diff --git a/extensions-builtin/sd_forge_controlnet/scripts/controlnet.py b/extensions-builtin/sd_forge_controlnet/scripts/controlnet.py
index f3619d68..1bd9562f 100644
--- a/extensions-builtin/sd_forge_controlnet/scripts/controlnet.py
+++ b/extensions-builtin/sd_forge_controlnet/scripts/controlnet.py
@@ -459,15 +459,34 @@ class ControlNetForForgeOfficial(scripts.Script):
         params.preprocessor = preprocessor
 
         model_filename = global_state.get_controlnet_filename(unit.model)
-        controlnet_model = cached_controlnet_loader(model_filename)
+        params.model = cached_controlnet_loader(model_filename)
 
-        controlnet_model.strength = float(unit.weight)
-        controlnet_model.start_percent = float(unit.guidance_start)
-        controlnet_model.end_percent = float(unit.guidance_end)
-        controlnet_model.positive_advanced_weighting = None
-        controlnet_model.negative_advanced_weighting = None
-        controlnet_model.advanced_frame_weighting = None
-        controlnet_model.advanced_sigma_weighting = None
+        logger.info(f"Current ControlNet: {model_filename}")
+
+        return
+
+    def process_unit_before_every_sampling(self,
+                                           p: StableDiffusionProcessing,
+                                           unit: external_code.ControlNetUnit,
+                                           params: ControlNetCachedParameters,
+                                           *args, **kwargs):
+
+        is_hr_pass = getattr(p, 'is_hr_pass', False)
+
+        if is_hr_pass:
+            cond = params.control_cond_for_hr_fix
+        else:
+            cond = params.control_cond
+
+        kwargs.update(dict(unit=unit, params=params))
+
+        params.model.strength = float(unit.weight)
+        params.model.start_percent = float(unit.guidance_start)
+        params.model.end_percent = float(unit.guidance_end)
+        params.model.positive_advanced_weighting = None
+        params.model.negative_advanced_weighting = None
+        params.model.advanced_frame_weighting = None
+        params.model.advanced_sigma_weighting = None
 
         soft_weighting = {
             'input': [0.09941396206337118, 0.12050177219802567, 0.14606275417942507, 0.17704576264172736,
@@ -488,30 +507,13 @@ class ControlNetForForgeOfficial(scripts.Script):
         }
 
         if unit.control_mode == external_code.ControlMode.CONTROL.value:
-            controlnet_model.positive_advanced_weighting = soft_weighting.copy()
-            controlnet_model.negative_advanced_weighting = zero_weighting.copy()
+            params.model.positive_advanced_weighting = soft_weighting.copy()
+            params.model.negative_advanced_weighting = zero_weighting.copy()
 
-        if unit.control_mode == external_code.ControlMode.PROMPT.value:
-            controlnet_model.positive_advanced_weighting = soft_weighting.copy()
-            controlnet_model.negative_advanced_weighting = soft_weighting.copy()
-
-        params.model = controlnet_model
-        logger.info(f"Current ControlNet: {model_filename}")
-
-        return
-
-    def process_unit_before_every_sampling(self,
-                                           p: StableDiffusionProcessing,
-                                           unit: external_code.ControlNetUnit,
-                                           params: ControlNetCachedParameters,
-                                           *args, **kwargs):
-
-        if getattr(p, 'is_hr_pass', False):
-            cond = params.control_cond_for_hr_fix
-        else:
-            cond = params.control_cond
-
-        kwargs.update(dict(unit=unit, params=params))
+        # high-ref fix pass always use softer injections
+        if is_hr_pass or unit.control_mode == external_code.ControlMode.PROMPT.value:
+            params.model.positive_advanced_weighting = soft_weighting.copy()
+            params.model.negative_advanced_weighting = soft_weighting.copy()
 
         params.preprocessor.process_before_every_sampling(process=p, cond=cond, **kwargs)
         params.model.process_before_every_sampling(process=p, cond=cond, **kwargs)