[Official callbacks] SDXL Controlnet CFG Cutoff (#9311)

* initial proposal

* style

src/diffusers/callbacks.py

@@ -97,13 +97,17 @@ class SDCFGCutoffCallback(PipelineCallback):
 
 class SDXLCFGCutoffCallback(PipelineCallback):
     """
-    Callback function for Stable Diffusion XL Pipelines. After certain number of steps (set by `cutoff_step_ratio` or
-    `cutoff_step_index`), this callback will disable the CFG.
+    Callback function for the base Stable Diffusion XL Pipelines. After certain number of steps (set by
+    `cutoff_step_ratio` or `cutoff_step_index`), this callback will disable the CFG.
 
     Note: This callback mutates the pipeline by changing the `_guidance_scale` attribute to 0.0 after the cutoff step.
     """
 
-    tensor_inputs = ["prompt_embeds", "add_text_embeds", "add_time_ids"]
+    tensor_inputs = [
+        "prompt_embeds",
+        "add_text_embeds",
+        "add_time_ids",
+    ]
 
     def callback_fn(self, pipeline, step_index, timestep, callback_kwargs) -> Dict[str, Any]:
         cutoff_step_ratio = self.config.cutoff_step_ratio
@@ -129,6 +133,55 @@ class SDXLCFGCutoffCallback(PipelineCallback):
             callback_kwargs[self.tensor_inputs[0]] = prompt_embeds
             callback_kwargs[self.tensor_inputs[1]] = add_text_embeds
             callback_kwargs[self.tensor_inputs[2]] = add_time_ids
+
+        return callback_kwargs
+
+
+class SDXLControlnetCFGCutoffCallback(PipelineCallback):
+    """
+    Callback function for the Controlnet Stable Diffusion XL Pipelines. After certain number of steps (set by
+    `cutoff_step_ratio` or `cutoff_step_index`), this callback will disable the CFG.
+
+    Note: This callback mutates the pipeline by changing the `_guidance_scale` attribute to 0.0 after the cutoff step.
+    """
+
+    tensor_inputs = [
+        "prompt_embeds",
+        "add_text_embeds",
+        "add_time_ids",
+        "image",
+    ]
+
+    def callback_fn(self, pipeline, step_index, timestep, callback_kwargs) -> Dict[str, Any]:
+        cutoff_step_ratio = self.config.cutoff_step_ratio
+        cutoff_step_index = self.config.cutoff_step_index
+
+        # Use cutoff_step_index if it's not None, otherwise use cutoff_step_ratio
+        cutoff_step = (
+            cutoff_step_index if cutoff_step_index is not None else int(pipeline.num_timesteps * cutoff_step_ratio)
+        )
+
+        if step_index == cutoff_step:
+            prompt_embeds = callback_kwargs[self.tensor_inputs[0]]
+            prompt_embeds = prompt_embeds[-1:]  # "-1" denotes the embeddings for conditional text tokens.
+
+            add_text_embeds = callback_kwargs[self.tensor_inputs[1]]
+            add_text_embeds = add_text_embeds[-1:]  # "-1" denotes the embeddings for conditional pooled text tokens
+
+            add_time_ids = callback_kwargs[self.tensor_inputs[2]]
+            add_time_ids = add_time_ids[-1:]  # "-1" denotes the embeddings for conditional added time vector
+
+            # For Controlnet
+            image = callback_kwargs[self.tensor_inputs[3]]
+            image = image[-1:]
+
+            pipeline._guidance_scale = 0.0
+
+            callback_kwargs[self.tensor_inputs[0]] = prompt_embeds
+            callback_kwargs[self.tensor_inputs[1]] = add_text_embeds
+            callback_kwargs[self.tensor_inputs[2]] = add_time_ids
+            callback_kwargs[self.tensor_inputs[3]] = image
+
         return callback_kwargs
 
 

src/diffusers/pipelines/controlnet/pipeline_controlnet_sd_xl.py

@@ -242,6 +242,7 @@ class StableDiffusionXLControlNetPipeline(
         "add_time_ids",
         "negative_pooled_prompt_embeds",
         "negative_add_time_ids",
+        "image",
     ]
 
     def __init__(
@@ -1540,6 +1541,7 @@ class StableDiffusionXLControlNetPipeline(
                     )
                     add_time_ids = callback_outputs.pop("add_time_ids", add_time_ids)
                     negative_add_time_ids = callback_outputs.pop("negative_add_time_ids", negative_add_time_ids)
+                    image = callback_outputs.pop("image", image)
 
                 # call the callback, if provided
                 if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):