diff --git a/vllm/attention/layer.py b/vllm/attention/layer.py
index 05d997279893b..792fbd23c8428 100644
--- a/vllm/attention/layer.py
+++ b/vllm/attention/layer.py
@@ -227,7 +227,8 @@ def forward(
                                                  key,
                                                  value,
                                                  scale=self.scale)
-            out = out.transpose(1, 2)
+            # TODO (cmq): revert me after op `transpose` in torch-npu is fixed
+            out = out.transpose(1, 2).contiguous()
         return out.view(bsz, q_len, -1)
 
 
diff --git a/vllm/model_executor/models/intern_vit.py b/vllm/model_executor/models/intern_vit.py
index 7ff68bd60e8ad..1dec65dce2cbe 100644
--- a/vllm/model_executor/models/intern_vit.py
+++ b/vllm/model_executor/models/intern_vit.py
@@ -271,7 +271,8 @@ def forward(self, x: torch.Tensor) -> torch.Tensor:
         v = v.transpose(1, 2)
 
         x = F.scaled_dot_product_attention(q, k, v, scale=self.scale)
-        x = x.transpose(1, 2).view(B, N, -1)
+        # TODO (cmq): revert me after op `transpose` in torch-npu is fixed
+        x = x.transpose(1, 2).contiguous().view(B, N, -1)
 
         x = self.proj(x)
         return x