482 lines
18 KiB
Python
482 lines
18 KiB
Python
import torch.nn as nn
|
|
import torch
|
|
import cv2
|
|
import numpy as np
|
|
import comfy.model_management
|
|
|
|
from comfy.model_patcher import ModelPatcher
|
|
from enum import Enum
|
|
from tqdm import tqdm
|
|
from typing import Optional, Tuple
|
|
|
|
class LayerMethod(Enum):
|
|
FG_ONLY_ATTN = "Attention Injection"
|
|
FG_ONLY_CONV = "Conv Injection"
|
|
FG_TO_BLEND = "Foreground"
|
|
FG_BLEND_TO_BG = "Foreground to Background"
|
|
BG_TO_BLEND = "Background"
|
|
BG_BLEND_TO_FG = "Background to Foreground"
|
|
|
|
try:
|
|
from diffusers.configuration_utils import ConfigMixin, register_to_config
|
|
from diffusers.models.modeling_utils import ModelMixin
|
|
from diffusers.models.unet_2d_blocks import UNetMidBlock2D, get_down_block, get_up_block
|
|
import functools
|
|
|
|
def zero_module(module):
|
|
"""
|
|
Zero out the parameters of a module and return it.
|
|
"""
|
|
for p in module.parameters():
|
|
p.detach().zero_()
|
|
return module
|
|
|
|
|
|
class LatentTransparencyOffsetEncoder(torch.nn.Module):
|
|
def __init__(self, *args, **kwargs):
|
|
super().__init__(*args, **kwargs)
|
|
self.blocks = torch.nn.Sequential(
|
|
torch.nn.Conv2d(4, 32, kernel_size=3, padding=1, stride=1),
|
|
nn.SiLU(),
|
|
torch.nn.Conv2d(32, 32, kernel_size=3, padding=1, stride=1),
|
|
nn.SiLU(),
|
|
torch.nn.Conv2d(32, 64, kernel_size=3, padding=1, stride=2),
|
|
nn.SiLU(),
|
|
torch.nn.Conv2d(64, 64, kernel_size=3, padding=1, stride=1),
|
|
nn.SiLU(),
|
|
torch.nn.Conv2d(64, 128, kernel_size=3, padding=1, stride=2),
|
|
nn.SiLU(),
|
|
torch.nn.Conv2d(128, 128, kernel_size=3, padding=1, stride=1),
|
|
nn.SiLU(),
|
|
torch.nn.Conv2d(128, 256, kernel_size=3, padding=1, stride=2),
|
|
nn.SiLU(),
|
|
torch.nn.Conv2d(256, 256, kernel_size=3, padding=1, stride=1),
|
|
nn.SiLU(),
|
|
zero_module(torch.nn.Conv2d(256, 4, kernel_size=3, padding=1, stride=1)),
|
|
)
|
|
|
|
def __call__(self, x):
|
|
return self.blocks(x)
|
|
|
|
|
|
# 1024 * 1024 * 3 -> 16 * 16 * 512 -> 1024 * 1024 * 3
|
|
class UNet1024(ModelMixin, ConfigMixin):
|
|
@register_to_config
|
|
def __init__(
|
|
self,
|
|
in_channels: int = 3,
|
|
out_channels: int = 3,
|
|
down_block_types: Tuple[str] = (
|
|
"DownBlock2D",
|
|
"DownBlock2D",
|
|
"DownBlock2D",
|
|
"DownBlock2D",
|
|
"AttnDownBlock2D",
|
|
"AttnDownBlock2D",
|
|
"AttnDownBlock2D",
|
|
),
|
|
up_block_types: Tuple[str] = (
|
|
"AttnUpBlock2D",
|
|
"AttnUpBlock2D",
|
|
"AttnUpBlock2D",
|
|
"UpBlock2D",
|
|
"UpBlock2D",
|
|
"UpBlock2D",
|
|
"UpBlock2D",
|
|
),
|
|
block_out_channels: Tuple[int] = (32, 32, 64, 128, 256, 512, 512),
|
|
layers_per_block: int = 2,
|
|
mid_block_scale_factor: float = 1,
|
|
downsample_padding: int = 1,
|
|
downsample_type: str = "conv",
|
|
upsample_type: str = "conv",
|
|
dropout: float = 0.0,
|
|
act_fn: str = "silu",
|
|
attention_head_dim: Optional[int] = 8,
|
|
norm_num_groups: int = 4,
|
|
norm_eps: float = 1e-5,
|
|
):
|
|
super().__init__()
|
|
|
|
# input
|
|
self.conv_in = nn.Conv2d(
|
|
in_channels, block_out_channels[0], kernel_size=3, padding=(1, 1)
|
|
)
|
|
self.latent_conv_in = zero_module(
|
|
nn.Conv2d(4, block_out_channels[2], kernel_size=1)
|
|
)
|
|
|
|
self.down_blocks = nn.ModuleList([])
|
|
self.mid_block = None
|
|
self.up_blocks = nn.ModuleList([])
|
|
|
|
# down
|
|
output_channel = block_out_channels[0]
|
|
for i, down_block_type in enumerate(down_block_types):
|
|
input_channel = output_channel
|
|
output_channel = block_out_channels[i]
|
|
is_final_block = i == len(block_out_channels) - 1
|
|
|
|
down_block = get_down_block(
|
|
down_block_type,
|
|
num_layers=layers_per_block,
|
|
in_channels=input_channel,
|
|
out_channels=output_channel,
|
|
temb_channels=None,
|
|
add_downsample=not is_final_block,
|
|
resnet_eps=norm_eps,
|
|
resnet_act_fn=act_fn,
|
|
resnet_groups=norm_num_groups,
|
|
attention_head_dim=(
|
|
attention_head_dim
|
|
if attention_head_dim is not None
|
|
else output_channel
|
|
),
|
|
downsample_padding=downsample_padding,
|
|
resnet_time_scale_shift="default",
|
|
downsample_type=downsample_type,
|
|
dropout=dropout,
|
|
)
|
|
self.down_blocks.append(down_block)
|
|
|
|
# mid
|
|
self.mid_block = UNetMidBlock2D(
|
|
in_channels=block_out_channels[-1],
|
|
temb_channels=None,
|
|
dropout=dropout,
|
|
resnet_eps=norm_eps,
|
|
resnet_act_fn=act_fn,
|
|
output_scale_factor=mid_block_scale_factor,
|
|
resnet_time_scale_shift="default",
|
|
attention_head_dim=(
|
|
attention_head_dim
|
|
if attention_head_dim is not None
|
|
else block_out_channels[-1]
|
|
),
|
|
resnet_groups=norm_num_groups,
|
|
attn_groups=None,
|
|
add_attention=True,
|
|
)
|
|
|
|
# up
|
|
reversed_block_out_channels = list(reversed(block_out_channels))
|
|
output_channel = reversed_block_out_channels[0]
|
|
for i, up_block_type in enumerate(up_block_types):
|
|
prev_output_channel = output_channel
|
|
output_channel = reversed_block_out_channels[i]
|
|
input_channel = reversed_block_out_channels[
|
|
min(i + 1, len(block_out_channels) - 1)
|
|
]
|
|
|
|
is_final_block = i == len(block_out_channels) - 1
|
|
|
|
up_block = get_up_block(
|
|
up_block_type,
|
|
num_layers=layers_per_block + 1,
|
|
in_channels=input_channel,
|
|
out_channels=output_channel,
|
|
prev_output_channel=prev_output_channel,
|
|
temb_channels=None,
|
|
add_upsample=not is_final_block,
|
|
resnet_eps=norm_eps,
|
|
resnet_act_fn=act_fn,
|
|
resnet_groups=norm_num_groups,
|
|
attention_head_dim=(
|
|
attention_head_dim
|
|
if attention_head_dim is not None
|
|
else output_channel
|
|
),
|
|
resnet_time_scale_shift="default",
|
|
upsample_type=upsample_type,
|
|
dropout=dropout,
|
|
)
|
|
self.up_blocks.append(up_block)
|
|
prev_output_channel = output_channel
|
|
|
|
# out
|
|
self.conv_norm_out = nn.GroupNorm(
|
|
num_channels=block_out_channels[0], num_groups=norm_num_groups, eps=norm_eps
|
|
)
|
|
self.conv_act = nn.SiLU()
|
|
self.conv_out = nn.Conv2d(
|
|
block_out_channels[0], out_channels, kernel_size=3, padding=1
|
|
)
|
|
|
|
def forward(self, x, latent):
|
|
sample_latent = self.latent_conv_in(latent)
|
|
sample = self.conv_in(x)
|
|
emb = None
|
|
|
|
down_block_res_samples = (sample,)
|
|
for i, downsample_block in enumerate(self.down_blocks):
|
|
if i == 3:
|
|
sample = sample + sample_latent
|
|
|
|
sample, res_samples = downsample_block(hidden_states=sample, temb=emb)
|
|
down_block_res_samples += res_samples
|
|
|
|
sample = self.mid_block(sample, emb)
|
|
|
|
for upsample_block in self.up_blocks:
|
|
res_samples = down_block_res_samples[-len(upsample_block.resnets):]
|
|
down_block_res_samples = down_block_res_samples[
|
|
: -len(upsample_block.resnets)
|
|
]
|
|
sample = upsample_block(sample, res_samples, emb)
|
|
|
|
sample = self.conv_norm_out(sample)
|
|
sample = self.conv_act(sample)
|
|
sample = self.conv_out(sample)
|
|
return sample
|
|
|
|
|
|
def checkerboard(shape):
|
|
return np.indices(shape).sum(axis=0) % 2
|
|
|
|
|
|
def fill_checkerboard_bg(y: torch.Tensor) -> torch.Tensor:
|
|
alpha = y[..., :1]
|
|
fg = y[..., 1:]
|
|
B, H, W, C = fg.shape
|
|
cb = checkerboard(shape=(H // 64, W // 64))
|
|
cb = cv2.resize(cb, (W, H), interpolation=cv2.INTER_NEAREST)
|
|
cb = (0.5 + (cb - 0.5) * 0.1)[None, ..., None]
|
|
cb = torch.from_numpy(cb).to(fg)
|
|
vis = fg * alpha + cb * (1 - alpha)
|
|
return vis
|
|
|
|
|
|
class TransparentVAEDecoder:
|
|
def __init__(self, sd, device, dtype):
|
|
self.load_device = device
|
|
self.dtype = dtype
|
|
|
|
model = UNet1024(in_channels=3, out_channels=4)
|
|
model.load_state_dict(sd, strict=True)
|
|
model.to(self.load_device, dtype=self.dtype)
|
|
model.eval()
|
|
self.model = model
|
|
|
|
@torch.no_grad()
|
|
def estimate_single_pass(self, pixel, latent):
|
|
y = self.model(pixel, latent)
|
|
return y
|
|
|
|
@torch.no_grad()
|
|
def estimate_augmented(self, pixel, latent):
|
|
args = [
|
|
[False, 0],
|
|
[False, 1],
|
|
[False, 2],
|
|
[False, 3],
|
|
[True, 0],
|
|
[True, 1],
|
|
[True, 2],
|
|
[True, 3],
|
|
]
|
|
|
|
result = []
|
|
|
|
for flip, rok in tqdm(args):
|
|
feed_pixel = pixel.clone()
|
|
feed_latent = latent.clone()
|
|
|
|
if flip:
|
|
feed_pixel = torch.flip(feed_pixel, dims=(3,))
|
|
feed_latent = torch.flip(feed_latent, dims=(3,))
|
|
|
|
feed_pixel = torch.rot90(feed_pixel, k=rok, dims=(2, 3))
|
|
feed_latent = torch.rot90(feed_latent, k=rok, dims=(2, 3))
|
|
|
|
eps = self.estimate_single_pass(feed_pixel, feed_latent).clip(0, 1)
|
|
eps = torch.rot90(eps, k=-rok, dims=(2, 3))
|
|
|
|
if flip:
|
|
eps = torch.flip(eps, dims=(3,))
|
|
|
|
result += [eps]
|
|
|
|
result = torch.stack(result, dim=0)
|
|
median = torch.median(result, dim=0).values
|
|
return median
|
|
|
|
@torch.no_grad()
|
|
def decode_pixel(
|
|
self, pixel: torch.TensorType, latent: torch.TensorType
|
|
) -> torch.TensorType:
|
|
# pixel.shape = [B, C=3, H, W]
|
|
assert pixel.shape[1] == 3
|
|
pixel_device = pixel.device
|
|
pixel_dtype = pixel.dtype
|
|
|
|
pixel = pixel.to(device=self.load_device, dtype=self.dtype)
|
|
latent = latent.to(device=self.load_device, dtype=self.dtype)
|
|
# y.shape = [B, C=4, H, W]
|
|
y = self.estimate_augmented(pixel, latent)
|
|
y = y.clip(0, 1)
|
|
assert y.shape[1] == 4
|
|
# Restore image to original device of input image.
|
|
return y.to(pixel_device, dtype=pixel_dtype)
|
|
|
|
|
|
def calculate_weight_adjust_channel(func):
|
|
@functools.wraps(func)
|
|
def calculate_weight(
|
|
self: ModelPatcher, patches, weight: torch.Tensor, key: str
|
|
) -> torch.Tensor:
|
|
weight = func(self, patches, weight, key)
|
|
|
|
for p in patches:
|
|
alpha = p[0]
|
|
v = p[1]
|
|
|
|
if isinstance(v, list):
|
|
v = (func(v[1:], v[0].clone(), key),)
|
|
|
|
if len(v) == 1:
|
|
patch_type = "diff"
|
|
elif len(v) == 2:
|
|
patch_type = v[0]
|
|
v = v[1]
|
|
|
|
if patch_type == "diff":
|
|
w1 = v[0]
|
|
if all(
|
|
(
|
|
alpha != 0.0,
|
|
w1.shape != weight.shape,
|
|
w1.ndim == weight.ndim == 4,
|
|
)
|
|
):
|
|
new_shape = [max(n, m) for n, m in zip(weight.shape, w1.shape)]
|
|
print(
|
|
f"Merged with {key} channel changed from {weight.shape} to {new_shape}"
|
|
)
|
|
new_diff = alpha * comfy.model_management.cast_to_device(
|
|
w1, weight.device, weight.dtype
|
|
)
|
|
new_weight = torch.zeros(size=new_shape).to(weight)
|
|
new_weight[
|
|
: weight.shape[0],
|
|
: weight.shape[1],
|
|
: weight.shape[2],
|
|
: weight.shape[3],
|
|
] = weight
|
|
new_weight[
|
|
: new_diff.shape[0],
|
|
: new_diff.shape[1],
|
|
: new_diff.shape[2],
|
|
: new_diff.shape[3],
|
|
] += new_diff
|
|
new_weight = new_weight.contiguous().clone()
|
|
weight = new_weight
|
|
return weight
|
|
|
|
return calculate_weight
|
|
|
|
except ImportError:
|
|
ModelMixin = None
|
|
ConfigMixin = None
|
|
TransparentVAEDecoder = None
|
|
calculate_weight_adjust_channel = None
|
|
print("\33[31mModule 'diffusers' not installed. Please install it via:\033[0m")
|
|
print("\33[31mpip install diffusers\033[0m")
|
|
|
|
|
|
from comfy.utils import load_torch_file
|
|
from comfy.conds import CONDRegular
|
|
from comfy_extras.nodes_compositing import JoinImageWithAlpha
|
|
from .config import LAYER_DIFFUSION, LAYER_DIFFUSION_DIR, LAYER_DIFFUSION_VAE
|
|
from .libs.utils import to_lora_patch_dict, get_local_filepath
|
|
class LayerDiffuse:
|
|
|
|
def __init__(self) -> None:
|
|
self.vae_transparent_decoder = None
|
|
self.vae_transparent_encoder = None
|
|
|
|
def get_layer_diffusion_method(self, method, has_blend_latent):
|
|
method = LayerMethod(method)
|
|
if method == LayerMethod.BG_TO_BLEND and has_blend_latent:
|
|
method = LayerMethod.BG_BLEND_TO_FG
|
|
elif method == LayerMethod.FG_TO_BLEND and has_blend_latent:
|
|
method = LayerMethod.FG_BLEND_TO_BG
|
|
return method
|
|
|
|
def apply_layer_c_concat(self, cond, uncond, c_concat):
|
|
def write_c_concat(cond):
|
|
new_cond = []
|
|
for t in cond:
|
|
n = [t[0], t[1].copy()]
|
|
if "model_conds" not in n[1]:
|
|
n[1]["model_conds"] = {}
|
|
n[1]["model_conds"]["c_concat"] = CONDRegular(c_concat)
|
|
new_cond.append(n)
|
|
return new_cond
|
|
|
|
return (write_c_concat(cond), write_c_concat(uncond))
|
|
|
|
def apply_layer_diffusion(self, model: ModelPatcher, method, weight, samples, blend_samples, positive, negative):
|
|
|
|
model_file = get_local_filepath(LAYER_DIFFUSION[method.value]["model_url"], LAYER_DIFFUSION_DIR)
|
|
layer_lora_state_dict = load_torch_file(model_file)
|
|
layer_lora_patch_dict = to_lora_patch_dict(layer_lora_state_dict)
|
|
work_model = model.clone()
|
|
work_model.add_patches(layer_lora_patch_dict, weight)
|
|
|
|
# cond_contact
|
|
if method in [LayerMethod.FG_ONLY_ATTN, LayerMethod.FG_ONLY_CONV]:
|
|
samp_model = work_model
|
|
else:
|
|
if method in [LayerMethod.BG_TO_BLEND, LayerMethod.FG_TO_BLEND]:
|
|
c_concat = model.model.latent_format.process_in(samples["samples"])
|
|
else:
|
|
c_concat = model.model.latent_format.process_in(torch.cat([samples["samples"], blend_samples["samples"]], dim=1))
|
|
samp_model, positive, negative = (work_model,) + self.apply_layer_c_concat(positive, negative, c_concat)
|
|
|
|
return samp_model, positive, negative
|
|
|
|
def join_image_with_alpha(self, image, alpha):
|
|
out = image.movedim(-1, 1)
|
|
if out.shape[1] == 3: # RGB
|
|
out = torch.cat([out, torch.ones_like(out[:, :1, :, :])], dim=1)
|
|
for i in range(out.shape[0]):
|
|
out[i, 3, :, :] = alpha
|
|
return out.movedim(1, -1)
|
|
|
|
def layer_diffusion_decode(self, layer_diffusion_method, latent, blend_samples, samp_images):
|
|
alpha = None
|
|
if layer_diffusion_method is not None:
|
|
method = self.get_layer_diffusion_method(layer_diffusion_method, blend_samples is not None)
|
|
print(method.value)
|
|
if method in [LayerMethod.FG_ONLY_CONV, LayerMethod.FG_ONLY_ATTN, LayerMethod.BG_BLEND_TO_FG]:
|
|
if self.vae_transparent_decoder is None:
|
|
decoder_file = get_local_filepath(LAYER_DIFFUSION_VAE['decode']["model_url"], LAYER_DIFFUSION_DIR)
|
|
self.vae_transparent_decoder = TransparentVAEDecoder(
|
|
load_torch_file(decoder_file),
|
|
device=comfy.model_management.get_torch_device(),
|
|
dtype=(torch.float16 if comfy.model_management.should_use_fp16() else torch.float32),
|
|
)
|
|
|
|
pixel = samp_images.movedim(-1, 1) # [B, H, W, C] => [B, C, H, W]
|
|
decoded = []
|
|
sub_batch_size = 16
|
|
for start_idx in range(0, latent.shape[0], sub_batch_size):
|
|
decoded.append(
|
|
self.vae_transparent_decoder.decode_pixel(
|
|
pixel[start_idx: start_idx + sub_batch_size],
|
|
latent[start_idx: start_idx + sub_batch_size],
|
|
)
|
|
)
|
|
pixel_with_alpha = torch.cat(decoded, dim=0)
|
|
# [B, C, H, W] => [B, H, W, C]
|
|
pixel_with_alpha = pixel_with_alpha.movedim(1, -1)
|
|
image = pixel_with_alpha[..., 1:]
|
|
alpha = pixel_with_alpha[..., 0]
|
|
|
|
alpha = 1.0 - alpha
|
|
new_images, = JoinImageWithAlpha().join_image_with_alpha(image, alpha)
|
|
else:
|
|
new_images = samp_images
|
|
else:
|
|
new_images = samp_images
|
|
return (new_images, samp_images, alpha) |