42
+ # install aws cli for llava_example.py
43
+ - pip install awscli
44
+ - python3 offline_inference.py
45
+ - python3 offline_inference_with_prefix.py
46
+ - python3 llm_engine_example.py
47
+ - python3 llava_example.py
48
+
49
- label: Kernels Test %N
50
command: pytest -v -s kernels --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
51
parallelism: 4
examples/llava_example.py
+9
-3
@@ -78,7 +78,13 @@ if __name__ == "__main__":
78
# Make sure the local directory exists or create it
79
os.makedirs(local_directory, exist_ok=True)
80
81
- # Use AWS CLI to sync the directory
82
- subprocess.check_call(
83
- ["aws", "s3", "sync", s3_bucket_path, local_directory])
81
+ # Use AWS CLI to sync the directory, assume anonymous access
82
+ subprocess.check_call([
83
+ "aws",
84
+ "s3",
85
+ "sync",
86
+ s3_bucket_path,
87
+ local_directory,
88
+ "--no-sign-request",
89
+ ])
90
main(args)