[Feature] Support batch invariance on ROCm (#52231)

Signed-off-by: Matthew Wong <Matthew.Wong2@amd.com>
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-authored-by: Andreas Karatzas <akaratza@amd.com>
This commit is contained in:
Matt
2026-10-12 10:53:26 -05:00
committed by GitHub
co-authored by Claude Opus 5.5 Andreas Karatzas
parent 50b21a2667
commit e0317b28ce
31 changed files with 1106 additions and 122 deletions
@@ -180,6 +180,7 @@ def test_logprobs_bitwise_batch_invariance_bs1_vs_bsN(
llm = LLM(
model=TEST_MODEL,
tensor_parallel_size=tp_size,
disable_custom_all_reduce=os.getenv("VLLM_TEST_DISABLE_CUSTOM_AR") == "1",
max_num_seqs=128,
max_model_len=8192,
dtype="auto", # not everything is supported