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, 15:45 UTC
[CI/Build] Re-enabling Entrypoints tests on ROCm, excluding ones that fail (#8551)
Alexey Kondratiev(AMD) committed
Sep 19, 2024 at 16:06 UTC
6cb748e190a94e20987314025614b8bd806602f2
2 files changed
+10
-1
.buildkite/run-amd-test.sh
+9
@@ -94,6 +94,15 @@ if [[ $commands == *" kernels "* ]]; then
94
--ignore=kernels/test_sampler.py"
95
fi
96
97
+#ignore certain Entrypoints tests
98
+if [[ $commands == *" entrypoints/openai "* ]]; then
99
+ commands=${commands//" entrypoints/openai "/" entrypoints/openai \
100
+ --ignore=entrypoints/openai/test_accuracy.py \
101
+ --ignore=entrypoints/openai/test_audio.py \
102
+ --ignore=entrypoints/openai/test_encoder_decoder.py \
103
+ --ignore=entrypoints/openai/test_oot_registration.py "}
104
+fi
105
+
106
PARALLEL_JOB_COUNT=8
107
# check if the command contains shard flag, we will run all shards in parallel because the host have 8 GPUs.
108
if [[ $commands == *"--shard-id="* ]]; then
.buildkite/test-pipeline.yaml
+1
-1
@@ -84,7 +84,7 @@ steps:
84
- label: Entrypoints Test # 20min
85
working_dir: "/vllm-workspace/tests"
86
fast_check: true
87
- #mirror_hardwares: [amd]
87
+ mirror_hardwares: [amd]
88
source_file_dependencies:
89
- vllm/
90
commands: