diff --git a/src/diffusers/models/unet_conditional.py b/src/diffusers/models/unet_conditional.py
index fe25e2baea..293542f587 100644
--- a/src/diffusers/models/unet_conditional.py
+++ b/src/diffusers/models/unet_conditional.py
@@ -65,6 +65,7 @@ class UNetConditionalModel(ModelMixin, ConfigMixin):
         center_input_sample=False,
         resnet_num_groups=30,
     ):
+        super().__init__()
         self.image_size = image_size
         time_embed_dim = block_channels[0] * 4
 
diff --git a/src/diffusers/models/unet_unconditional.py b/src/diffusers/models/unet_unconditional.py
index dcd673abd8..c809374a6f 100644
--- a/src/diffusers/models/unet_unconditional.py
+++ b/src/diffusers/models/unet_unconditional.py
@@ -61,6 +61,7 @@ class UNetUnconditionalModel(ModelMixin, ConfigMixin):
         center_input_sample=False,
         resnet_num_groups=32,
     ):
+        super().__init__()
         self.image_size = image_size
         time_embed_dim = block_channels[0] * 4
 
diff --git a/src/diffusers/models/vae.py b/src/diffusers/models/vae.py
index 58d85800f6..95b22fcd59 100644
--- a/src/diffusers/models/vae.py
+++ b/src/diffusers/models/vae.py
@@ -400,6 +400,7 @@ class VQModel(ModelMixin, ConfigMixin):
         resamp_with_conv=True,
         give_pre_end=False,
     ):
+        super().__init__()
 
         # pass init params to Encoder
         self.encoder = Encoder(
@@ -477,6 +478,7 @@ class AutoencoderKL(ModelMixin, ConfigMixin):
         resamp_with_conv=True,
         give_pre_end=False,
     ):
+        super().__init__()
 
         # pass init params to Encoder
         self.encoder = Encoder(