Skip to content

Looking Ahead: Modeling Future Evolution in Autoregressive Video Generation

2.8 KB
Raw Blame
# Modified for Looking-Ahead: selected TF/TI2V components; unrelated workflows removed.
# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
import torch
import torch.nn as nn
from .attention import attention


class WanRMSNorm(nn.Module):

    def __init__(self, dim, eps=1e-5):
        super().__init__()
        self.dim = dim
        self.eps = eps
        self.weight = nn.Parameter(torch.ones(dim))

    def forward(self, x):
        r"""
        Args:
            x(Tensor): Shape [B, L, C]
        """
        return self._norm(x.float()).type_as(x) * self.weight

    def _norm(self, x):
        return x * torch.rsqrt(x.pow(2).mean(dim=-1, keepdim=True) + self.eps)

    def reset_parameters(self):
        nn.init.ones_(self.weight)


class WanSelfAttention(nn.Module):

    def __init__(self,
                 dim,
                 num_heads,
                 window_size=(-1, -1),
                 qk_norm=True,
                 eps=1e-6):
        assert dim % num_heads == 0
        super().__init__()
        self.dim = dim
        self.num_heads = num_heads
        self.head_dim = dim // num_heads
        self.window_size = window_size
        self.qk_norm = qk_norm
        self.eps = eps

        # layers
        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 = WanRMSNorm(dim, eps=eps) if qk_norm else nn.Identity()
        self.norm_k = WanRMSNorm(dim, eps=eps) if qk_norm else nn.Identity()



class WanT2VCrossAttention(WanSelfAttention):

    def forward(self, x, context, context_lens, crossattn_cache=None):
        r"""
        Args:
            x(Tensor): Shape [B, L1, C]
            context(Tensor): Shape [B, L2, C]
            context_lens(Tensor): Shape [B]
            crossattn_cache (List[dict], *optional*): Contains the cached key and value tensors for context embedding.
        """
        b, n, d = x.size(0), self.num_heads, self.head_dim

        # compute query, key, value
        q = self.norm_q(self.q(x)).view(b, -1, n, d)

        if crossattn_cache is not None:
            if not crossattn_cache["is_init"]:
                crossattn_cache["is_init"] = True
                k = self.norm_k(self.k(context)).view(b, -1, n, d)
                v = self.v(context).view(b, -1, n, d)
                crossattn_cache["k"] = k
                crossattn_cache["v"] = v
            else:
                k = crossattn_cache["k"]
                v = crossattn_cache["v"]
        else:
            k = self.norm_k(self.k(context)).view(b, -1, n, d)
            v = self.v(context).view(b, -1, n, d)

        # compute attention
        x = attention(q, k, v, k_lens=context_lens)

        # output
        x = x.flatten(2)
        x = self.o(x)
        return x

WAN_CROSSATTENTION_CLASSES = {"t2v_cross_attn": WanT2VCrossAttention}