Add missing `_version` to the MLPBlock (#6113)

* Add missing `_version` to the MLPBlock

* fix linter

torchvision/models/vision_transformer.py

@@ -40,6 +40,8 @@ class ConvStemConfig(NamedTuple):
 class MLPBlock(MLP):
     """Transformer MLP block."""
 
+    _version = 2
+
     def __init__(self, in_dim: int, mlp_dim: int, dropout: float):
         super().__init__(in_dim, [mlp_dim, in_dim], activation_layer=nn.GELU, inplace=None, dropout=dropout)