Add support for FLUX controlnet weight, begin_step_percent and end_step_percent.

invokeai/backend/flux/controlnet_extension.py

@@ -1,3 +1,4 @@
+import math
 from typing import List, Union
 
 import torch
@@ -55,6 +56,9 @@ class ControlNetExtension:
             resize_mode=resize_mode,
         )
 
+        # Map pixel values from [0, 1] to [-1, 1].
+        controlnet_cond = controlnet_cond * 2 - 1
+
         return cls(
             model=model,
             controlnet_cond=controlnet_cond,
@@ -65,6 +69,8 @@ class ControlNetExtension:
 
     def run_controlnet(
         self,
+        timestep_index: int,
+        total_num_timesteps: int,
         img: torch.Tensor,
         img_ids: torch.Tensor,
         txt: torch.Tensor,
@@ -72,8 +78,12 @@ class ControlNetExtension:
         y: torch.Tensor,
         timesteps: torch.Tensor,
         guidance: torch.Tensor | None,
-    ) -> list[torch.Tensor]:
-        # TODO(ryand): Handle weight, begin_step_percent, end_step_percent.
+    ) -> list[torch.Tensor] | None:
+        first_step = math.floor(self._begin_step_percent * total_num_timesteps)
+        last_step = math.ceil(self._end_step_percent * total_num_timesteps)
+        if timestep_index < first_step or timestep_index > last_step:
+            return
+        weight = self._weight
 
         controlnet_block_res_samples = self._model(
             img=img,
@@ -85,4 +95,9 @@ class ControlNetExtension:
             y=y,
             guidance=guidance,
         )
+
+        # Apply weight to the residuals.
+        for block_res_sample in controlnet_block_res_samples:
+            block_res_sample *= weight
+
         return controlnet_block_res_samples

invokeai/backend/flux/denoise.py

@@ -43,10 +43,12 @@ def denoise(
 
         # Run ControlNet models.
         # controlnet_block_residuals[i][j] is the residual of the j-th block of the i-th ControlNet model.
-        controlnet_block_residuals: list[list[torch.Tensor]] = []
+        controlnet_block_residuals: list[list[torch.Tensor] | None] = []
         for controlnet_extension in controlnet_extensions or []:
             controlnet_block_residuals.append(
                 controlnet_extension.run_controlnet(
+                    timestep_index=step - 1,
+                    total_num_timesteps=total_steps,
                     img=img,
                     img_ids=img_ids,
                     txt=txt,
@@ -65,7 +67,7 @@ def denoise(
             y=vec,
             timesteps=t_vec,
             guidance=guidance_vec,
-            block_controlnet_hidden_states=controlnet_block_residuals,
+            controlnet_block_residuals=controlnet_block_residuals,
         )
 
         preview_img = img - t_curr * pred

invokeai/backend/flux/model.py

@@ -88,7 +88,7 @@ class Flux(nn.Module):
         timesteps: Tensor,
         y: Tensor,
         guidance: Tensor | None = None,
-        block_controlnet_hidden_states: list[Tensor] | None = None,
+        controlnet_block_residuals: list[list[Tensor] | None] | None = None,
     ) -> Tensor:
         if img.ndim != 3 or txt.ndim != 3:
             raise ValueError("Input img and txt tensors must have 3 dimensions.")
@@ -109,9 +109,11 @@ class Flux(nn.Module):
         for block_index, block in enumerate(self.double_blocks):
             img, txt = block(img=img, txt=txt, vec=vec, pe=pe)
 
-            # Apply ControlNet residual.
-            if block_controlnet_hidden_states is not None:
-                img = img + block_controlnet_hidden_states[block_index % len(block_controlnet_hidden_states)]
+            # Apply ControlNet residuals.
+            if controlnet_block_residuals is not None:
+                for single_controlnet_block_residuals in controlnet_block_residuals:
+                    if single_controlnet_block_residuals:
+                        img += single_controlnet_block_residuals[block_index % len(single_controlnet_block_residuals)]
 
         img = torch.cat((txt, img), 1)
         for block in self.single_blocks: