[CI][LoRA] Drop flaky all-None batch from multi-LoRA parity test (#23287)

This commit is contained in:
jsheng_Linkedin
2026-04-20 14:43:25 -07:00
committed by GitHub
parent b65799cf83
commit 575fdc2c4c
-16
View File
@@ -672,9 +672,6 @@ def create_multiple_batch_test_samples(
prompts: List[str], lora_adapter_paths: List[str]
):
random.seed(42)
from sglang.srt.utils.common import get_bool_env_var, is_hip
_use_aiter = get_bool_env_var("SGLANG_USE_AITER") and is_hip()
test_cases = [
(
@@ -721,19 +718,6 @@ def create_multiple_batch_test_samples(
# ),
]
# [AMD] Aiter may fail this case but the model quality doesn't drop
# Skip this flaky case for now
if not _use_aiter:
test_cases.append(
(
[
random.choice(prompts),
random.choice(prompts),
random.choice(prompts),
],
[None, None, None],
)
)
return test_cases