Skip to content

GitLab

Unverified Commit 94866d7c authored by Jee Jee Li's avatar Jee Jee Li Committed by GitHub
Browse files

[Misc] Slight improve deepgemm print (#24085) 


Signed-off-by: default avatarJee Jee Li <pandaleefree@gmail.com>
parent 83609ca9
Show whitespace changes
Inline Side-by-side
Showing with 3 additions and 2 deletions
vllm/model_executor/layers/fused_moe/deep_gemm_moe.py
View file @ 94866d7c
...@@ -57,13 +57,14 @@ def _valid_deep_gemm(hidden_states: torch.Tensor, w1: torch.Tensor, ...@@ -57,13 +57,14 @@ def _valid_deep_gemm(hidden_states: torch.Tensor, w1: torch.Tensor,
if not _valid_deep_gemm_shape(M, N, K): if not _valid_deep_gemm_shape(M, N, K):
logger.debug_once( logger.debug_once(
"DeepGemm disabled due to unaligned problem size. " "DeepGemm disabled due to unaligned problem size. "
"M: %s, N: %s, K: %s. M should >= align size " "M: %s, N: %s, K: %s. M should >= %s "
"and N and K must be multiples of %s." "and N and K must be multiples of %s. "
"This is not an error and we will fall back to triton.", "This is not an error and we will fall back to triton.",
M, M,
N, N,
K, K,
align, align,
align,
) )
return False return False
elif N <= 512: elif N <= 512:
......
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