[ROCm][CI] Prep Tests For Change To ROCM_ATTN As New Default Backend On ROCm (#36025)
Signed-off-by: Micah Williamson <micah.williamson@amd.com>
This commit is contained in:
@@ -24,7 +24,7 @@ if command -v rocm-smi &> /dev/null || [[ -d /opt/rocm ]] || [[ -n "${ROCM_PATH:
|
||||
BACKENDS=("allgather_reducescatter")
|
||||
# Disable MOE padding for ROCm since it is causing eplb to fail
|
||||
export VLLM_ROCM_MOE_PADDING=0
|
||||
PLATFORM_ARGS=("--no-async-scheduling")
|
||||
PLATFORM_ARGS=("--no-async-scheduling" "--attention-backend=TRITON_ATTN")
|
||||
echo "Disabled async scheduling for ROCm platform due to issues with spec decode."
|
||||
else
|
||||
# Non-ROCm platform (CUDA/other)
|
||||
|
||||
Reference in New Issue
Block a user