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 15, 11:00 UTC
[BugFix] Fix 4-GPU RLHF tests (#18007)
Signed-off-by: Nick Hill <nhill@redhat.com>
Nick Hill committed
May 12, 2025 at 23:03 UTC
ee5be834e7dcdea02512cffc125add6961f71b82
1 file changed
+2
-2
.buildkite/test-pipeline.yaml
+2
-2
@@ -158,8 +158,8 @@ steps:
158
# TODO: create a dedicated test section for multi-GPU example tests
159
# when we have multiple distributed example tests
160
- pushd ../examples/offline_inference
161
- - python3 rlhf.py
162
- - RAY_DEDUP_LOGS=0 python3 rlhf_colocate.py
161
+ - VLLM_ALLOW_INSECURE_SERIALIZATION=1 python3 rlhf.py
162
+ - VLLM_ALLOW_INSECURE_SERIALIZATION=1 RAY_DEDUP_LOGS=0 python3 rlhf_colocate.py
163
- popd
164
165
- label: Metrics, Tracing Test # 10min