[Misc] Fix Current vLLM config is not set. warnings, assert to avoid issues in the future (#31747)

Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
This commit is contained in:
Lucas Wilkinson
2026-01-08 15:20:49 -08:00
committed by GitHub
co-authored by Luka Govedič
parent 5d3b6097ad
commit 6cdf015c3c
48 changed files with 380 additions and 240 deletions
@@ -124,7 +124,12 @@ def _quantize_dequantize_fp8_ds_mla(
reason="FlashMLASparseBackend requires CUDA 9.0 or higher",
)
def test_sparse_backend_decode_correctness(
dist_init, batch_name, kv_cache_dtype, tensor_parallel_size, workspace_init
default_vllm_config,
dist_init,
batch_name,
kv_cache_dtype,
tensor_parallel_size,
workspace_init,
):
if current_platform.is_rocm():
pytest.skip("ROCm does not support fp8_ds_mla data type for kv cache.")