[Qwen][Bugfix] Fixes sigmoid activation in torch impl of RMSNormGated.

vllm/model_executor/layers/layernorm.py

@@ -478,9 +478,12 @@ def forward_native(
         weight = self.weight.float()
         z = z.float() if z is not None else None
 
+        assert self.activation in ["silu", "sigmoid", "swish"]
+        act_fn = F.sigmoid if self.activation == "sigmoid" else F.silu
+
         # Apply gating before normalization if needed
         if z is not None and not self.norm_before_gate:
-            x = x * F.silu(z)
+            x = x * act_fn(z)
 
         # RMS Normalization
         if self.group_size is None:
@@ -499,7 +502,7 @@ def forward_native(
 
         # Apply gating after normalization if needed
         if z is not None and self.norm_before_gate:
-            out = out * F.silu(z)
+            out = out * act_fn(z)
 
         return out.to(orig_dtype)
 

vllm/model_executor/layers/mamba/gdn_linear_attn.py

@@ -357,11 +357,19 @@ def __init__(
         set_weight_attrs(self.A_log, {"weight_loader": sharded_weight_loader(0)})
         set_weight_attrs(self.dt_bias, {"weight_loader": sharded_weight_loader(0)})
 
+        output_gate_type = getattr(config, "output_gate_type", "silu")
+        if output_gate_type == "swish":
+            output_gate_type = "silu"
+        assert output_gate_type in ["silu", "swish", "sigmoid"], (
+            f"unsupported {output_gate_type=}"
+        )
+
         self.norm = RMSNormGated(
             self.head_v_dim,
             eps=self.layer_norm_epsilon,
             group_size=None,
             norm_before_gate=True,
+            activation=output_gate_type,
             device=current_platform.current_device(),
         )