[BugFix] Fix false assertion with spec-decode=[2,4,..] and TP>2 (#29036)

Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com> (cherry picked from commit 8f4f77a727)
2025-11-19 16:43:54 -05:00
parent fa3ffb4365
commit 275de34170
1 changed files with 1 additions and 1 deletions
--- a/vllm/config/compilation.py
+++ b/vllm/config/compilation.py
@@ -917,7 +917,7 @@ class CompilationConfig:
        self, uniform_decode_query_len: int, tensor_parallel_size: int
    ):
        multiple_of = uniform_decode_query_len
-        if tensor_parallel_size > 1:
+        if tensor_parallel_size > 1 and self.pass_config.enable_sequence_parallelism:
            multiple_of = max(uniform_decode_query_len, tensor_parallel_size)
            if (
                multiple_of % uniform_decode_query_len != 0