[V1] Enable Mamba2 layers other than MambaMixer2 in the v1 engine (#20660)

Signed-off-by: nopperl <54780682+nopperl@users.noreply.github.com>
2025-07-11 14:53:31 +09:00
parent 31d5c1797f
commit 5d09152ff1
11 changed files with 68 additions and 45 deletions
--- a/vllm/model_executor/models/granitemoehybrid.py
+++ b/vllm/model_executor/models/granitemoehybrid.py
@@ -69,8 +69,7 @@ class GraniteMoeHybridMambaDecoderLayer(nn.Module):
                                rms_norm_eps=config.rms_norm_eps,
                                activation=config.hidden_act,
                                quant_config=quant_config,
-                                prefix=f"{prefix}.mixer",
-                                chunk_size=config.mamba_chunk_size)
+                                prefix=f"{prefix}.mixer")

        self.block_sparse_moe = None
        if getattr(config, "num_local_experts", 0) > 0: