Handle the "medium" and "small" SDXL controlnet models

invokeai/backend/util/hotfixes.py

@@ -98,6 +98,7 @@ class ControlNetModel(ModelMixin, ConfigMixin):
         norm_num_groups: Optional[int] = 32,
         norm_eps: float = 1e-5,
         cross_attention_dim: int = 1280,
+        transformer_layers_per_block: Union[int, Tuple[int]] = 1,
         attention_head_dim: Union[int, Tuple[int]] = 8,
         num_attention_heads: Optional[Union[int, Tuple[int]]] = None,
         use_linear_projection: bool = False,
@@ -135,6 +136,8 @@ class ControlNetModel(ModelMixin, ConfigMixin):
             raise ValueError(
                 f"Must provide the same number of `num_attention_heads` as `down_block_types`. `num_attention_heads`: {num_attention_heads}. `down_block_types`: {down_block_types}."
             )
+        if isinstance(transformer_layers_per_block, int):
+            transformer_layers_per_block = [transformer_layers_per_block] * len(down_block_types)
 
         # input
         conv_in_kernel = 3
@@ -212,6 +215,7 @@ class ControlNetModel(ModelMixin, ConfigMixin):
             down_block = get_down_block(
                 down_block_type,
                 num_layers=layers_per_block,
+                transformer_layers_per_block=transformer_layers_per_block[i],
                 in_channels=input_channel,
                 out_channels=output_channel,
                 temb_channels=time_embed_dim,
@@ -248,6 +252,7 @@ class ControlNetModel(ModelMixin, ConfigMixin):
         self.controlnet_mid_block = controlnet_block
 
         self.mid_block = UNetMidBlock2DCrossAttn(
+            transformer_layers_per_block=transformer_layers_per_block[-1],
             in_channels=mid_block_channel,
             temb_channels=time_embed_dim,
             resnet_eps=norm_eps,