Merge pull request #869 from rewbs/rewbs/a1111-1.6-cn-hack

Workaround for 'unhashable type: 'slice' /  'No ControlNetUnit detected in args' since 1.6 upgrade, and for drop_path issue related to timm upgrade.

.github/workflows/run_tests.yaml

@@ -14,7 +14,7 @@ jobs:
         uses: actions/checkout@v3
         with:
           repository: AUTOMATIC1111/stable-diffusion-webui
-          ref: v1.5.1      
+          ref: v1.6.0 
       - name: Checkout Controlnet extension
         uses: actions/checkout@v3
         with:

scripts/deforum_helpers/deforum_controlnet.py

@@ -18,6 +18,7 @@
 # https://github.com/Mikubill/sd-webui-controlnet — controlnet repo
 
 import os
+import copy
 import gradio as gr
 import scripts
 from PIL import Image
@@ -264,7 +265,27 @@ def read_cn_data(cn_idx):
     if not any(os.path.exists(cn_inputframes) for cn_inputframes in cn_inputframes_list) and not any_loopback_mode:
         print(f'\033[33mNeither the base nor the masking frames for ControlNet were found. Using the regular pipeline\033[0m')
 
-    p.scripts = scripts.scripts_img2img if is_img2img else scripts.scripts_txt2img
+    # Remove all scripts except controlnet.
+    #
+    # This is required because controlnet's access to p.script_args invokes @script_args.setter, 
+    # which triggers *all* alwayson_scripts' setup() functions, with whatever happens to be in script_args.
+    # In the case of seed.py (which we really don't need with deforum), this ovewrites our p.seed & co, which we
+    # had carefully prepared previously. So let's remove the scripts to avoid the problem.
+    #
+    # An alternative would be to populate all the args with the correct values
+    # for all scripts, but this seems even more fragile, as it would break
+    # if a1111 adds or removed scripts.
+    #
+    # Note that we must copy scripts.scripts_img2img or scripts.scripts_txt2img before mutating it
+    # because it persists across requests. Shallow-copying is sufficient because we only mutate a top-level
+    # reference (scripts.alwayson_scripts)
+    #
+    p.scripts = copy.copy(scripts.scripts_img2img if is_img2img else scripts.scripts_txt2img)
+    controlnet_script = find_controlnet_script(p)
+    p.scripts.alwayson_scripts =  [controlnet_script]
+    # Filling the list with None is safe because only the length will be considered,
+    # and all cn args will be replaced.
+    p.script_args_value = [None] * controlnet_script.args_to
 
     def create_cnu_dict(cn_args, prefix, img_np, mask_np, frame_idx, CnSchKeys):
 
@@ -292,9 +313,14 @@ def create_cnu_dict(cn_args, prefix, img_np, mask_np, frame_idx, CnSchKeys):
     cn_units = [cnet.ControlNetUnit(**create_cnu_dict(controlnet_args, f"cn_{i + 1}", img_np, mask_np, frame_idx, CnSchKeys))
                 for i, (img_np, mask_np) in enumerate(zip(images_np, masks_np))]
 
-    p.script_args = {"enabled": True}
     cnet.update_cn_script_in_processing(p, cn_units, is_img2img=is_img2img, is_ui=False)
 
+def find_controlnet_script(p):
+    controlnet_script = next((script for script in p.scripts.alwayson_scripts if script.title().lower()  == "controlnet"), None)
+    if not controlnet_script:
+        raise Exception("ControlNet script not found.")
+    return controlnet_script
+
 def process_controlnet_input_frames(args, anim_args, controlnet_args, video_path, mask_path, outdir_suffix, id):
     if (video_path or mask_path) and getattr(controlnet_args, f'cn_{id}_enabled'):
         frame_path = os.path.join(args.outdir, f'controlnet_{id}_{outdir_suffix}')

scripts/deforum_helpers/src/midas/backbones/beit.py

@@ -95,16 +95,23 @@ def block_forward(self, x, resolution, shared_rel_pos_bias: Optional[torch.Tenso
     """
     Modification of timm.models.beit.py: Block.forward to support arbitrary window sizes.
     """
+
+    if hasattr(self, 'drop_path1'):
+        drop_path_compat = self.drop_path1
+    elif hasattr(self.target, 'drop_path'):
+        drop_path_compat = self.drop_path
+    else:
+        raise AttributeError("Neither drop_path1 nor drop_path exists on the target.")        
+
     if self.gamma_1 is None:
-        x = x + self.drop_path(self.attn(self.norm1(x), resolution, shared_rel_pos_bias=shared_rel_pos_bias))
-        x = x + self.drop_path(self.mlp(self.norm2(x)))
+        x = x + drop_path_compat(self.attn(self.norm1(x), resolution, shared_rel_pos_bias=shared_rel_pos_bias))
+        x = x + drop_path_compat(self.mlp(self.norm2(x)))
     else:
-        x = x + self.drop_path(self.gamma_1 * self.attn(self.norm1(x), resolution,
+        x = x + drop_path_compat(self.gamma_1 * self.attn(self.norm1(x), resolution,
                                                         shared_rel_pos_bias=shared_rel_pos_bias))
-        x = x + self.drop_path(self.gamma_2 * self.mlp(self.norm2(x)))
+        x = x + drop_path_compat(self.gamma_2 * self.mlp(self.norm2(x)))
     return x
 
-
 def beit_forward_features(self, x):
     """
     Modification of timm.models.beit.py: Beit.forward_features to support arbitrary window sizes.