Commit ee405280 authored by zhuwenwen's avatar zhuwenwen
Browse files

set VLLM_ROCM_MOE_PADDING=0

parent 45366531
...@@ -629,7 +629,7 @@ environment_variables: dict[str, Callable[[], Any]] = { ...@@ -629,7 +629,7 @@ environment_variables: dict[str, Callable[[], Any]] = {
# Pad the weights for the moe kernel # Pad the weights for the moe kernel
"VLLM_ROCM_MOE_PADDING": "VLLM_ROCM_MOE_PADDING":
lambda: bool(int(os.getenv("VLLM_ROCM_MOE_PADDING", "1"))), lambda: bool(int(os.getenv("VLLM_ROCM_MOE_PADDING", "0"))),
# custom paged attention kernel for MI3* cards # custom paged attention kernel for MI3* cards
"VLLM_ROCM_CUSTOM_PAGED_ATTN": "VLLM_ROCM_CUSTOM_PAGED_ATTN":
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment