[Kernel] Add expert_map support to Cutlass FP8 MOE (#16861)

Signed-off-by: varun sundar rabindranath <vsundarr@redhat.com> Co-authored-by: varun sundar rabindranath <vsundarr@redhat.com>
2025-04-21 23:44:32 -04:00
parent c9acbf1141
commit 7b8a2ab76f
5 changed files with 331 additions and 171 deletions
--- a/vllm/config.py
+++ b/vllm/config.py
@@ -1693,6 +1693,7 @@ class ParallelConfig:
        factors: list[Any] = []
        factors.append(self.pipeline_parallel_size)
        factors.append(self.tensor_parallel_size)
+        factors.append(self.enable_expert_parallel)
        return hashlib.sha256(str(factors).encode()).hexdigest()

    def __post_init__(self) -> None: