comfyui-inpaint-nodes
comfyui-inpaint-nodes copied to clipboard
Published
20 hours ago •
Acly
Acly
(IMPORT FAILED) ComfyUI Inpaint Nodes?
1.update ComfyUI and
pip install -r requirements.txt
2.reinstall this node
Use this way I solve this problem, I hope it will also help you guys!
pip install -r requirements.txt
1.update ComfyUI and
pip install -r requirements.txt2.reinstall this nodeUse this way I solve this problem, I hope it will also help you guys!
1.update ComfyUI and
pip install -r requirements.txt2.reinstall this nodeUse this way I solve this problem, I hope it will also help you guys!
Thanks! Unfortunately, it didn't work for me. Maybe there is some dependency on the Python version, I'm not sure.
You will find the reason for failed import in ComfyUI's output (terminal window).
You will find the reason for failed import in ComfyUI's output (terminal window).
Thanks! I have the following records:
File "C:\ai\ComfyUI\ComfyUI\custom_nodes\comfyui-inpaint-nodes\nodes.py", line 63, in <module>
raise RuntimeError(too_old_msg)
RuntimeError: comfyui-inpaint-nodes requires a newer version of ComfyUI (v0.1.1 or later), please update!
Cannot import C:\ai\ComfyUI\ComfyUI\custom_nodes\comfyui-inpaint-nodes module for custom nodes: comfyui-inpaint-nodes requires a newer version of ComfyUI (v0.1.1 or later), please update!
But ComfyUI has already been updated to the latest version (ComfyUI v0.2.2 Release).
I was able to fix it. I can't be 100% sure about the correct sequence, because some of my actions were repetitions of found solutions, and not conscious actions. I performed these actions:
python -m pip uninstall torch
pip install h5py
pip install typing-extensions
pip install wheel
pip install tensorflow==2.13.0rc2
python -m pip install typing_extensions==4.11
python -m pip install torch==2.3.1 torchvision==0.18.1 torchaudio==2.3.1 --extra-index-url https://download.pytorch.org/whl/cu121
pip install insightface==0.7.3
I also encountered this problem, but after I updated comfyui, the problem was not solved. Finally, I found that the root cause of the problem was that the code in the comfy/lora.py file was changed (probably a custom node). Paste the following code at the end of the lora.py file to solve the problem.
def weight_decompose(dora_scale, weight, lora_diff, alpha, strength, intermediate_dtype):
dora_scale = comfy.model_management.cast_to_device(dora_scale, weight.device, intermediate_dtype)
lora_diff *= alpha
weight_calc = weight + lora_diff.type(weight.dtype)
weight_norm = (
weight_calc.transpose(0, 1)
.reshape(weight_calc.shape[1], -1)
.norm(dim=1, keepdim=True)
.reshape(weight_calc.shape[1], *[1] * (weight_calc.dim() - 1))
.transpose(0, 1)
)
weight_calc *= (dora_scale / weight_norm).type(weight.dtype)
if strength != 1.0:
weight_calc -= weight
weight += strength * (weight_calc)
else:
weight[:] = weight_calc
return weight
def pad_tensor_to_shape(tensor: torch.Tensor, new_shape: list[int]) -> torch.Tensor:
"""
Pad a tensor to a new shape with zeros.
Args:
tensor (torch.Tensor): The original tensor to be padded.
new_shape (List[int]): The desired shape of the padded tensor.
Returns:
torch.Tensor: A new tensor padded with zeros to the specified shape.
Note:
If the new shape is smaller than the original tensor in any dimension,
the original tensor will be truncated in that dimension.
"""
if any([new_shape[i] < tensor.shape[i] for i in range(len(new_shape))]):
raise ValueError("The new shape must be larger than the original tensor in all dimensions")
if len(new_shape) != len(tensor.shape):
raise ValueError("The new shape must have the same number of dimensions as the original tensor")
# Create a new tensor filled with zeros
padded_tensor = torch.zeros(new_shape, dtype=tensor.dtype, device=tensor.device)
# Create slicing tuples for both tensors
orig_slices = tuple(slice(0, dim) for dim in tensor.shape)
new_slices = tuple(slice(0, dim) for dim in tensor.shape)
# Copy the original tensor into the new tensor
padded_tensor[new_slices] = tensor[orig_slices]
return padded_tensor
def calculate_weight(patches, weight, key, intermediate_dtype=torch.float32):
for p in patches:
strength = p[0]
v = p[1]
strength_model = p[2]
offset = p[3]
function = p[4]
if function is None:
function = lambda a: a
old_weight = None
if offset is not None:
old_weight = weight
weight = weight.narrow(offset[0], offset[1], offset[2])
if strength_model != 1.0:
weight *= strength_model
if isinstance(v, list):
v = (calculate_weight(v[1:], comfy.model_management.cast_to_device(v[0], weight.device, intermediate_dtype, copy=True), key, intermediate_dtype=intermediate_dtype), )
if len(v) == 1:
patch_type = "diff"
elif len(v) == 2:
patch_type = v[0]
v = v[1]
if patch_type == "diff":
diff: torch.Tensor = v[0]
# An extra flag to pad the weight if the diff's shape is larger than the weight
do_pad_weight = len(v) > 1 and v[1]['pad_weight']
if do_pad_weight and diff.shape != weight.shape:
logging.info("Pad weight {} from {} to shape: {}".format(key, weight.shape, diff.shape))
weight = pad_tensor_to_shape(weight, diff.shape)
if strength != 0.0:
if diff.shape != weight.shape:
logging.warning("WARNING SHAPE MISMATCH {} WEIGHT NOT MERGED {} != {}".format(key, diff.shape, weight.shape))
else:
weight += function(strength * comfy.model_management.cast_to_device(diff, weight.device, weight.dtype))
elif patch_type == "lora": #lora/locon
mat1 = comfy.model_management.cast_to_device(v[0], weight.device, intermediate_dtype)
mat2 = comfy.model_management.cast_to_device(v[1], weight.device, intermediate_dtype)
dora_scale = v[4]
if v[2] is not None:
alpha = v[2] / mat2.shape[0]
else:
alpha = 1.0
if v[3] is not None:
#locon mid weights, hopefully the math is fine because I didn't properly test it
mat3 = comfy.model_management.cast_to_device(v[3], weight.device, intermediate_dtype)
final_shape = [mat2.shape[1], mat2.shape[0], mat3.shape[2], mat3.shape[3]]
mat2 = torch.mm(mat2.transpose(0, 1).flatten(start_dim=1), mat3.transpose(0, 1).flatten(start_dim=1)).reshape(final_shape).transpose(0, 1)
try:
lora_diff = torch.mm(mat1.flatten(start_dim=1), mat2.flatten(start_dim=1)).reshape(weight.shape)
if dora_scale is not None:
weight = function(weight_decompose(dora_scale, weight, lora_diff, alpha, strength, intermediate_dtype))
else:
weight += function(((strength * alpha) * lora_diff).type(weight.dtype))
except Exception as e:
logging.error("ERROR {} {} {}".format(patch_type, key, e))
elif patch_type == "lokr":
w1 = v[0]
w2 = v[1]
w1_a = v[3]
w1_b = v[4]
w2_a = v[5]
w2_b = v[6]
t2 = v[7]
dora_scale = v[8]
dim = None
if w1 is None:
dim = w1_b.shape[0]
w1 = torch.mm(comfy.model_management.cast_to_device(w1_a, weight.device, intermediate_dtype),
comfy.model_management.cast_to_device(w1_b, weight.device, intermediate_dtype))
else:
w1 = comfy.model_management.cast_to_device(w1, weight.device, intermediate_dtype)
if w2 is None:
dim = w2_b.shape[0]
if t2 is None:
w2 = torch.mm(comfy.model_management.cast_to_device(w2_a, weight.device, intermediate_dtype),
comfy.model_management.cast_to_device(w2_b, weight.device, intermediate_dtype))
else:
w2 = torch.einsum('i j k l, j r, i p -> p r k l',
comfy.model_management.cast_to_device(t2, weight.device, intermediate_dtype),
comfy.model_management.cast_to_device(w2_b, weight.device, intermediate_dtype),
comfy.model_management.cast_to_device(w2_a, weight.device, intermediate_dtype))
else:
w2 = comfy.model_management.cast_to_device(w2, weight.device, intermediate_dtype)
if len(w2.shape) == 4:
w1 = w1.unsqueeze(2).unsqueeze(2)
if v[2] is not None and dim is not None:
alpha = v[2] / dim
else:
alpha = 1.0
try:
lora_diff = torch.kron(w1, w2).reshape(weight.shape)
if dora_scale is not None:
weight = function(weight_decompose(dora_scale, weight, lora_diff, alpha, strength, intermediate_dtype))
else:
weight += function(((strength * alpha) * lora_diff).type(weight.dtype))
except Exception as e:
logging.error("ERROR {} {} {}".format(patch_type, key, e))
elif patch_type == "loha":
w1a = v[0]
w1b = v[1]
if v[2] is not None:
alpha = v[2] / w1b.shape[0]
else:
alpha = 1.0
w2a = v[3]
w2b = v[4]
dora_scale = v[7]
if v[5] is not None: #cp decomposition
t1 = v[5]
t2 = v[6]
m1 = torch.einsum('i j k l, j r, i p -> p r k l',
comfy.model_management.cast_to_device(t1, weight.device, intermediate_dtype),
comfy.model_management.cast_to_device(w1b, weight.device, intermediate_dtype),
comfy.model_management.cast_to_device(w1a, weight.device, intermediate_dtype))
m2 = torch.einsum('i j k l, j r, i p -> p r k l',
comfy.model_management.cast_to_device(t2, weight.device, intermediate_dtype),
comfy.model_management.cast_to_device(w2b, weight.device, intermediate_dtype),
comfy.model_management.cast_to_device(w2a, weight.device, intermediate_dtype))
else:
m1 = torch.mm(comfy.model_management.cast_to_device(w1a, weight.device, intermediate_dtype),
comfy.model_management.cast_to_device(w1b, weight.device, intermediate_dtype))
m2 = torch.mm(comfy.model_management.cast_to_device(w2a, weight.device, intermediate_dtype),
comfy.model_management.cast_to_device(w2b, weight.device, intermediate_dtype))
try:
lora_diff = (m1 * m2).reshape(weight.shape)
if dora_scale is not None:
weight = function(weight_decompose(dora_scale, weight, lora_diff, alpha, strength, intermediate_dtype))
else:
weight += function(((strength * alpha) * lora_diff).type(weight.dtype))
except Exception as e:
logging.error("ERROR {} {} {}".format(patch_type, key, e))
elif patch_type == "glora":
dora_scale = v[5]
old_glora = False
if v[3].shape[1] == v[2].shape[0] == v[0].shape[0] == v[1].shape[1]:
rank = v[0].shape[0]
old_glora = True
if v[3].shape[0] == v[2].shape[1] == v[0].shape[1] == v[1].shape[0]:
if old_glora and v[1].shape[0] == weight.shape[0] and weight.shape[0] == weight.shape[1]:
pass
else:
old_glora = False
rank = v[1].shape[0]
a1 = comfy.model_management.cast_to_device(v[0].flatten(start_dim=1), weight.device, intermediate_dtype)
a2 = comfy.model_management.cast_to_device(v[1].flatten(start_dim=1), weight.device, intermediate_dtype)
b1 = comfy.model_management.cast_to_device(v[2].flatten(start_dim=1), weight.device, intermediate_dtype)
b2 = comfy.model_management.cast_to_device(v[3].flatten(start_dim=1), weight.device, intermediate_dtype)
if v[4] is not None:
alpha = v[4] / rank
else:
alpha = 1.0
try:
if old_glora:
lora_diff = (torch.mm(b2, b1) + torch.mm(torch.mm(weight.flatten(start_dim=1).to(dtype=intermediate_dtype), a2), a1)).reshape(weight.shape) #old lycoris glora
else:
if weight.dim() > 2:
lora_diff = torch.einsum("o i ..., i j -> o j ...", torch.einsum("o i ..., i j -> o j ...", weight.to(dtype=intermediate_dtype), a1), a2).reshape(weight.shape)
else:
lora_diff = torch.mm(torch.mm(weight.to(dtype=intermediate_dtype), a1), a2).reshape(weight.shape)
lora_diff += torch.mm(b1, b2).reshape(weight.shape)
if dora_scale is not None:
weight = function(weight_decompose(dora_scale, weight, lora_diff, alpha, strength, intermediate_dtype))
else:
weight += function(((strength * alpha) * lora_diff).type(weight.dtype))
except Exception as e:
logging.error("ERROR {} {} {}".format(patch_type, key, e))
else:
logging.warning("patch type not recognized {} {}".format(patch_type, key))
if old_weight is not None:
weight = old_weight
return weight
I also encountered this problem, but after I updated comfyui, the problem was not solved. Finally, I found that the root cause of the problem was that the code in the comfy/lora.py file was changed (probably a custom node). Paste the following code at the end of the lora.py file to solve the problem.
def weight_decompose(dora_scale, weight, lora_diff, alpha, strength, intermediate_dtype): dora_scale = comfy.model_management.cast_to_device(dora_scale, weight.device, intermediate_dtype) lora_diff *= alpha weight_calc = weight + lora_diff.type(weight.dtype) weight_norm = ( weight_calc.transpose(0, 1) .reshape(weight_calc.shape[1], -1) .norm(dim=1, keepdim=True) .reshape(weight_calc.shape[1], *[1] * (weight_calc.dim() - 1)) .transpose(0, 1) ) weight_calc *= (dora_scale / weight_norm).type(weight.dtype) if strength != 1.0: weight_calc -= weight weight += strength * (weight_calc) else: weight[:] = weight_calc return weight def pad_tensor_to_shape(tensor: torch.Tensor, new_shape: list[int]) -> torch.Tensor: """ Pad a tensor to a new shape with zeros. Args: tensor (torch.Tensor): The original tensor to be padded. new_shape (List[int]): The desired shape of the padded tensor. Returns: torch.Tensor: A new tensor padded with zeros to the specified shape. Note: If the new shape is smaller than the original tensor in any dimension, the original tensor will be truncated in that dimension. """ if any([new_shape[i] < tensor.shape[i] for i in range(len(new_shape))]): raise ValueError("The new shape must be larger than the original tensor in all dimensions") if len(new_shape) != len(tensor.shape): raise ValueError("The new shape must have the same number of dimensions as the original tensor") # Create a new tensor filled with zeros padded_tensor = torch.zeros(new_shape, dtype=tensor.dtype, device=tensor.device) # Create slicing tuples for both tensors orig_slices = tuple(slice(0, dim) for dim in tensor.shape) new_slices = tuple(slice(0, dim) for dim in tensor.shape) # Copy the original tensor into the new tensor padded_tensor[new_slices] = tensor[orig_slices] return padded_tensor def calculate_weight(patches, weight, key, intermediate_dtype=torch.float32): for p in patches: strength = p[0] v = p[1] strength_model = p[2] offset = p[3] function = p[4] if function is None: function = lambda a: a old_weight = None if offset is not None: old_weight = weight weight = weight.narrow(offset[0], offset[1], offset[2]) if strength_model != 1.0: weight *= strength_model if isinstance(v, list): v = (calculate_weight(v[1:], comfy.model_management.cast_to_device(v[0], weight.device, intermediate_dtype, copy=True), key, intermediate_dtype=intermediate_dtype), ) if len(v) == 1: patch_type = "diff" elif len(v) == 2: patch_type = v[0] v = v[1] if patch_type == "diff": diff: torch.Tensor = v[0] # An extra flag to pad the weight if the diff's shape is larger than the weight do_pad_weight = len(v) > 1 and v[1]['pad_weight'] if do_pad_weight and diff.shape != weight.shape: logging.info("Pad weight {} from {} to shape: {}".format(key, weight.shape, diff.shape)) weight = pad_tensor_to_shape(weight, diff.shape) if strength != 0.0: if diff.shape != weight.shape: logging.warning("WARNING SHAPE MISMATCH {} WEIGHT NOT MERGED {} != {}".format(key, diff.shape, weight.shape)) else: weight += function(strength * comfy.model_management.cast_to_device(diff, weight.device, weight.dtype)) elif patch_type == "lora": #lora/locon mat1 = comfy.model_management.cast_to_device(v[0], weight.device, intermediate_dtype) mat2 = comfy.model_management.cast_to_device(v[1], weight.device, intermediate_dtype) dora_scale = v[4] if v[2] is not None: alpha = v[2] / mat2.shape[0] else: alpha = 1.0 if v[3] is not None: #locon mid weights, hopefully the math is fine because I didn't properly test it mat3 = comfy.model_management.cast_to_device(v[3], weight.device, intermediate_dtype) final_shape = [mat2.shape[1], mat2.shape[0], mat3.shape[2], mat3.shape[3]] mat2 = torch.mm(mat2.transpose(0, 1).flatten(start_dim=1), mat3.transpose(0, 1).flatten(start_dim=1)).reshape(final_shape).transpose(0, 1) try: lora_diff = torch.mm(mat1.flatten(start_dim=1), mat2.flatten(start_dim=1)).reshape(weight.shape) if dora_scale is not None: weight = function(weight_decompose(dora_scale, weight, lora_diff, alpha, strength, intermediate_dtype)) else: weight += function(((strength * alpha) * lora_diff).type(weight.dtype)) except Exception as e: logging.error("ERROR {} {} {}".format(patch_type, key, e)) elif patch_type == "lokr": w1 = v[0] w2 = v[1] w1_a = v[3] w1_b = v[4] w2_a = v[5] w2_b = v[6] t2 = v[7] dora_scale = v[8] dim = None if w1 is None: dim = w1_b.shape[0] w1 = torch.mm(comfy.model_management.cast_to_device(w1_a, weight.device, intermediate_dtype), comfy.model_management.cast_to_device(w1_b, weight.device, intermediate_dtype)) else: w1 = comfy.model_management.cast_to_device(w1, weight.device, intermediate_dtype) if w2 is None: dim = w2_b.shape[0] if t2 is None: w2 = torch.mm(comfy.model_management.cast_to_device(w2_a, weight.device, intermediate_dtype), comfy.model_management.cast_to_device(w2_b, weight.device, intermediate_dtype)) else: w2 = torch.einsum('i j k l, j r, i p -> p r k l', comfy.model_management.cast_to_device(t2, weight.device, intermediate_dtype), comfy.model_management.cast_to_device(w2_b, weight.device, intermediate_dtype), comfy.model_management.cast_to_device(w2_a, weight.device, intermediate_dtype)) else: w2 = comfy.model_management.cast_to_device(w2, weight.device, intermediate_dtype) if len(w2.shape) == 4: w1 = w1.unsqueeze(2).unsqueeze(2) if v[2] is not None and dim is not None: alpha = v[2] / dim else: alpha = 1.0 try: lora_diff = torch.kron(w1, w2).reshape(weight.shape) if dora_scale is not None: weight = function(weight_decompose(dora_scale, weight, lora_diff, alpha, strength, intermediate_dtype)) else: weight += function(((strength * alpha) * lora_diff).type(weight.dtype)) except Exception as e: logging.error("ERROR {} {} {}".format(patch_type, key, e)) elif patch_type == "loha": w1a = v[0] w1b = v[1] if v[2] is not None: alpha = v[2] / w1b.shape[0] else: alpha = 1.0 w2a = v[3] w2b = v[4] dora_scale = v[7] if v[5] is not None: #cp decomposition t1 = v[5] t2 = v[6] m1 = torch.einsum('i j k l, j r, i p -> p r k l', comfy.model_management.cast_to_device(t1, weight.device, intermediate_dtype), comfy.model_management.cast_to_device(w1b, weight.device, intermediate_dtype), comfy.model_management.cast_to_device(w1a, weight.device, intermediate_dtype)) m2 = torch.einsum('i j k l, j r, i p -> p r k l', comfy.model_management.cast_to_device(t2, weight.device, intermediate_dtype), comfy.model_management.cast_to_device(w2b, weight.device, intermediate_dtype), comfy.model_management.cast_to_device(w2a, weight.device, intermediate_dtype)) else: m1 = torch.mm(comfy.model_management.cast_to_device(w1a, weight.device, intermediate_dtype), comfy.model_management.cast_to_device(w1b, weight.device, intermediate_dtype)) m2 = torch.mm(comfy.model_management.cast_to_device(w2a, weight.device, intermediate_dtype), comfy.model_management.cast_to_device(w2b, weight.device, intermediate_dtype)) try: lora_diff = (m1 * m2).reshape(weight.shape) if dora_scale is not None: weight = function(weight_decompose(dora_scale, weight, lora_diff, alpha, strength, intermediate_dtype)) else: weight += function(((strength * alpha) * lora_diff).type(weight.dtype)) except Exception as e: logging.error("ERROR {} {} {}".format(patch_type, key, e)) elif patch_type == "glora": dora_scale = v[5] old_glora = False if v[3].shape[1] == v[2].shape[0] == v[0].shape[0] == v[1].shape[1]: rank = v[0].shape[0] old_glora = True if v[3].shape[0] == v[2].shape[1] == v[0].shape[1] == v[1].shape[0]: if old_glora and v[1].shape[0] == weight.shape[0] and weight.shape[0] == weight.shape[1]: pass else: old_glora = False rank = v[1].shape[0] a1 = comfy.model_management.cast_to_device(v[0].flatten(start_dim=1), weight.device, intermediate_dtype) a2 = comfy.model_management.cast_to_device(v[1].flatten(start_dim=1), weight.device, intermediate_dtype) b1 = comfy.model_management.cast_to_device(v[2].flatten(start_dim=1), weight.device, intermediate_dtype) b2 = comfy.model_management.cast_to_device(v[3].flatten(start_dim=1), weight.device, intermediate_dtype) if v[4] is not None: alpha = v[4] / rank else: alpha = 1.0 try: if old_glora: lora_diff = (torch.mm(b2, b1) + torch.mm(torch.mm(weight.flatten(start_dim=1).to(dtype=intermediate_dtype), a2), a1)).reshape(weight.shape) #old lycoris glora else: if weight.dim() > 2: lora_diff = torch.einsum("o i ..., i j -> o j ...", torch.einsum("o i ..., i j -> o j ...", weight.to(dtype=intermediate_dtype), a1), a2).reshape(weight.shape) else: lora_diff = torch.mm(torch.mm(weight.to(dtype=intermediate_dtype), a1), a2).reshape(weight.shape) lora_diff += torch.mm(b1, b2).reshape(weight.shape) if dora_scale is not None: weight = function(weight_decompose(dora_scale, weight, lora_diff, alpha, strength, intermediate_dtype)) else: weight += function(((strength * alpha) * lora_diff).type(weight.dtype)) except Exception as e: logging.error("ERROR {} {} {}".format(patch_type, key, e)) else: logging.warning("patch type not recognized {} {}".format(patch_type, key)) if old_weight is not None: weight = old_weight return weight
This one works.
But you also have to add
import torch
at beginning of comfy/lora.py
