Error occurred when executing ADE_AnimateDiffLoaderWithContext:

[zqlong123]

Error occurred when executing ADE_AnimateDiffLoaderWithContext:

module 'comfy.ops' has no attribute 'Linear'

File "H:\Blender_ComfyUI\ComfyUI\execution.py", line 153, in recursive_execute output_data, output_ui = get_output_data(obj, input_data_all) File "H:\Blender_ComfyUI\ComfyUI\execution.py", line 83, in get_output_data return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True) File "H:\Blender_ComfyUI\ComfyUI\execution.py", line 76, in map_node_over_list results.append(getattr(obj, func)(*slice_dict(input_data_all, i))) File "H:\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\nodes.py", line 204, in load_mm_and_inject_params mm = load_motion_module(model_name, motion_lora, model=model, motion_model_settings=motion_model_settings) File "H:\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\motion_module.py", line 216, in load_motion_module motion_module = AnimDiffMotionWrapper(mm_state_dict=mm_state_dict, mm_hash=model_hash, mm_name=model_name, loras=loras) File "H:\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\motion_module_ad.py", line 44, in init self.down_blocks.append(MotionModule(c, temporal_position_encoding_max_len=self.encoding_max_len, block_type=BlockType.DOWN)) File "H:\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\motion_module_ad.py", line 97, in init get_motion_module(in_channels, temporal_position_encoding_max_len), File "H:\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\motion_module_ad.py", line 119, in get_motion_module return VanillaTemporalModule(in_channels=in_channels, temporal_position_encoding_max_len=temporal_position_encoding_max_len) File "H:\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\motion_module_ad.py", line 137, in init self.temporal_transformer = TemporalTransformer3DModel( File "H:\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\motion_module_ad.py", line 199, in init [ File "H:\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\motion_module_ad.py", line 200, in TemporalTransformerBlock( File "H:\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\motion_module_ad.py", line 292, in init VersatileAttention( File "H:\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\motion_module_ad.py", line 386, in init super().init(args, **kwargs) File "H:\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\motion_utils.py", line 40, in init self.to_q = operations.Linear(query_dim, inner_dim, bias=False, dtype=dtype, device=device)

TemporalTransformerBlock（ 文件“H：\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\motion_module_ad.py”，第 292 行，init VersatileAttention（ 文件“H：\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\motion_module_ad.py”，第 386 行，init super（）.init（*args， **kwargs） 文件“H：\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI-AnimateDiff-Evolved\animatediff\motion_utils.py”，第 40 行，init self.to_q = 操作。线性（query_dim， inner_dim， bias=False， dtype=dtype， device=device）

[Kosinkadink]

Update AnimateDiff-Evolved and then the error will go away.

[zqlong123]

Thanks! Let me try

[zqlong123]

The second post is about the new one: AnimateDiff-Evolved, but the problem has not been solved yet.

[zqlong123]

The second update is: AnimateDiff-Evolved, the problem has not been solved yet

[Kosinkadink]

There is no need to make a second post, we can keep it on this one. And if updating didn't fix it, that means your update didn't work. We can confirm this by you looking at the contents of the comfy_nodes/ComfyUI-AnimateDiff-Evolved/animatediff/motion_utils.py file. If it is updated, CrossAttentionMM should look like this:

If it doesn't look like that, then whatever you did to try to update did not work. How did you go about trying to update?

[zqlong123]

[zqlong123]

There is no such file ，/motion_utils.py

[Kosinkadink]

Woops, I had a typo, motion_utils is in animatediff folder there.

[zqlong123]

motion_utils.py，What should I use to open this file? When I double-click to open it, nothing happens.

[zqlong123]

I am Chinese, which country are you from?

[Kosinkadink]

You can look into the file with any text editor

[zqlong123]

import hashlib import os from pathlib import Path from typing import Callable

import numpy as np

import folder_paths from comfy.model_base import SDXL, BaseModel, model_sampling from comfy.model_management import xformers_enabled from comfy.model_patcher import ModelPatcher

class IsChangedHelper: def init(self): self.val = 0

def no_change(self):
    return self.val

def change(self):
    self.val = (self.val + 1) % 100

class ModelSamplingConfig: def init(self, beta_schedule: str): self.beta_schedule = beta_schedule

class BetaSchedules: SQRT_LINEAR = "sqrt_linear (AnimateDiff)" LINEAR = "linear (HotshotXL/default)" SQRT = "sqrt" COSINE = "cosine" SQUAREDCOS_CAP_V2 = "squaredcos_cap_v2"

ALIAS_LIST = [SQRT_LINEAR, LINEAR, SQRT, COSINE, SQUAREDCOS_CAP_V2]

ALIAS_MAP = {
    SQRT_LINEAR: "sqrt_linear",
    LINEAR: "linear",
    SQRT: "sqrt",
    COSINE: "cosine",
    SQUAREDCOS_CAP_V2: "squaredcos_cap_v2",
}

@classmethod
def to_name(cls, alias: str):
    return cls.ALIAS_MAP[alias]

@classmethod
def to_config(cls, alias: str) -> ModelSamplingConfig:
    return ModelSamplingConfig(cls.to_name(alias))

@classmethod
def to_model_sampling(cls, alias: str, model: ModelPatcher):
    return model_sampling(cls.to_config(alias), model_type=model.model.model_type)

@staticmethod
def get_alias_list_with_first_element(first_element: str):
    new_list = BetaSchedules.ALIAS_LIST.copy()
    element_index = new_list.index(first_element)
    new_list[0], new_list[element_index] = new_list[element_index], new_list[0]
    return new_list

class BetaScheduleCache: def init(self, model: ModelPatcher): self.model_sampling = model.model.model_sampling

def use_cached_beta_schedule_and_clean(self, model: ModelPatcher):
    model.model.model_sampling = self.model_sampling
    self.clean()

def clean(self):
    self.model_sampling = None

class Folders: ANIMATEDIFF_MODELS = "AnimateDiffEvolved_Models" MOTION_LORA = "AnimateDiffMotion_LoRA"

register motion models folder(s)

folder_paths.folder_names_and_paths[Folders.ANIMATEDIFF_MODELS] = ( [ str(Path(file).parent.parent / "models") ], folder_paths.supported_pt_extensions )

register motion LoRA folder(s)

folder_paths.folder_names_and_paths[Folders.MOTION_LORA] = ( [ str(Path(file).parent.parent / "motion_lora") ], folder_paths.supported_pt_extensions )

Register video_formats folder

folder_paths.folder_names_and_paths["video_formats"] = ( [ os.path.join(os.path.dirname(os.path.abspath(file)), "..", "video_formats"), ], [".json"] )

def get_available_motion_models(): return folder_paths.get_filename_list(Folders.ANIMATEDIFF_MODELS)

def get_motion_model_path(model_name: str): return folder_paths.get_full_path(Folders.ANIMATEDIFF_MODELS, model_name)

def get_available_motion_loras(): return folder_paths.get_filename_list(Folders.MOTION_LORA)

def get_motion_lora_path(lora_name: str): return folder_paths.get_full_path(Folders.MOTION_LORA, lora_name)

modified from https://stackoverflow.com/questions/22058048/hashing-a-file-in-python

def calculate_file_hash(filename: str, hash_every_n: int = 50): h = hashlib.sha256() b = bytearray(1024*1024) mv = memoryview(b) with open(filename, 'rb', buffering=0) as f: i = 0

don't hash entire file, only portions of it

    while n := f.readinto(mv):
        if i%hash_every_n == 0:
            h.update(mv[:n])
        i += 1
return h.hexdigest()

def calculate_model_hash(model: ModelPatcher): unet = model.model.diff t = unet.input_blocks[1] m = hashlib.sha256() for buf in t.buffers(): m.update(buf.cpu().numpy().view(np.uint8)) return m.hexdigest()

class ModelTypesSD: SD1_5 = "sd1_5" SDXL = "sdxl"

def get_sd_model_type(model: ModelPatcher) -> str: if model is None: return None if is_checkpoint_sd1_5(model): return ModelTypesSD.SD1_5 elif is_checkpoint_sdxl(model): return ModelTypesSD.SDXL return False

def is_checkpoint_sd1_5(model: ModelPatcher): if model is None: return False model_type = type(model.model) return model_type == BaseModel

def is_checkpoint_sdxl(model: ModelPatcher): if model is None: return False model_type = type(model.model) return model_type == SDXL

def raise_if_not_checkpoint_sd1_5(model: ModelPatcher): if not is_checkpoint_sd1_5(model): raise ValueError(f"For AnimateDiff, SD Checkpoint (model) is expected to be SD1.5-based (BaseModel), but was: {type(model.model).name}")

TODO: remove this filth when xformers bug gets fixed in future xformers version

def wrap_function_to_inject_xformers_bug_info(function_to_wrap: Callable) -> Callable: if not xformers_enabled: return function_to_wrap else: def wrapped_function(*args, *kwargs): try: return function_to_wrap(args, **kwargs) except RuntimeError as e: if str(e).startswith("CUDA error: invalid configuration argument"): raise RuntimeError(f"An xformers bug was encountered in AnimateDiff - this is unexpected, \ report this to Kosinkadink/ComfyUI-AnimateDiff-Evolved repo as an issue, \ and a workaround for now is to run ComfyUI with the --disable-xformers argument.") raise return wrapped_function

TODO: possibly add configuration file in future when needed?

Load config settings

ADE_DIR = Path(file).parent.parent

ADE_CONFIG_FILE = ADE_DIR / "ade_config.json"

class ADE_Settings:

USE_XFORMERS_IN_VERSATILE_ATTENTION = "use_xformers_in_VersatileAttention"

Create ADE config if not present

ABS_CONFIG = {

ADE_Settings.USE_XFORMERS_IN_VERSATILE_ATTENTION: True

}

if not ADE_CONFIG_FILE.exists():

with ADE_CONFIG_FILE.open("w") as f:

json.dumps(ABS_CONFIG, indent=4)

otherwise, load it and use values

else:

loaded_values: dict = None

with ADE_CONFIG_FILE.open("r") as f:

loaded_values = json.load(f)

if loaded_values is not None:

for key, value in loaded_values.items():

if key in ABS_CONFIG:

ABS_CONFIG[key] = value

[Kosinkadink]

That's the model_utils file instead of motion_utils, but even from that, I can tell you're not on the latest version of AnimateDiff-Evolved. Update AnimateDiff-Evolved either though the manager (Update All in the menu), or manually through the git pull command inside the ComfyUI-AnimateDiff-Evolved directory.

[zqlong123]

Thanks! I will try again

[bparrott78]

Updated everything, even did a git reset --hard, didn't work, same error. Rolled back to d8eddc785633877cb186470c9ecd6c19916396bb and it worked.

[Kosinkadink]

@bparrott78 what was the error you had? The commit you list there is the one that introduced the fix, and the latest version has the same code there in it. So you may have been outdated somehow, and your attempt at a rollback actually brought you forward.

[bparrott78]

Same error as the guy above, I'll try to move back to current and see if that works.

[Kosinkadink]

There is no issue with the latest version of the code. You just need to be sure you are actually using it.

[bparrott78]

Figured it out. I was on the new injection branch, rejoining the main branch solved it

[Kosinkadink]

Gotcha. New injection was never a good branch to be in tbh other than testing things out a few weeks ago, the real fixes for model patching stuff got introduced to main branch last week

[Kosinkadink]

@zqlong123 assuming you're still trying to update, also make sure you are in main brach

[bparrott78]

Yeah I hadn't used animatediff for a week so that makes sense.

[Kosinkadink]

I'm gonna go ahead and delete that branch to avoid confusion

[zqlong123]

I'll try again today, it was too late yesterday and I had to rest and sleep.

[zqlong123]

Error occurred when executing IPAdapterApply:

Error(s) in loading state_dict for Resampler: size mismatch for proj_in.weight: copying a param with shape torch.Size([768, 1280]) from checkpoint, the shape in current model is torch.Size([768, 1024]).

File "F:\Blender_ComfyUI\ComfyUI\execution.py", line 153, in recursive_execute output_data, output_ui = get_output_data(obj, input_data_all) File "F:\Blender_ComfyUI\ComfyUI\execution.py", line 83, in get_output_data return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True) File "F:\Blender_ComfyUI\ComfyUI\execution.py", line 76, in map_node_over_list results.append(getattr(obj, func)(**slice_dict(input_data_all, i))) File "F:\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI_IPAdapter_plus\IPAdapterPlus.py", line 426, in apply_ipadapter self.ipadapter = IPAdapter( File "F:\Blender_ComfyUI\ComfyUI\custom_nodes\ComfyUI_IPAdapter_plus\IPAdapterPlus.py", line 175, in init self.image_proj_model.load_state_dict(ipadapter_model["image_proj"]) File "F:\Blender_ComfyUI\python_embeded\lib\site-packages\torch\nn\modules\module.py", line 2152, in load_state_dict raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(

[zqlong123]

hi Can you help me figure out what the problem is?

[johndpope]

@zqlong123 - when you boot up comfyui - there's errors that occur. Check the logs - disable problematic extensions. I found other ones that had nothing todo with this repo - and was disabled it worked.