fix: fix rmsnorm -> layernorm in qwen3 omni (sgl-project#11791)

Co-authored-by: Brayden Zhong <b8zhong@users.noreply.github.com>

Author: 2 people

python/sglang/srt/models/qwen3_omni_moe.py

@@ -31,7 +31,6 @@
 )
 from sglang.srt.configs.qwen3_vl import Qwen3VLMoeConfig
 from sglang.srt.layers.attention.vision import VisionAttention
-from sglang.srt.layers.layernorm import RMSNorm
 from sglang.srt.layers.linear import ColumnParallelLinear, RowParallelLinear
 from sglang.srt.layers.moe.fused_moe_triton.layer import FusedMoE
 from sglang.srt.layers.quantization.base_config import QuantizationConfig
@@ -318,7 +317,7 @@ def __init__(
         super().__init__()
         self.hidden_size = context_dim * (spatial_merge_size**2)
         self.use_postshuffle_norm = use_postshuffle_norm
-        self.ln_q = RMSNorm(
+        self.ln_q = nn.LayerNorm(
             self.hidden_size if use_postshuffle_norm else context_dim, eps=1e-6
         )
         self.mlp = nn.ModuleList(