diff --git a/configs/models/cc12m_1024x1024.yaml b/configs/models/cc12m_1024x1024.yaml
index 2b83827..b32d352 100644
--- a/configs/models/cc12m_1024x1024.yaml
+++ b/configs/models/cc12m_1024x1024.yaml
@@ -102,6 +102,21 @@ unet_config:
   temporal_mode: false
   temporal_positional_encoding: false
   temporal_spatial_ds: false
+diffusion_config:
+  sampler_config:
+    num_diffusion_steps: 1000
+    reproject_signal: False
+    schedule_type: DEEPFLOYD
+    prediction_type: V_PREDICTION
+    loss_target_type: DDPM
+    beta_start: 0.0001
+    beta_end: 0.02
+    threshold_function: CLIP
+    rescale_schedule: 1.0
+    schedule_shifted: True
+  model_output_scale: 0.0
+  use_vdm_loss_weights: False
+  no_use_residual: true
 
 # import defaults
 # reader-config-file: configs/datasets/reader_config.yaml
diff --git a/configs/models/cc12m_256x256.yaml b/configs/models/cc12m_256x256.yaml
index 96c0236..046742c 100644
--- a/configs/models/cc12m_256x256.yaml
+++ b/configs/models/cc12m_256x256.yaml
@@ -36,7 +36,7 @@ unet_config:
   conditioning_feature_dim: -1
   conditioning_feature_proj_dim: -1
   freeze_inner_unet: false
-  initialize_inner_with_pretrained: None
+  initialize_inner_with_pretrained: null
   inner_config:
     attention_levels: [1, 2]
     conditioning_feature_dim: -1
@@ -76,6 +76,21 @@ unet_config:
   temporal_mode: false
   temporal_positional_encoding: false
   temporal_spatial_ds: false
+diffusion_config:
+  sampler_config:
+    num_diffusion_steps: 1000
+    reproject_signal: False
+    schedule_type: DEEPFLOYD
+    prediction_type: V_PREDICTION
+    loss_target_type: DDPM
+    beta_start: 0.0001
+    beta_end: 0.02
+    threshold_function: CLIP
+    rescale_schedule: 1.0
+    schedule_shifted: True
+  model_output_scale: 0.0
+  use_vdm_loss_weights: False
+  no_use_residual: true
 
 reader_config:
   image_size: 256