File size: 38,327 Bytes
d234621 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 |
import torch
import torch.nn as nn
import torch.nn.functional as F
import math
from typing import Tuple, Optional
from einops import rearrange
from .utils import hash_state_dict_keys
try:
import flash_attn_interface
FLASH_ATTN_3_AVAILABLE = True
except ModuleNotFoundError:
FLASH_ATTN_3_AVAILABLE = False
try:
import flash_attn
FLASH_ATTN_2_AVAILABLE = True
except ModuleNotFoundError:
FLASH_ATTN_2_AVAILABLE = False
try:
from sageattention import sageattn
SAGE_ATTN_AVAILABLE = True
except ModuleNotFoundError:
SAGE_ATTN_AVAILABLE = False
def flash_attention(q: torch.Tensor, k: torch.Tensor, v: torch.Tensor, num_heads: int, compatibility_mode=False, causal=False):
if compatibility_mode:
q = rearrange(q, "b s (n d) -> b n s d", n=num_heads)
k = rearrange(k, "b s (n d) -> b n s d", n=num_heads)
v = rearrange(v, "b s (n d) -> b n s d", n=num_heads)
x = F.scaled_dot_product_attention(q, k, v)
x = rearrange(x, "b n s d -> b s (n d)", n=num_heads)
elif FLASH_ATTN_3_AVAILABLE:
q = rearrange(q, "b s (n d) -> b s n d", n=num_heads)
k = rearrange(k, "b s (n d) -> b s n d", n=num_heads)
v = rearrange(v, "b s (n d) -> b s n d", n=num_heads)
x = flash_attn_interface.flash_attn_func(q, k, v)
x = rearrange(x, "b s n d -> b s (n d)", n=num_heads)
elif FLASH_ATTN_2_AVAILABLE:
q = rearrange(q, "b s (n d) -> b s n d", n=num_heads)
k = rearrange(k, "b s (n d) -> b s n d", n=num_heads)
v = rearrange(v, "b s (n d) -> b s n d", n=num_heads)
x = flash_attn.flash_attn_func(q, k, v)
x = rearrange(x, "b s n d -> b s (n d)", n=num_heads)
elif SAGE_ATTN_AVAILABLE:
q = rearrange(q, "b s (n d) -> b n s d", n=num_heads)
k = rearrange(k, "b s (n d) -> b n s d", n=num_heads)
v = rearrange(v, "b s (n d) -> b n s d", n=num_heads)
x = sageattn(q, k, v)
x = rearrange(x, "b n s d -> b s (n d)", n=num_heads)
else:
q = rearrange(q, "b s (n d) -> b n s d", n=num_heads)
k = rearrange(k, "b s (n d) -> b n s d", n=num_heads)
v = rearrange(v, "b s (n d) -> b n s d", n=num_heads)
x = F.scaled_dot_product_attention(q, k, v)
x = rearrange(x, "b n s d -> b s (n d)", n=num_heads)
return x
def modulate(x: torch.Tensor, shift: torch.Tensor, scale: torch.Tensor):
return (x * (1 + scale) + shift)
def sinusoidal_embedding_1d(dim, position):
sinusoid = torch.outer(position.type(torch.float64), torch.pow(
10000, -torch.arange(dim//2, dtype=torch.float64, device=position.device).div(dim//2)))
x = torch.cat([torch.cos(sinusoid), torch.sin(sinusoid)], dim=1)
return x.to(position.dtype)
def precompute_freqs_cis_3d(dim: int, end: int = 1024, theta: float = 10000.0):
# 3d rope precompute
f_freqs_cis = precompute_freqs_cis(dim - 2 * (dim // 3), end, theta)
h_freqs_cis = precompute_freqs_cis(dim // 3, end, theta)
w_freqs_cis = precompute_freqs_cis(dim // 3, end, theta)
return f_freqs_cis, h_freqs_cis, w_freqs_cis
def precompute_freqs_cis(dim: int, end: int = 1024, theta: float = 10000.0):
# 1d rope precompute
freqs = 1.0 / (theta ** (torch.arange(0, dim, 2)
[: (dim // 2)].double() / dim))
freqs = torch.outer(torch.arange(end, device=freqs.device), freqs)
freqs_cis = torch.polar(torch.ones_like(freqs), freqs) # complex64
return freqs_cis
def rope_apply(x, freqs, num_heads):
x = rearrange(x, "b s (n d) -> b s n d", n=num_heads)
x_out = torch.view_as_complex(x.to(torch.float64).reshape(
x.shape[0], x.shape[1], x.shape[2], -1, 2))
x_out = torch.view_as_real(x_out * freqs).flatten(2)
return x_out.to(x.dtype)
class RMSNorm(nn.Module):
def __init__(self, dim, eps=1e-5):
super().__init__()
self.eps = eps
self.weight = nn.Parameter(torch.ones(dim))
def norm(self, x):
return x * torch.rsqrt(x.pow(2).mean(dim=-1, keepdim=True) + self.eps)
def forward(self, x):
dtype = x.dtype
return self.norm(x.float()).to(dtype) * self.weight
class AttentionModule(nn.Module):
def __init__(self, num_heads, causal=False):
super().__init__()
self.num_heads = num_heads
def forward(self, q, k, v):
x = flash_attention(q=q, k=k, v=v, num_heads=self.num_heads)
return x
class SelfAttention(nn.Module):
def __init__(self, dim: int, num_heads: int, eps: float = 1e-6, causal: bool = False):
super().__init__()
self.dim = dim
self.num_heads = num_heads
self.head_dim = dim // num_heads
self.q = nn.Linear(dim, dim)
self.k = nn.Linear(dim, dim)
self.v = nn.Linear(dim, dim)
self.o = nn.Linear(dim, dim)
self.norm_q = RMSNorm(dim, eps=eps)
self.norm_k = RMSNorm(dim, eps=eps)
self.attn = AttentionModule(self.num_heads)
def forward(self, x, freqs):
x = x.to(self.q.weight.dtype)
q = self.norm_q(self.q(x))
k = self.norm_k(self.k(x))
v = self.v(x)
q = rope_apply(q, freqs, self.num_heads)
k = rope_apply(k, freqs, self.num_heads)
x = self.attn(q, k, v)
return self.o(x)
class CrossAttention(nn.Module):
def __init__(self, dim: int, num_heads: int, eps: float = 1e-6, has_image_input: bool = False):
super().__init__()
self.dim = dim
self.num_heads = num_heads
self.head_dim = dim // num_heads
self.q = nn.Linear(dim, dim)
self.k = nn.Linear(dim, dim)
self.v = nn.Linear(dim, dim)
self.o = nn.Linear(dim, dim)
self.norm_q = RMSNorm(dim, eps=eps)
self.norm_k = RMSNorm(dim, eps=eps)
self.has_image_input = has_image_input
if has_image_input:
self.k_img = nn.Linear(dim, dim)
self.v_img = nn.Linear(dim, dim)
self.norm_k_img = RMSNorm(dim, eps=eps)
self.attn = AttentionModule(self.num_heads)
def forward(self, x: torch.Tensor, y: torch.Tensor):
if self.has_image_input:
img = y[:, :257]
ctx = y[:, 257:]
else:
ctx = y
q = self.norm_q(self.q(x))
k = self.norm_k(self.k(ctx))
v = self.v(ctx)
x = self.attn(q, k, v)
if self.has_image_input:
k_img = self.norm_k_img(self.k_img(img))
v_img = self.v_img(img)
y = flash_attention(q, k_img, v_img, num_heads=self.num_heads)
x = x + y
return self.o(x)
class DiTBlock(nn.Module):
def __init__(self, has_image_input: bool, dim: int, num_heads: int, ffn_dim: int, eps: float = 1e-6):
super().__init__()
self.dim = dim
self.num_heads = num_heads
self.ffn_dim = ffn_dim
#self.self_attn = SelfAttention(dim, num_heads, eps, causal=True) # Enable causal masking
self.self_attn = SelfAttention(dim, num_heads, eps)
self.cross_attn = CrossAttention(
dim, num_heads, eps, has_image_input=has_image_input)
self.norm1 = nn.LayerNorm(dim, eps=eps, elementwise_affine=False)
self.norm2 = nn.LayerNorm(dim, eps=eps, elementwise_affine=False)
self.norm3 = nn.LayerNorm(dim, eps=eps)
self.ffn = nn.Sequential(nn.Linear(dim, ffn_dim), nn.GELU(
approximate='tanh'), nn.Linear(ffn_dim, dim))
self.modulation = nn.Parameter(torch.randn(1, 6, dim) / dim**0.5)
def forward(self, x, context, cam_emb, t_mod, freqs):
# msa: multi-head self-attention mlp: multi-layer perceptron
shift_msa, scale_msa, gate_msa, shift_mlp, scale_mlp, gate_mlp = (
self.modulation.to(dtype=t_mod.dtype, device=t_mod.device) + t_mod).chunk(6, dim=1)
input_x = modulate(self.norm1(x), shift_msa, scale_msa)
if cam_emb is not None:
# 🔧 简化:cam_emb已经在process_input_hidden_states中处理过空间扩展和重排列
cam_emb = cam_emb.to(self.cam_encoder.weight.dtype)
cam_emb = self.cam_encoder(cam_emb) # [batch, seq_len, dim]
input_x = input_x + cam_emb
# Ensure input_x dtype matches self.projector.weight dtype
input_x = input_x.to(self.projector.weight.dtype)
# Ensure self.self_attn output dtype matches self.projector.weight dtype
attn_output = self.self_attn(input_x, freqs)
attn_output = attn_output.to(self.projector.weight.dtype)
x = x + gate_msa * self.projector(attn_output)
x = x.to(self.norm3.weight.dtype)
x = x + self.cross_attn(self.norm3(x), context)
input_x = modulate(self.norm2(x), shift_mlp, scale_mlp)
x = x + gate_mlp * self.ffn(input_x)
return x
class MLP(torch.nn.Module):
def __init__(self, in_dim, out_dim):
super().__init__()
self.proj = torch.nn.Sequential(
nn.LayerNorm(in_dim),
nn.Linear(in_dim, in_dim),
nn.GELU(),
nn.Linear(in_dim, out_dim),
nn.LayerNorm(out_dim)
)
def forward(self, x):
return self.proj(x)
class Head(nn.Module):
def __init__(self, dim: int, out_dim: int, patch_size: Tuple[int, int, int], eps: float):
super().__init__()
self.dim = dim
self.patch_size = patch_size
self.norm = nn.LayerNorm(dim, eps=eps, elementwise_affine=False)
self.head = nn.Linear(dim, out_dim * math.prod(patch_size))
self.modulation = nn.Parameter(torch.randn(1, 2, dim) / dim**0.5)
def forward(self, x, t_mod):
shift, scale = (self.modulation.to(dtype=t_mod.dtype, device=t_mod.device) + t_mod).chunk(2, dim=1)
x = (self.head(self.norm(x) * (1 + scale) + shift))
return x
class WanModelFuture4(torch.nn.Module):
def __init__(
self,
dim: int,
in_dim: int,
ffn_dim: int,
out_dim: int,
text_dim: int,
freq_dim: int,
eps: float,
patch_size: Tuple[int, int, int],
num_heads: int,
num_layers: int,
has_image_input: bool,
):
super().__init__()
self.dim = dim
self.freq_dim = freq_dim
self.has_image_input = has_image_input
self.patch_size = patch_size
self.patch_embedding = nn.Conv3d(
in_dim, dim, kernel_size=patch_size, stride=patch_size)
self.text_embedding = nn.Sequential(
nn.Linear(text_dim, dim),
nn.GELU(approximate='tanh'),
nn.Linear(dim, dim)
)
self.time_embedding = nn.Sequential(
nn.Linear(freq_dim, dim),
nn.SiLU(),
nn.Linear(dim, dim)
)
self.time_projection = nn.Sequential(
nn.SiLU(), nn.Linear(dim, dim * 6))
self.blocks = nn.ModuleList([
DiTBlock(has_image_input, dim, num_heads, ffn_dim, eps)
for _ in range(num_layers)
])
self.head = Head(dim, out_dim, patch_size, eps)
head_dim = dim // num_heads
self.freqs = precompute_freqs_cis_3d(head_dim)
if has_image_input:
self.img_emb = MLP(1280, dim) # clip_feature_dim = 1280
def patchify(self, x: torch.Tensor):
x = self.patch_embedding(x)
grid_size = x.shape[2:]
x = rearrange(x, 'b c f h w -> b (f h w) c').contiguous()
return x, grid_size # x, grid_size: (f, h, w)
def unpatchify(self, x: torch.Tensor, grid_size: torch.Tensor):
return rearrange(
x, 'b (f h w) (x y z c) -> b c (f x) (h y) (w z)',
f=grid_size[0], h=grid_size[1], w=grid_size[2],
x=self.patch_size[0], y=self.patch_size[1], z=self.patch_size[2]
)
def create_clean_x_embedder(self):
"""创建类似FramePack的clean_x_embedder"""
class CleanXEmbedder(nn.Module):
def __init__(self, inner_dim):
super().__init__()
# 参考hunyuan_video_packed.py的设计
self.proj = nn.Conv3d(16, inner_dim, kernel_size=(1, 2, 2), stride=(1, 2, 2))
self.proj_2x = nn.Conv3d(16, inner_dim, kernel_size=(2, 4, 4), stride=(2, 4, 4))
self.proj_4x = nn.Conv3d(16, inner_dim, kernel_size=(4, 8, 8), stride=(4, 8, 8))
def forward(self, x, scale="1x"):
if scale == "1x":
return self.proj(x)
elif scale == "2x":
return self.proj_2x(x)
elif scale == "4x":
return self.proj_4x(x)
else:
raise ValueError(f"Unsupported scale: {scale}")
return CleanXEmbedder(self.dim)
def rope(self, frame_indices, height, width, device):
"""🔧 模仿HunyuanVideo的rope方法"""
batch_size = frame_indices.shape[0]
seq_len = frame_indices.shape[1]
# 使用frame_indices生成时间维度的频率
f_freqs = self.freqs[0][frame_indices.to("cpu")] # [batch, seq_len, freq_dim]
# 为每个spatial位置生成频率
h_positions = torch.arange(height, device=device).unsqueeze(0).unsqueeze(0).expand(batch_size, seq_len, -1)
w_positions = torch.arange(width, device=device).unsqueeze(0).unsqueeze(0).expand(batch_size, seq_len, -1)
# 获取h和w的频率
h_freqs = self.freqs[1][h_positions.to("cpu")] # [batch, seq_len, height, h_freq_dim]
w_freqs = self.freqs[2][w_positions.to("cpu")] # [batch, seq_len, width, w_freq_dim]
# 扩展到完整的spatial grid
f_freqs_expanded = f_freqs.unsqueeze(2).unsqueeze(3).expand(-1, -1, height, width, -1)
h_freqs_expanded = h_freqs.unsqueeze(3).expand(-1, -1, -1, width, -1)
w_freqs_expanded = w_freqs.unsqueeze(2).expand(-1, -1, height, -1, -1)
# 合并所有频率
rope_freqs = torch.cat([f_freqs_expanded, h_freqs_expanded, w_freqs_expanded], dim=-1)
return rope_freqs # [batch, seq_len, height, width, total_freq_dim]
def pad_for_3d_conv(self, x, kernel_size):
"""3D卷积的padding - 参考hunyuan实现"""
if len(x.shape) == 5: # [B, C, T, H, W]
b, c, t, h, w = x.shape
pt, ph, pw = kernel_size
pad_t = (pt - (t % pt)) % pt
pad_h = (ph - (h % ph)) % ph
pad_w = (pw - (w % pw)) % pw
return torch.nn.functional.pad(x, (0, pad_w, 0, pad_h, 0, pad_t), mode='replicate')
elif len(x.shape) == 6: # [B, T, H, W, C] (RoPE频率)
b, t, h, w, c = x.shape
pt, ph, pw = kernel_size
pad_t = (pt - (t % pt)) % pt
pad_h = (ph - (h % ph)) % ph
pad_w = (pw - (w % pw)) % pw
return torch.nn.functional.pad(x, (0, 0, 0, pad_w, 0, pad_h, 0, pad_t), mode='replicate')
else:
raise ValueError(f"Unsupported tensor shape: {x.shape}")
def center_down_sample_3d(self, x, scale_factor):
"""🔧 模仿HunyuanVideo的center_down_sample_3d"""
if len(x.shape) == 6: # [B, T, H, W, C] (RoPE频率)
st, sh, sw = scale_factor
return x[:, ::st, ::sh, ::sw, :]
elif len(x.shape) == 5: # [B, C, T, H, W]
st, sh, sw = scale_factor
return x[:, :, ::st, ::sh, ::sw]
else:
raise ValueError(f"Unsupported tensor shape: {x.shape}")
def process_input_hidden_states(self,
latents, latent_indices=None,
clean_latents=None, clean_latent_indices=None,
clean_latents_2x=None, clean_latent_2x_indices=None,
clean_latents_4x=None, clean_latent_4x_indices=None,
cam_emb=None):
"""🔧 处理FramePack风格的多尺度输入 - 修改clean_latents为起始4帧+最后1帧"""
# 主要latents处理
hidden_states, grid_size = self.patchify(latents)
B, T_patches, C = hidden_states.shape
f, h, w = grid_size
# 🔧 修正:使用latent_indices指定的时间位置计算RoPE频率
if latent_indices is None:
latent_indices = torch.arange(0, f, device=hidden_states.device).unsqueeze(0).expand(B, -1)
# 为主要latents计算RoPE频率
main_rope_freqs_list = []
for b in range(B):
batch_rope_freqs = []
for t_idx in latent_indices[b]:
f_freq = self.freqs[0][t_idx:t_idx+1]
h_freq = self.freqs[1][:h]
w_freq = self.freqs[2][:w]
spatial_freqs = torch.cat([
f_freq.view(1, 1, 1, -1).expand(1, h, w, -1),
h_freq.view(1, h, 1, -1).expand(1, h, w, -1),
w_freq.view(1, 1, w, -1).expand(1, h, w, -1)
], dim=-1).reshape(h * w, -1)
batch_rope_freqs.append(spatial_freqs)
batch_rope_freqs = torch.cat(batch_rope_freqs, dim=0)
main_rope_freqs_list.append(batch_rope_freqs)
rope_freqs = torch.stack(main_rope_freqs_list, dim=0)
# 🔧 准备camera embeddings - 直接用真实索引
combined_camera_embeddings = None
if cam_emb is not None and clean_latent_indices is not None:
start_indice = clean_latent_indices[0][0].item()
# 提取target部分的camera(基于latent_indices)
target_start = latent_indices[0].min().item() - start_indice
target_end = latent_indices[0].max().item() + 1 - start_indice
target_camera = cam_emb[:, target_start:target_end, :]
# 为主要latents处理camera空间扩展
target_camera_spatial = target_camera.unsqueeze(2).unsqueeze(3).repeat(1, 1, h, w, 1)
target_camera_spatial = rearrange(target_camera_spatial, 'b f h w d -> b (f h w) d')
combined_camera_embeddings = target_camera_spatial
# 🔧 处理clean_latents (1x scale) - 修改为起始4帧+最后1帧结构
if clean_latents is not None and clean_latent_indices is not None:
clean_latents = clean_latents.to(hidden_states)
clean_hidden_states = self.clean_x_embedder(clean_latents, scale="1x")
clean_hidden_states = rearrange(clean_hidden_states, 'b c f h w -> b (f h w) c')
# 🔧 为clean_latents计算RoPE频率 - 现在clean_latents是5帧(起始4帧+最后1帧)
clean_rope_freqs_list = []
for b in range(B):
clean_batch_rope_freqs = []
# 🔧 处理clean_latent_indices:应该包含5帧的索引
# 前4帧是起始帧,第5帧是最后1帧
for i, t_idx in enumerate(clean_latent_indices[b]):
if t_idx >= 0: # 有效索引
f_freq = self.freqs[0][t_idx:t_idx+1].to(hidden_states.device) # 🔧 确保设备一致
h_freq = self.freqs[1][:h].to(hidden_states.device) # 🔧 确保设备一致
w_freq = self.freqs[2][:w].to(hidden_states.device) # 🔧 确保设备一致
spatial_freqs = torch.cat([
f_freq.view(1, 1, 1, -1).expand(1, h, w, -1),
h_freq.view(1, h, 1, -1).expand(1, h, w, -1),
w_freq.view(1, 1, w, -1).expand(1, h, w, -1)
], dim=-1).reshape(h * w, -1)
else:
# 无效索引(-1),使用0频率
spatial_freqs = torch.zeros(h * w, f_freq.shape[-1] + h_freq.shape[-1] + w_freq.shape[-1],
device=hidden_states.device, dtype=hidden_states.dtype)
clean_batch_rope_freqs.append(spatial_freqs)
clean_batch_rope_freqs = torch.cat(clean_batch_rope_freqs, dim=0)
clean_rope_freqs_list.append(clean_batch_rope_freqs)
clean_rope_freqs = torch.stack(clean_rope_freqs_list, dim=0)
# 处理clean camera embeddings
if cam_emb is not None:
# 🔧 直接用真实索引提取camera - 现在是5帧的camera embedding
clean_camera_indices = []
for idx in clean_latent_indices[0]:
if idx >= 0: # 有效索引
clean_camera_indices.append(idx.item() - start_indice)
else:
# 无效索引,后面会用0填充
clean_camera_indices.append(0) # 临时占位
clean_camera = cam_emb[:, clean_camera_indices, :]
# 🔧 对无效位置清零
for i, idx in enumerate(clean_latent_indices[0]):
if idx < 0:
clean_camera[:, i, :] = 0
clean_camera_spatial = clean_camera.unsqueeze(2).unsqueeze(3).repeat(1, 1, h, w, 1)
clean_camera_spatial = rearrange(clean_camera_spatial, 'b f h w d -> b (f h w) d')
combined_camera_embeddings = torch.cat([clean_camera_spatial, combined_camera_embeddings], dim=1)
hidden_states = torch.cat([clean_hidden_states, hidden_states], dim=1)
rope_freqs = torch.cat([clean_rope_freqs.to(hidden_states.device), rope_freqs.to(hidden_states.device)], dim=1)
# 🔧 处理clean_latents_2x (2x scale) - 修正RoPE频率计算
if clean_latents_2x is not None and clean_latent_2x_indices is not None and clean_latent_2x_indices.numel() > 0:
# 过滤有效索引(非-1)
valid_2x_indices = clean_latent_2x_indices[clean_latent_2x_indices >= 0]
if len(valid_2x_indices) > 0:
clean_latents_2x = clean_latents_2x.to(hidden_states)
clean_latents_2x = self.pad_for_3d_conv(clean_latents_2x, (2, 4, 4))
clean_hidden_states_2x = self.clean_x_embedder(clean_latents_2x, scale="2x")
_, _, clean_2x_f, clean_2x_h, clean_2x_w = clean_hidden_states_2x.shape
clean_hidden_states_2x = rearrange(clean_hidden_states_2x, 'b c f h w -> b (f h w) c')
# 🔧 为2x latents计算RoPE频率 - 基于实际的下采样结果
clean_2x_rope_freqs_list = []
for b in range(B):
clean_2x_batch_rope_freqs = []
# 🔧 修正:使用clean_2x_f作为实际的时间帧数
for frame_idx in range(clean_2x_f):
# 计算对应的原始时间索引
if frame_idx < len(valid_2x_indices):
t_idx = valid_2x_indices[frame_idx]
else:
# 如果超出有效索引,使用0频率
t_idx = valid_2x_indices[-1] if len(valid_2x_indices) > 0 else 0
f_freq = self.freqs[0][t_idx:t_idx+1]
h_freq = self.freqs[1][:clean_2x_h]
w_freq = self.freqs[2][:clean_2x_w]
spatial_freqs = torch.cat([
f_freq.view(1, 1, 1, -1).expand(1, clean_2x_h, clean_2x_w, -1),
h_freq.view(1, clean_2x_h, 1, -1).expand(1, clean_2x_h, clean_2x_w, -1),
w_freq.view(1, 1, clean_2x_w, -1).expand(1, clean_2x_h, clean_2x_w, -1)
], dim=-1).reshape(clean_2x_h * clean_2x_w, -1)
clean_2x_batch_rope_freqs.append(spatial_freqs)
clean_2x_batch_rope_freqs = torch.cat(clean_2x_batch_rope_freqs, dim=0)
clean_2x_rope_freqs_list.append(clean_2x_batch_rope_freqs)
clean_2x_rope_freqs = torch.stack(clean_2x_rope_freqs_list, dim=0)
# 🔧 处理2x camera embeddings
if cam_emb is not None:
# 创建2x camera,0填充无效部分
clean_2x_camera = torch.zeros(B, clean_2x_f, cam_emb.shape[-1], dtype=cam_emb.dtype, device=cam_emb.device)
for frame_idx in range(min(clean_2x_f, len(valid_2x_indices))):
cam_idx = valid_2x_indices[frame_idx].item() - start_indice
if 0 <= cam_idx < cam_emb.shape[1]:
clean_2x_camera[:, frame_idx, :] = cam_emb[:, cam_idx, :]
clean_2x_camera_spatial = clean_2x_camera.unsqueeze(2).unsqueeze(3).repeat(1, 1, clean_2x_h, clean_2x_w, 1)
clean_2x_camera_spatial = rearrange(clean_2x_camera_spatial, 'b f h w d -> b (f h w) d')
combined_camera_embeddings = torch.cat([clean_2x_camera_spatial, combined_camera_embeddings], dim=1)
hidden_states = torch.cat([clean_hidden_states_2x, hidden_states], dim=1)
rope_freqs = torch.cat([clean_2x_rope_freqs.to(rope_freqs.device), rope_freqs], dim=1)
# 🔧 处理clean_latents_4x (4x scale) - 修正RoPE频率计算
if clean_latents_4x is not None and clean_latent_4x_indices is not None and clean_latent_4x_indices.numel() > 0:
# 过滤有效索引(非-1)
valid_4x_indices = clean_latent_4x_indices[clean_latent_4x_indices >= 0]
if len(valid_4x_indices) > 0:
clean_latents_4x = clean_latents_4x.to(hidden_states)
clean_latents_4x = self.pad_for_3d_conv(clean_latents_4x, (4, 8, 8))
clean_hidden_states_4x = self.clean_x_embedder(clean_latents_4x, scale="4x")
_, _, clean_4x_f, clean_4x_h, clean_4x_w = clean_hidden_states_4x.shape
clean_hidden_states_4x = rearrange(clean_hidden_states_4x, 'b c f h w -> b (f h w) c')
# 🔧 为4x latents计算RoPE频率 - 基于实际的下采样结果
clean_4x_rope_freqs_list = []
for b in range(B):
clean_4x_batch_rope_freqs = []
# 🔧 修正:使用clean_4x_f作为实际的时间帧数
for frame_idx in range(clean_4x_f):
# 计算对应的原始时间索引
if frame_idx < len(valid_4x_indices):
t_idx = valid_4x_indices[frame_idx]
else:
# 如果超出有效索引,使用0频率
t_idx = valid_4x_indices[-1] if len(valid_4x_indices) > 0 else 0
f_freq = self.freqs[0][t_idx:t_idx+1]
h_freq = self.freqs[1][:clean_4x_h]
w_freq = self.freqs[2][:clean_4x_w]
spatial_freqs = torch.cat([
f_freq.view(1, 1, 1, -1).expand(1, clean_4x_h, clean_4x_w, -1),
h_freq.view(1, clean_4x_h, 1, -1).expand(1, clean_4x_h, clean_4x_w, -1),
w_freq.view(1, 1, clean_4x_w, -1).expand(1, clean_4x_h, clean_4x_w, -1)
], dim=-1).reshape(clean_4x_h * clean_4x_w, -1)
clean_4x_batch_rope_freqs.append(spatial_freqs)
clean_4x_batch_rope_freqs = torch.cat(clean_4x_batch_rope_freqs, dim=0)
clean_4x_rope_freqs_list.append(clean_4x_batch_rope_freqs)
clean_4x_rope_freqs = torch.stack(clean_4x_rope_freqs_list, dim=0)
# 🔧 处理4x camera embeddings
if cam_emb is not None:
# 创建4x camera,0填充无效部分
clean_4x_camera = torch.zeros(B, clean_4x_f, cam_emb.shape[-1], dtype=cam_emb.dtype, device=cam_emb.device)
for frame_idx in range(min(clean_4x_f, len(valid_4x_indices))):
cam_idx = valid_4x_indices[frame_idx].item() - start_indice
if 0 <= cam_idx < cam_emb.shape[1]:
clean_4x_camera[:, frame_idx, :] = cam_emb[:, cam_idx, :]
clean_4x_camera_spatial = clean_4x_camera.unsqueeze(2).unsqueeze(3).repeat(1, 1, clean_4x_h, clean_4x_w, 1)
clean_4x_camera_spatial = rearrange(clean_4x_camera_spatial, 'b f h w d -> b (f h w) d')
combined_camera_embeddings = torch.cat([clean_4x_camera_spatial, combined_camera_embeddings], dim=1)
hidden_states = torch.cat([clean_hidden_states_4x, hidden_states], dim=1)
rope_freqs = torch.cat([clean_4x_rope_freqs.to(rope_freqs.device), rope_freqs], dim=1)
rope_freqs = rope_freqs.unsqueeze(2).to(device=hidden_states.device)
return hidden_states, rope_freqs, grid_size, combined_camera_embeddings
def forward(self,
latents, timestep, cam_emb,
# 🔧 FramePack参数
latent_indices=None,
clean_latents=None, clean_latent_indices=None,
clean_latents_2x=None, clean_latent_2x_indices=None,
clean_latents_4x=None, clean_latent_4x_indices=None,
**kwargs):
# 🔧 使用新的处理方法来处理多尺度输入和RoPE频率
hidden_states, rope_freqs, grid_size, processed_cam_emb = self.process_input_hidden_states(
latents, latent_indices,
clean_latents, clean_latent_indices,
clean_latents_2x, clean_latent_2x_indices,
clean_latents_4x, clean_latent_4x_indices,
cam_emb
)
# 计算原始latent序列长度(用于最后提取)
batch_size, num_channels, num_frames, height, width = latents.shape
p, p_t = self.patch_size[2], self.patch_size[0] # [t, h, w]
post_patch_num_frames = num_frames // p_t
post_patch_height = height // p
post_patch_width = width // p
original_context_length = post_patch_num_frames * post_patch_height * post_patch_width
# 处理其他embeddings
context = kwargs.get("context", None)
if context is not None:
context = self.text_embedding(context)
t = self.time_embedding(
sinusoidal_embedding_1d(self.freq_dim, timestep))
t_mod = self.time_projection(t).unflatten(1, (6, self.dim))
# 确保rope_freqs与hidden_states的序列长度匹配
assert rope_freqs.shape[1] == hidden_states.shape[1], \
f"RoPE频率序列长度 {rope_freqs.shape[1]} 与 hidden_states序列长度 {hidden_states.shape[1]} 不匹配"
# Transformer blocks
for block in self.blocks:
hidden_states = block(hidden_states, context, processed_cam_emb, t_mod, rope_freqs)
# 🔧 只对原始预测目标部分进行输出投影
# 提取最后original_context_length个tokens(对应原始latents)
hidden_states = hidden_states[:, -original_context_length:, :]
hidden_states = self.head(hidden_states, t)
hidden_states = self.unpatchify(hidden_states, grid_size)
return hidden_states
@staticmethod
def state_dict_converter():
return WanModelStateDictConverter()
class WanModelStateDictConverter:
def __init__(self):
pass
def from_diffusers(self, state_dict):
rename_dict = {
"blocks.0.attn1.norm_k.weight": "blocks.0.self_attn.norm_k.weight",
"blocks.0.attn1.norm_q.weight": "blocks.0.self_attn.norm_q.weight",
"blocks.0.attn1.to_k.bias": "blocks.0.self_attn.k.bias",
"blocks.0.attn1.to_k.weight": "blocks.0.self_attn.k.weight",
"blocks.0.attn1.to_out.0.bias": "blocks.0.self_attn.o.bias",
"blocks.0.attn1.to_out.0.weight": "blocks.0.self_attn.o.weight",
"blocks.0.attn1.to_q.bias": "blocks.0.self_attn.q.bias",
"blocks.0.attn1.to_q.weight": "blocks.0.self_attn.q.weight",
"blocks.0.attn1.to_v.bias": "blocks.0.self_attn.v.bias",
"blocks.0.attn1.to_v.weight": "blocks.0.self_attn.v.weight",
"blocks.0.attn2.norm_k.weight": "blocks.0.cross_attn.norm_k.weight",
"blocks.0.attn2.norm_q.weight": "blocks.0.cross_attn.norm_q.weight",
"blocks.0.attn2.to_k.bias": "blocks.0.cross_attn.k.bias",
"blocks.0.attn2.to_k.weight": "blocks.0.cross_attn.k.weight",
"blocks.0.attn2.to_out.0.bias": "blocks.0.cross_attn.o.bias",
"blocks.0.attn2.to_out.0.weight": "blocks.0.cross_attn.o.weight",
"blocks.0.attn2.to_q.bias": "blocks.0.cross_attn.q.bias",
"blocks.0.attn2.to_q.weight": "blocks.0.cross_attn.q.weight",
"blocks.0.attn2.to_v.bias": "blocks.0.cross_attn.v.bias",
"blocks.0.attn2.to_v.weight": "blocks.0.cross_attn.v.weight",
"blocks.0.ffn.net.0.proj.bias": "blocks.0.ffn.0.bias",
"blocks.0.ffn.net.0.proj.weight": "blocks.0.ffn.0.weight",
"blocks.0.ffn.net.2.bias": "blocks.0.ffn.2.bias",
"blocks.0.ffn.net.2.weight": "blocks.0.ffn.2.weight",
"blocks.0.norm2.bias": "blocks.0.norm3.bias",
"blocks.0.norm2.weight": "blocks.0.norm3.weight",
"blocks.0.scale_shift_table": "blocks.0.modulation",
"condition_embedder.text_embedder.linear_1.bias": "text_embedding.0.bias",
"condition_embedder.text_embedder.linear_1.weight": "text_embedding.0.weight",
"condition_embedder.text_embedder.linear_2.bias": "text_embedding.2.bias",
"condition_embedder.text_embedder.linear_2.weight": "text_embedding.2.weight",
"condition_embedder.time_embedder.linear_1.bias": "time_embedding.0.bias",
"condition_embedder.time_embedder.linear_1.weight": "time_embedding.0.weight",
"condition_embedder.time_embedder.linear_2.bias": "time_embedding.2.bias",
"condition_embedder.time_embedder.linear_2.weight": "time_embedding.2.weight",
"condition_embedder.time_proj.bias": "time_projection.1.bias",
"condition_embedder.time_proj.weight": "time_projection.1.weight",
"patch_embedding.bias": "patch_embedding.bias",
"patch_embedding.weight": "patch_embedding.weight",
"scale_shift_table": "head.modulation",
"proj_out.bias": "head.head.bias",
"proj_out.weight": "head.head.weight",
}
state_dict_ = {}
for name, param in state_dict.items():
if name in rename_dict:
state_dict_[rename_dict[name]] = param
else:
name_ = ".".join(name.split(".")[:1] + ["0"] + name.split(".")[2:])
if name_ in rename_dict:
name_ = rename_dict[name_]
name_ = ".".join(name_.split(".")[:1] + [name.split(".")[1]] + name_.split(".")[2:])
state_dict_[name_] = param
if hash_state_dict_keys(state_dict) == "cb104773c6c2cb6df4f9529ad5c60d0b":
config = {
"model_type": "t2v",
"patch_size": (1, 2, 2),
"text_len": 512,
"in_dim": 16,
"dim": 5120,
"ffn_dim": 13824,
"freq_dim": 256,
"text_dim": 4096,
"out_dim": 16,
"num_heads": 40,
"num_layers": 40,
"window_size": (-1, -1),
"qk_norm": True,
"cross_attn_norm": True,
"eps": 1e-6,
}
else:
config = {}
return state_dict_, config
def from_civitai(self, state_dict):
if hash_state_dict_keys(state_dict) == "9269f8db9040a9d860eaca435be61814":
config = {
"has_image_input": False,
"patch_size": [1, 2, 2],
"in_dim": 16,
"dim": 1536,
"ffn_dim": 8960,
"freq_dim": 256,
"text_dim": 4096,
"out_dim": 16,
"num_heads": 12,
"num_layers": 30,
"eps": 1e-6
}
elif hash_state_dict_keys(state_dict) == "aafcfd9672c3a2456dc46e1cb6e52c70":
config = {
"has_image_input": False,
"patch_size": [1, 2, 2],
"in_dim": 16,
"dim": 5120,
"ffn_dim": 13824,
"freq_dim": 256,
"text_dim": 4096,
"out_dim": 16,
"num_heads": 40,
"num_layers": 40,
"eps": 1e-6
}
elif hash_state_dict_keys(state_dict) == "6bfcfb3b342cb286ce886889d519a77e":
config = {
"has_image_input": True,
"patch_size": [1, 2, 2],
"in_dim": 36,
"dim": 5120,
"ffn_dim": 13824,
"freq_dim": 256,
"text_dim": 4096,
"out_dim": 16,
"num_heads": 40,
"num_layers": 40,
"eps": 1e-6
}
else:
config = {}
return state_dict, config
|