Modified ControlNet/T2IAdapter get_control function to receive transf…

…ormer_options as additional parameter, made the model_options stored in extra_args in inner_sample be a clone of the original model_options instead of same ref
comfyanonymous · Sep 25, 2024 · fd2d572 · fd2d572
1 parent d3229cb
commit fd2d572
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 8 deletions.
diff --git a/comfy/controlnet.py b/comfy/controlnet.py
@@ -201,10 +201,10 @@ def __init__(self, control_model=None, global_average_pooling=False, compression
         self.strength_type = strength_type
         self.concat_mask = concat_mask
 
-    def get_control(self, x_noisy, t, cond, batched_number):
+    def get_control(self, x_noisy, t, cond, batched_number, transformer_options):
         control_prev = None
         if self.previous_controlnet is not None:
-            control_prev = self.previous_controlnet.get_control(x_noisy, t, cond, batched_number)
+            control_prev = self.previous_controlnet.get_control(x_noisy, t, cond, batched_number, transformer_options)
 
         if self.timestep_range is not None:
             if t[0] > self.timestep_range[0] or t[0] < self.timestep_range[1]:
@@ -674,10 +674,10 @@ def scale_image_to(self, width, height):
         height = math.ceil(height / unshuffle_amount) * unshuffle_amount
         return width, height
 
-    def get_control(self, x_noisy, t, cond, batched_number):
+    def get_control(self, x_noisy, t, cond, batched_number, transformer_options):
         control_prev = None
         if self.previous_controlnet is not None:
-            control_prev = self.previous_controlnet.get_control(x_noisy, t, cond, batched_number)
+            control_prev = self.previous_controlnet.get_control(x_noisy, t, cond, batched_number, transformer_options)
 
         if self.timestep_range is not None:
             if t[0] > self.timestep_range[0] or t[0] < self.timestep_range[1]:

diff --git a/comfy/samplers.py b/comfy/samplers.py
@@ -271,9 +271,6 @@ def outer_calc_cond_batch(model: 'BaseModel', conds: List[List[Dict]], x_in: tor
             c = cond_cat(c)
             timestep_ = torch.cat([timestep] * batch_chunks)
 
-            if control is not None:
-                c['control'] = control.get_control(input_x, timestep_, c, len(cond_or_uncond))
-
             transformer_options = {}
             if 'transformer_options' in model_options:
                 transformer_options = model_options['transformer_options'].copy()
@@ -295,6 +292,9 @@ def outer_calc_cond_batch(model: 'BaseModel', conds: List[List[Dict]], x_in: tor
 
             c['transformer_options'] = transformer_options
 
+            if control is not None:
+                c['control'] = control.get_control(input_x, timestep_, c, len(cond_or_uncond), transformer_options)
+
             if 'model_function_wrapper' in model_options:
                 output = model_options['model_function_wrapper'](model.apply_model, {"input": input_x, "timestep": timestep_, "c": c, "cond_or_uncond": cond_or_uncond}).chunk(batch_chunks)
             else:
@@ -769,7 +769,7 @@ def inner_sample(self, noise, latent_image, device, sampler, sigmas, denoise_mas
 
         self.conds = process_conds(self.inner_model, noise, self.conds, device, latent_image, denoise_mask, seed)
 
-        extra_args = {"model_options": self.model_options, "seed":seed}
+        extra_args = {"model_options": comfy.model_patcher.create_model_options_clone(self.model_options), "seed": seed}
 
         samples = sampler.sample(self, sigmas, extra_args, callback, noise, latent_image, denoise_mask, disable_pbar)
         return self.inner_model.process_latent_out(samples.to(torch.float32))