mirror of
https://github.com/comfyanonymous/ComfyUI.git
synced 2025-05-10 14:36:07 +00:00
798 lines
33 KiB
Python
798 lines
33 KiB
Python
from typing import Optional, Tuple, List
|
|
|
|
import torch
|
|
import torch.nn as nn
|
|
import einops
|
|
from einops import repeat
|
|
|
|
from comfy.ldm.lightricks.model import TimestepEmbedding, Timesteps
|
|
import torch.nn.functional as F
|
|
|
|
from comfy.ldm.flux.math import apply_rope, rope
|
|
from comfy.ldm.flux.layers import LastLayer
|
|
|
|
from comfy.ldm.modules.attention import optimized_attention
|
|
import comfy.model_management
|
|
|
|
|
|
# Copied from https://github.com/black-forest-labs/flux/blob/main/src/flux/modules/layers.py
|
|
class EmbedND(nn.Module):
|
|
def __init__(self, theta: int, axes_dim: List[int]):
|
|
super().__init__()
|
|
self.theta = theta
|
|
self.axes_dim = axes_dim
|
|
|
|
def forward(self, ids: torch.Tensor) -> torch.Tensor:
|
|
n_axes = ids.shape[-1]
|
|
emb = torch.cat(
|
|
[rope(ids[..., i], self.axes_dim[i], self.theta) for i in range(n_axes)],
|
|
dim=-3,
|
|
)
|
|
return emb.unsqueeze(2)
|
|
|
|
|
|
class PatchEmbed(nn.Module):
|
|
def __init__(
|
|
self,
|
|
patch_size=2,
|
|
in_channels=4,
|
|
out_channels=1024,
|
|
dtype=None, device=None, operations=None
|
|
):
|
|
super().__init__()
|
|
self.patch_size = patch_size
|
|
self.out_channels = out_channels
|
|
self.proj = operations.Linear(in_channels * patch_size * patch_size, out_channels, bias=True, dtype=dtype, device=device)
|
|
|
|
def forward(self, latent):
|
|
latent = self.proj(latent)
|
|
return latent
|
|
|
|
|
|
class PooledEmbed(nn.Module):
|
|
def __init__(self, text_emb_dim, hidden_size, dtype=None, device=None, operations=None):
|
|
super().__init__()
|
|
self.pooled_embedder = TimestepEmbedding(in_channels=text_emb_dim, time_embed_dim=hidden_size, dtype=dtype, device=device, operations=operations)
|
|
|
|
def forward(self, pooled_embed):
|
|
return self.pooled_embedder(pooled_embed)
|
|
|
|
|
|
class TimestepEmbed(nn.Module):
|
|
def __init__(self, hidden_size, frequency_embedding_size=256, dtype=None, device=None, operations=None):
|
|
super().__init__()
|
|
self.time_proj = Timesteps(num_channels=frequency_embedding_size, flip_sin_to_cos=True, downscale_freq_shift=0)
|
|
self.timestep_embedder = TimestepEmbedding(in_channels=frequency_embedding_size, time_embed_dim=hidden_size, dtype=dtype, device=device, operations=operations)
|
|
|
|
def forward(self, timesteps, wdtype):
|
|
t_emb = self.time_proj(timesteps).to(dtype=wdtype)
|
|
t_emb = self.timestep_embedder(t_emb)
|
|
return t_emb
|
|
|
|
|
|
def attention(query: torch.Tensor, key: torch.Tensor, value: torch.Tensor):
|
|
return optimized_attention(query.view(query.shape[0], -1, query.shape[-1] * query.shape[-2]), key.view(key.shape[0], -1, key.shape[-1] * key.shape[-2]), value.view(value.shape[0], -1, value.shape[-1] * value.shape[-2]), query.shape[2])
|
|
|
|
|
|
class HiDreamAttnProcessor_flashattn:
|
|
"""Attention processor used typically in processing the SD3-like self-attention projections."""
|
|
|
|
def __call__(
|
|
self,
|
|
attn,
|
|
image_tokens: torch.FloatTensor,
|
|
image_tokens_masks: Optional[torch.FloatTensor] = None,
|
|
text_tokens: Optional[torch.FloatTensor] = None,
|
|
rope: torch.FloatTensor = None,
|
|
*args,
|
|
**kwargs,
|
|
) -> torch.FloatTensor:
|
|
dtype = image_tokens.dtype
|
|
batch_size = image_tokens.shape[0]
|
|
|
|
query_i = attn.q_rms_norm(attn.to_q(image_tokens)).to(dtype=dtype)
|
|
key_i = attn.k_rms_norm(attn.to_k(image_tokens)).to(dtype=dtype)
|
|
value_i = attn.to_v(image_tokens)
|
|
|
|
inner_dim = key_i.shape[-1]
|
|
head_dim = inner_dim // attn.heads
|
|
|
|
query_i = query_i.view(batch_size, -1, attn.heads, head_dim)
|
|
key_i = key_i.view(batch_size, -1, attn.heads, head_dim)
|
|
value_i = value_i.view(batch_size, -1, attn.heads, head_dim)
|
|
if image_tokens_masks is not None:
|
|
key_i = key_i * image_tokens_masks.view(batch_size, -1, 1, 1)
|
|
|
|
if not attn.single:
|
|
query_t = attn.q_rms_norm_t(attn.to_q_t(text_tokens)).to(dtype=dtype)
|
|
key_t = attn.k_rms_norm_t(attn.to_k_t(text_tokens)).to(dtype=dtype)
|
|
value_t = attn.to_v_t(text_tokens)
|
|
|
|
query_t = query_t.view(batch_size, -1, attn.heads, head_dim)
|
|
key_t = key_t.view(batch_size, -1, attn.heads, head_dim)
|
|
value_t = value_t.view(batch_size, -1, attn.heads, head_dim)
|
|
|
|
num_image_tokens = query_i.shape[1]
|
|
num_text_tokens = query_t.shape[1]
|
|
query = torch.cat([query_i, query_t], dim=1)
|
|
key = torch.cat([key_i, key_t], dim=1)
|
|
value = torch.cat([value_i, value_t], dim=1)
|
|
else:
|
|
query = query_i
|
|
key = key_i
|
|
value = value_i
|
|
|
|
if query.shape[-1] == rope.shape[-3] * 2:
|
|
query, key = apply_rope(query, key, rope)
|
|
else:
|
|
query_1, query_2 = query.chunk(2, dim=-1)
|
|
key_1, key_2 = key.chunk(2, dim=-1)
|
|
query_1, key_1 = apply_rope(query_1, key_1, rope)
|
|
query = torch.cat([query_1, query_2], dim=-1)
|
|
key = torch.cat([key_1, key_2], dim=-1)
|
|
|
|
hidden_states = attention(query, key, value)
|
|
|
|
if not attn.single:
|
|
hidden_states_i, hidden_states_t = torch.split(hidden_states, [num_image_tokens, num_text_tokens], dim=1)
|
|
hidden_states_i = attn.to_out(hidden_states_i)
|
|
hidden_states_t = attn.to_out_t(hidden_states_t)
|
|
return hidden_states_i, hidden_states_t
|
|
else:
|
|
hidden_states = attn.to_out(hidden_states)
|
|
return hidden_states
|
|
|
|
class HiDreamAttention(nn.Module):
|
|
def __init__(
|
|
self,
|
|
query_dim: int,
|
|
heads: int = 8,
|
|
dim_head: int = 64,
|
|
upcast_attention: bool = False,
|
|
upcast_softmax: bool = False,
|
|
scale_qk: bool = True,
|
|
eps: float = 1e-5,
|
|
processor = None,
|
|
out_dim: int = None,
|
|
single: bool = False,
|
|
dtype=None, device=None, operations=None
|
|
):
|
|
# super(Attention, self).__init__()
|
|
super().__init__()
|
|
self.inner_dim = out_dim if out_dim is not None else dim_head * heads
|
|
self.query_dim = query_dim
|
|
self.upcast_attention = upcast_attention
|
|
self.upcast_softmax = upcast_softmax
|
|
self.out_dim = out_dim if out_dim is not None else query_dim
|
|
|
|
self.scale_qk = scale_qk
|
|
self.scale = dim_head**-0.5 if self.scale_qk else 1.0
|
|
|
|
self.heads = out_dim // dim_head if out_dim is not None else heads
|
|
self.sliceable_head_dim = heads
|
|
self.single = single
|
|
|
|
linear_cls = operations.Linear
|
|
self.linear_cls = linear_cls
|
|
self.to_q = linear_cls(query_dim, self.inner_dim, dtype=dtype, device=device)
|
|
self.to_k = linear_cls(self.inner_dim, self.inner_dim, dtype=dtype, device=device)
|
|
self.to_v = linear_cls(self.inner_dim, self.inner_dim, dtype=dtype, device=device)
|
|
self.to_out = linear_cls(self.inner_dim, self.out_dim, dtype=dtype, device=device)
|
|
self.q_rms_norm = operations.RMSNorm(self.inner_dim, eps, dtype=dtype, device=device)
|
|
self.k_rms_norm = operations.RMSNorm(self.inner_dim, eps, dtype=dtype, device=device)
|
|
|
|
if not single:
|
|
self.to_q_t = linear_cls(query_dim, self.inner_dim, dtype=dtype, device=device)
|
|
self.to_k_t = linear_cls(self.inner_dim, self.inner_dim, dtype=dtype, device=device)
|
|
self.to_v_t = linear_cls(self.inner_dim, self.inner_dim, dtype=dtype, device=device)
|
|
self.to_out_t = linear_cls(self.inner_dim, self.out_dim, dtype=dtype, device=device)
|
|
self.q_rms_norm_t = operations.RMSNorm(self.inner_dim, eps, dtype=dtype, device=device)
|
|
self.k_rms_norm_t = operations.RMSNorm(self.inner_dim, eps, dtype=dtype, device=device)
|
|
|
|
self.processor = processor
|
|
|
|
def forward(
|
|
self,
|
|
norm_image_tokens: torch.FloatTensor,
|
|
image_tokens_masks: torch.FloatTensor = None,
|
|
norm_text_tokens: torch.FloatTensor = None,
|
|
rope: torch.FloatTensor = None,
|
|
) -> torch.Tensor:
|
|
return self.processor(
|
|
self,
|
|
image_tokens = norm_image_tokens,
|
|
image_tokens_masks = image_tokens_masks,
|
|
text_tokens = norm_text_tokens,
|
|
rope = rope,
|
|
)
|
|
|
|
|
|
class FeedForwardSwiGLU(nn.Module):
|
|
def __init__(
|
|
self,
|
|
dim: int,
|
|
hidden_dim: int,
|
|
multiple_of: int = 256,
|
|
ffn_dim_multiplier: Optional[float] = None,
|
|
dtype=None, device=None, operations=None
|
|
):
|
|
super().__init__()
|
|
hidden_dim = int(2 * hidden_dim / 3)
|
|
# custom dim factor multiplier
|
|
if ffn_dim_multiplier is not None:
|
|
hidden_dim = int(ffn_dim_multiplier * hidden_dim)
|
|
hidden_dim = multiple_of * (
|
|
(hidden_dim + multiple_of - 1) // multiple_of
|
|
)
|
|
|
|
self.w1 = operations.Linear(dim, hidden_dim, bias=False, dtype=dtype, device=device)
|
|
self.w2 = operations.Linear(hidden_dim, dim, bias=False, dtype=dtype, device=device)
|
|
self.w3 = operations.Linear(dim, hidden_dim, bias=False, dtype=dtype, device=device)
|
|
|
|
def forward(self, x):
|
|
return self.w2(torch.nn.functional.silu(self.w1(x)) * self.w3(x))
|
|
|
|
|
|
# Modified from https://github.com/deepseek-ai/DeepSeek-V3/blob/main/inference/model.py
|
|
class MoEGate(nn.Module):
|
|
def __init__(self, embed_dim, num_routed_experts=4, num_activated_experts=2, aux_loss_alpha=0.01, dtype=None, device=None, operations=None):
|
|
super().__init__()
|
|
self.top_k = num_activated_experts
|
|
self.n_routed_experts = num_routed_experts
|
|
|
|
self.scoring_func = 'softmax'
|
|
self.alpha = aux_loss_alpha
|
|
self.seq_aux = False
|
|
|
|
# topk selection algorithm
|
|
self.norm_topk_prob = False
|
|
self.gating_dim = embed_dim
|
|
self.weight = nn.Parameter(torch.empty((self.n_routed_experts, self.gating_dim), dtype=dtype, device=device))
|
|
self.reset_parameters()
|
|
|
|
def reset_parameters(self) -> None:
|
|
pass
|
|
# import torch.nn.init as init
|
|
# init.kaiming_uniform_(self.weight, a=math.sqrt(5))
|
|
|
|
def forward(self, hidden_states):
|
|
bsz, seq_len, h = hidden_states.shape
|
|
|
|
### compute gating score
|
|
hidden_states = hidden_states.view(-1, h)
|
|
logits = F.linear(hidden_states, comfy.model_management.cast_to(self.weight, dtype=hidden_states.dtype, device=hidden_states.device), None)
|
|
if self.scoring_func == 'softmax':
|
|
scores = logits.softmax(dim=-1)
|
|
else:
|
|
raise NotImplementedError(f'insupportable scoring function for MoE gating: {self.scoring_func}')
|
|
|
|
### select top-k experts
|
|
topk_weight, topk_idx = torch.topk(scores, k=self.top_k, dim=-1, sorted=False)
|
|
|
|
### norm gate to sum 1
|
|
if self.top_k > 1 and self.norm_topk_prob:
|
|
denominator = topk_weight.sum(dim=-1, keepdim=True) + 1e-20
|
|
topk_weight = topk_weight / denominator
|
|
|
|
aux_loss = None
|
|
return topk_idx, topk_weight, aux_loss
|
|
|
|
|
|
# Modified from https://github.com/deepseek-ai/DeepSeek-V3/blob/main/inference/model.py
|
|
class MOEFeedForwardSwiGLU(nn.Module):
|
|
def __init__(
|
|
self,
|
|
dim: int,
|
|
hidden_dim: int,
|
|
num_routed_experts: int,
|
|
num_activated_experts: int,
|
|
dtype=None, device=None, operations=None
|
|
):
|
|
super().__init__()
|
|
self.shared_experts = FeedForwardSwiGLU(dim, hidden_dim // 2, dtype=dtype, device=device, operations=operations)
|
|
self.experts = nn.ModuleList([FeedForwardSwiGLU(dim, hidden_dim, dtype=dtype, device=device, operations=operations) for i in range(num_routed_experts)])
|
|
self.gate = MoEGate(
|
|
embed_dim = dim,
|
|
num_routed_experts = num_routed_experts,
|
|
num_activated_experts = num_activated_experts,
|
|
dtype=dtype, device=device, operations=operations
|
|
)
|
|
self.num_activated_experts = num_activated_experts
|
|
|
|
def forward(self, x):
|
|
wtype = x.dtype
|
|
identity = x
|
|
orig_shape = x.shape
|
|
topk_idx, topk_weight, aux_loss = self.gate(x)
|
|
x = x.view(-1, x.shape[-1])
|
|
flat_topk_idx = topk_idx.view(-1)
|
|
if True: # self.training: # TODO: check which branch performs faster
|
|
x = x.repeat_interleave(self.num_activated_experts, dim=0)
|
|
y = torch.empty_like(x, dtype=wtype)
|
|
for i, expert in enumerate(self.experts):
|
|
y[flat_topk_idx == i] = expert(x[flat_topk_idx == i]).to(dtype=wtype)
|
|
y = (y.view(*topk_weight.shape, -1) * topk_weight.unsqueeze(-1)).sum(dim=1)
|
|
y = y.view(*orig_shape).to(dtype=wtype)
|
|
#y = AddAuxiliaryLoss.apply(y, aux_loss)
|
|
else:
|
|
y = self.moe_infer(x, flat_topk_idx, topk_weight.view(-1, 1)).view(*orig_shape)
|
|
y = y + self.shared_experts(identity)
|
|
return y
|
|
|
|
@torch.no_grad()
|
|
def moe_infer(self, x, flat_expert_indices, flat_expert_weights):
|
|
expert_cache = torch.zeros_like(x)
|
|
idxs = flat_expert_indices.argsort()
|
|
tokens_per_expert = flat_expert_indices.bincount().cpu().numpy().cumsum(0)
|
|
token_idxs = idxs // self.num_activated_experts
|
|
for i, end_idx in enumerate(tokens_per_expert):
|
|
start_idx = 0 if i == 0 else tokens_per_expert[i-1]
|
|
if start_idx == end_idx:
|
|
continue
|
|
expert = self.experts[i]
|
|
exp_token_idx = token_idxs[start_idx:end_idx]
|
|
expert_tokens = x[exp_token_idx]
|
|
expert_out = expert(expert_tokens)
|
|
expert_out.mul_(flat_expert_weights[idxs[start_idx:end_idx]])
|
|
|
|
# for fp16 and other dtype
|
|
expert_cache = expert_cache.to(expert_out.dtype)
|
|
expert_cache.scatter_reduce_(0, exp_token_idx.view(-1, 1).repeat(1, x.shape[-1]), expert_out, reduce='sum')
|
|
return expert_cache
|
|
|
|
|
|
class TextProjection(nn.Module):
|
|
def __init__(self, in_features, hidden_size, dtype=None, device=None, operations=None):
|
|
super().__init__()
|
|
self.linear = operations.Linear(in_features=in_features, out_features=hidden_size, bias=False, dtype=dtype, device=device)
|
|
|
|
def forward(self, caption):
|
|
hidden_states = self.linear(caption)
|
|
return hidden_states
|
|
|
|
|
|
class BlockType:
|
|
TransformerBlock = 1
|
|
SingleTransformerBlock = 2
|
|
|
|
|
|
class HiDreamImageSingleTransformerBlock(nn.Module):
|
|
def __init__(
|
|
self,
|
|
dim: int,
|
|
num_attention_heads: int,
|
|
attention_head_dim: int,
|
|
num_routed_experts: int = 4,
|
|
num_activated_experts: int = 2,
|
|
dtype=None, device=None, operations=None
|
|
):
|
|
super().__init__()
|
|
self.num_attention_heads = num_attention_heads
|
|
self.adaLN_modulation = nn.Sequential(
|
|
nn.SiLU(),
|
|
operations.Linear(dim, 6 * dim, bias=True, dtype=dtype, device=device)
|
|
)
|
|
|
|
# 1. Attention
|
|
self.norm1_i = operations.LayerNorm(dim, eps = 1e-06, elementwise_affine = False, dtype=dtype, device=device)
|
|
self.attn1 = HiDreamAttention(
|
|
query_dim=dim,
|
|
heads=num_attention_heads,
|
|
dim_head=attention_head_dim,
|
|
processor = HiDreamAttnProcessor_flashattn(),
|
|
single = True,
|
|
dtype=dtype, device=device, operations=operations
|
|
)
|
|
|
|
# 3. Feed-forward
|
|
self.norm3_i = operations.LayerNorm(dim, eps = 1e-06, elementwise_affine = False, dtype=dtype, device=device)
|
|
if num_routed_experts > 0:
|
|
self.ff_i = MOEFeedForwardSwiGLU(
|
|
dim = dim,
|
|
hidden_dim = 4 * dim,
|
|
num_routed_experts = num_routed_experts,
|
|
num_activated_experts = num_activated_experts,
|
|
dtype=dtype, device=device, operations=operations
|
|
)
|
|
else:
|
|
self.ff_i = FeedForwardSwiGLU(dim = dim, hidden_dim = 4 * dim, dtype=dtype, device=device, operations=operations)
|
|
|
|
def forward(
|
|
self,
|
|
image_tokens: torch.FloatTensor,
|
|
image_tokens_masks: Optional[torch.FloatTensor] = None,
|
|
text_tokens: Optional[torch.FloatTensor] = None,
|
|
adaln_input: Optional[torch.FloatTensor] = None,
|
|
rope: torch.FloatTensor = None,
|
|
|
|
) -> torch.FloatTensor:
|
|
wtype = image_tokens.dtype
|
|
shift_msa_i, scale_msa_i, gate_msa_i, shift_mlp_i, scale_mlp_i, gate_mlp_i = \
|
|
self.adaLN_modulation(adaln_input)[:,None].chunk(6, dim=-1)
|
|
|
|
# 1. MM-Attention
|
|
norm_image_tokens = self.norm1_i(image_tokens).to(dtype=wtype)
|
|
norm_image_tokens = norm_image_tokens * (1 + scale_msa_i) + shift_msa_i
|
|
attn_output_i = self.attn1(
|
|
norm_image_tokens,
|
|
image_tokens_masks,
|
|
rope = rope,
|
|
)
|
|
image_tokens = gate_msa_i * attn_output_i + image_tokens
|
|
|
|
# 2. Feed-forward
|
|
norm_image_tokens = self.norm3_i(image_tokens).to(dtype=wtype)
|
|
norm_image_tokens = norm_image_tokens * (1 + scale_mlp_i) + shift_mlp_i
|
|
ff_output_i = gate_mlp_i * self.ff_i(norm_image_tokens.to(dtype=wtype))
|
|
image_tokens = ff_output_i + image_tokens
|
|
return image_tokens
|
|
|
|
|
|
class HiDreamImageTransformerBlock(nn.Module):
|
|
def __init__(
|
|
self,
|
|
dim: int,
|
|
num_attention_heads: int,
|
|
attention_head_dim: int,
|
|
num_routed_experts: int = 4,
|
|
num_activated_experts: int = 2,
|
|
dtype=None, device=None, operations=None
|
|
):
|
|
super().__init__()
|
|
self.num_attention_heads = num_attention_heads
|
|
self.adaLN_modulation = nn.Sequential(
|
|
nn.SiLU(),
|
|
operations.Linear(dim, 12 * dim, bias=True, dtype=dtype, device=device)
|
|
)
|
|
# nn.init.zeros_(self.adaLN_modulation[1].weight)
|
|
# nn.init.zeros_(self.adaLN_modulation[1].bias)
|
|
|
|
# 1. Attention
|
|
self.norm1_i = operations.LayerNorm(dim, eps = 1e-06, elementwise_affine = False, dtype=dtype, device=device)
|
|
self.norm1_t = operations.LayerNorm(dim, eps = 1e-06, elementwise_affine = False, dtype=dtype, device=device)
|
|
self.attn1 = HiDreamAttention(
|
|
query_dim=dim,
|
|
heads=num_attention_heads,
|
|
dim_head=attention_head_dim,
|
|
processor = HiDreamAttnProcessor_flashattn(),
|
|
single = False,
|
|
dtype=dtype, device=device, operations=operations
|
|
)
|
|
|
|
# 3. Feed-forward
|
|
self.norm3_i = operations.LayerNorm(dim, eps = 1e-06, elementwise_affine = False, dtype=dtype, device=device)
|
|
if num_routed_experts > 0:
|
|
self.ff_i = MOEFeedForwardSwiGLU(
|
|
dim = dim,
|
|
hidden_dim = 4 * dim,
|
|
num_routed_experts = num_routed_experts,
|
|
num_activated_experts = num_activated_experts,
|
|
dtype=dtype, device=device, operations=operations
|
|
)
|
|
else:
|
|
self.ff_i = FeedForwardSwiGLU(dim = dim, hidden_dim = 4 * dim, dtype=dtype, device=device, operations=operations)
|
|
self.norm3_t = operations.LayerNorm(dim, eps = 1e-06, elementwise_affine = False)
|
|
self.ff_t = FeedForwardSwiGLU(dim = dim, hidden_dim = 4 * dim, dtype=dtype, device=device, operations=operations)
|
|
|
|
def forward(
|
|
self,
|
|
image_tokens: torch.FloatTensor,
|
|
image_tokens_masks: Optional[torch.FloatTensor] = None,
|
|
text_tokens: Optional[torch.FloatTensor] = None,
|
|
adaln_input: Optional[torch.FloatTensor] = None,
|
|
rope: torch.FloatTensor = None,
|
|
) -> torch.FloatTensor:
|
|
wtype = image_tokens.dtype
|
|
shift_msa_i, scale_msa_i, gate_msa_i, shift_mlp_i, scale_mlp_i, gate_mlp_i, \
|
|
shift_msa_t, scale_msa_t, gate_msa_t, shift_mlp_t, scale_mlp_t, gate_mlp_t = \
|
|
self.adaLN_modulation(adaln_input)[:,None].chunk(12, dim=-1)
|
|
|
|
# 1. MM-Attention
|
|
norm_image_tokens = self.norm1_i(image_tokens).to(dtype=wtype)
|
|
norm_image_tokens = norm_image_tokens * (1 + scale_msa_i) + shift_msa_i
|
|
norm_text_tokens = self.norm1_t(text_tokens).to(dtype=wtype)
|
|
norm_text_tokens = norm_text_tokens * (1 + scale_msa_t) + shift_msa_t
|
|
|
|
attn_output_i, attn_output_t = self.attn1(
|
|
norm_image_tokens,
|
|
image_tokens_masks,
|
|
norm_text_tokens,
|
|
rope = rope,
|
|
)
|
|
|
|
image_tokens = gate_msa_i * attn_output_i + image_tokens
|
|
text_tokens = gate_msa_t * attn_output_t + text_tokens
|
|
|
|
# 2. Feed-forward
|
|
norm_image_tokens = self.norm3_i(image_tokens).to(dtype=wtype)
|
|
norm_image_tokens = norm_image_tokens * (1 + scale_mlp_i) + shift_mlp_i
|
|
norm_text_tokens = self.norm3_t(text_tokens).to(dtype=wtype)
|
|
norm_text_tokens = norm_text_tokens * (1 + scale_mlp_t) + shift_mlp_t
|
|
|
|
ff_output_i = gate_mlp_i * self.ff_i(norm_image_tokens)
|
|
ff_output_t = gate_mlp_t * self.ff_t(norm_text_tokens)
|
|
image_tokens = ff_output_i + image_tokens
|
|
text_tokens = ff_output_t + text_tokens
|
|
return image_tokens, text_tokens
|
|
|
|
|
|
class HiDreamImageBlock(nn.Module):
|
|
def __init__(
|
|
self,
|
|
dim: int,
|
|
num_attention_heads: int,
|
|
attention_head_dim: int,
|
|
num_routed_experts: int = 4,
|
|
num_activated_experts: int = 2,
|
|
block_type: BlockType = BlockType.TransformerBlock,
|
|
dtype=None, device=None, operations=None
|
|
):
|
|
super().__init__()
|
|
block_classes = {
|
|
BlockType.TransformerBlock: HiDreamImageTransformerBlock,
|
|
BlockType.SingleTransformerBlock: HiDreamImageSingleTransformerBlock,
|
|
}
|
|
self.block = block_classes[block_type](
|
|
dim,
|
|
num_attention_heads,
|
|
attention_head_dim,
|
|
num_routed_experts,
|
|
num_activated_experts,
|
|
dtype=dtype, device=device, operations=operations
|
|
)
|
|
|
|
def forward(
|
|
self,
|
|
image_tokens: torch.FloatTensor,
|
|
image_tokens_masks: Optional[torch.FloatTensor] = None,
|
|
text_tokens: Optional[torch.FloatTensor] = None,
|
|
adaln_input: torch.FloatTensor = None,
|
|
rope: torch.FloatTensor = None,
|
|
) -> torch.FloatTensor:
|
|
return self.block(
|
|
image_tokens,
|
|
image_tokens_masks,
|
|
text_tokens,
|
|
adaln_input,
|
|
rope,
|
|
)
|
|
|
|
|
|
class HiDreamImageTransformer2DModel(nn.Module):
|
|
def __init__(
|
|
self,
|
|
patch_size: Optional[int] = None,
|
|
in_channels: int = 64,
|
|
out_channels: Optional[int] = None,
|
|
num_layers: int = 16,
|
|
num_single_layers: int = 32,
|
|
attention_head_dim: int = 128,
|
|
num_attention_heads: int = 20,
|
|
caption_channels: List[int] = None,
|
|
text_emb_dim: int = 2048,
|
|
num_routed_experts: int = 4,
|
|
num_activated_experts: int = 2,
|
|
axes_dims_rope: Tuple[int, int] = (32, 32),
|
|
max_resolution: Tuple[int, int] = (128, 128),
|
|
llama_layers: List[int] = None,
|
|
image_model=None,
|
|
dtype=None, device=None, operations=None
|
|
):
|
|
self.patch_size = patch_size
|
|
self.num_attention_heads = num_attention_heads
|
|
self.attention_head_dim = attention_head_dim
|
|
self.num_layers = num_layers
|
|
self.num_single_layers = num_single_layers
|
|
|
|
self.gradient_checkpointing = False
|
|
|
|
super().__init__()
|
|
self.dtype = dtype
|
|
self.out_channels = out_channels or in_channels
|
|
self.inner_dim = self.num_attention_heads * self.attention_head_dim
|
|
self.llama_layers = llama_layers
|
|
|
|
self.t_embedder = TimestepEmbed(self.inner_dim, dtype=dtype, device=device, operations=operations)
|
|
self.p_embedder = PooledEmbed(text_emb_dim, self.inner_dim, dtype=dtype, device=device, operations=operations)
|
|
self.x_embedder = PatchEmbed(
|
|
patch_size = patch_size,
|
|
in_channels = in_channels,
|
|
out_channels = self.inner_dim,
|
|
dtype=dtype, device=device, operations=operations
|
|
)
|
|
self.pe_embedder = EmbedND(theta=10000, axes_dim=axes_dims_rope)
|
|
|
|
self.double_stream_blocks = nn.ModuleList(
|
|
[
|
|
HiDreamImageBlock(
|
|
dim = self.inner_dim,
|
|
num_attention_heads = self.num_attention_heads,
|
|
attention_head_dim = self.attention_head_dim,
|
|
num_routed_experts = num_routed_experts,
|
|
num_activated_experts = num_activated_experts,
|
|
block_type = BlockType.TransformerBlock,
|
|
dtype=dtype, device=device, operations=operations
|
|
)
|
|
for i in range(self.num_layers)
|
|
]
|
|
)
|
|
|
|
self.single_stream_blocks = nn.ModuleList(
|
|
[
|
|
HiDreamImageBlock(
|
|
dim = self.inner_dim,
|
|
num_attention_heads = self.num_attention_heads,
|
|
attention_head_dim = self.attention_head_dim,
|
|
num_routed_experts = num_routed_experts,
|
|
num_activated_experts = num_activated_experts,
|
|
block_type = BlockType.SingleTransformerBlock,
|
|
dtype=dtype, device=device, operations=operations
|
|
)
|
|
for i in range(self.num_single_layers)
|
|
]
|
|
)
|
|
|
|
self.final_layer = LastLayer(self.inner_dim, patch_size, self.out_channels, dtype=dtype, device=device, operations=operations)
|
|
|
|
caption_channels = [caption_channels[1], ] * (num_layers + num_single_layers) + [caption_channels[0], ]
|
|
caption_projection = []
|
|
for caption_channel in caption_channels:
|
|
caption_projection.append(TextProjection(in_features=caption_channel, hidden_size=self.inner_dim, dtype=dtype, device=device, operations=operations))
|
|
self.caption_projection = nn.ModuleList(caption_projection)
|
|
self.max_seq = max_resolution[0] * max_resolution[1] // (patch_size * patch_size)
|
|
|
|
def expand_timesteps(self, timesteps, batch_size, device):
|
|
if not torch.is_tensor(timesteps):
|
|
is_mps = device.type == "mps"
|
|
if isinstance(timesteps, float):
|
|
dtype = torch.float32 if is_mps else torch.float64
|
|
else:
|
|
dtype = torch.int32 if is_mps else torch.int64
|
|
timesteps = torch.tensor([timesteps], dtype=dtype, device=device)
|
|
elif len(timesteps.shape) == 0:
|
|
timesteps = timesteps[None].to(device)
|
|
# broadcast to batch dimension in a way that's compatible with ONNX/Core ML
|
|
timesteps = timesteps.expand(batch_size)
|
|
return timesteps
|
|
|
|
def unpatchify(self, x: torch.Tensor, img_sizes: List[Tuple[int, int]]) -> List[torch.Tensor]:
|
|
x_arr = []
|
|
for i, img_size in enumerate(img_sizes):
|
|
pH, pW = img_size
|
|
x_arr.append(
|
|
einops.rearrange(x[i, :pH*pW].reshape(1, pH, pW, -1), 'B H W (p1 p2 C) -> B C (H p1) (W p2)',
|
|
p1=self.patch_size, p2=self.patch_size)
|
|
)
|
|
x = torch.cat(x_arr, dim=0)
|
|
return x
|
|
|
|
def patchify(self, x, max_seq, img_sizes=None):
|
|
pz2 = self.patch_size * self.patch_size
|
|
if isinstance(x, torch.Tensor):
|
|
B = x.shape[0]
|
|
device = x.device
|
|
dtype = x.dtype
|
|
else:
|
|
B = len(x)
|
|
device = x[0].device
|
|
dtype = x[0].dtype
|
|
x_masks = torch.zeros((B, max_seq), dtype=dtype, device=device)
|
|
|
|
if img_sizes is not None:
|
|
for i, img_size in enumerate(img_sizes):
|
|
x_masks[i, 0:img_size[0] * img_size[1]] = 1
|
|
x = einops.rearrange(x, 'B C S p -> B S (p C)', p=pz2)
|
|
elif isinstance(x, torch.Tensor):
|
|
pH, pW = x.shape[-2] // self.patch_size, x.shape[-1] // self.patch_size
|
|
x = einops.rearrange(x, 'B C (H p1) (W p2) -> B (H W) (p1 p2 C)', p1=self.patch_size, p2=self.patch_size)
|
|
img_sizes = [[pH, pW]] * B
|
|
x_masks = None
|
|
else:
|
|
raise NotImplementedError
|
|
return x, x_masks, img_sizes
|
|
|
|
def forward(
|
|
self,
|
|
x: torch.Tensor,
|
|
t: torch.Tensor,
|
|
y: Optional[torch.Tensor] = None,
|
|
context: Optional[torch.Tensor] = None,
|
|
encoder_hidden_states_llama3=None,
|
|
control = None,
|
|
transformer_options = {},
|
|
) -> torch.Tensor:
|
|
hidden_states = x
|
|
timesteps = t
|
|
pooled_embeds = y
|
|
T5_encoder_hidden_states = context
|
|
|
|
img_sizes = None
|
|
|
|
# spatial forward
|
|
batch_size = hidden_states.shape[0]
|
|
hidden_states_type = hidden_states.dtype
|
|
|
|
# 0. time
|
|
timesteps = self.expand_timesteps(timesteps, batch_size, hidden_states.device)
|
|
timesteps = self.t_embedder(timesteps, hidden_states_type)
|
|
p_embedder = self.p_embedder(pooled_embeds)
|
|
adaln_input = timesteps + p_embedder
|
|
|
|
hidden_states, image_tokens_masks, img_sizes = self.patchify(hidden_states, self.max_seq, img_sizes)
|
|
if image_tokens_masks is None:
|
|
pH, pW = img_sizes[0]
|
|
img_ids = torch.zeros(pH, pW, 3, device=hidden_states.device)
|
|
img_ids[..., 1] = img_ids[..., 1] + torch.arange(pH, device=hidden_states.device)[:, None]
|
|
img_ids[..., 2] = img_ids[..., 2] + torch.arange(pW, device=hidden_states.device)[None, :]
|
|
img_ids = repeat(img_ids, "h w c -> b (h w) c", b=batch_size)
|
|
hidden_states = self.x_embedder(hidden_states)
|
|
|
|
# T5_encoder_hidden_states = encoder_hidden_states[0]
|
|
encoder_hidden_states = encoder_hidden_states_llama3.movedim(1, 0)
|
|
encoder_hidden_states = [encoder_hidden_states[k] for k in self.llama_layers]
|
|
|
|
if self.caption_projection is not None:
|
|
new_encoder_hidden_states = []
|
|
for i, enc_hidden_state in enumerate(encoder_hidden_states):
|
|
enc_hidden_state = self.caption_projection[i](enc_hidden_state)
|
|
enc_hidden_state = enc_hidden_state.view(batch_size, -1, hidden_states.shape[-1])
|
|
new_encoder_hidden_states.append(enc_hidden_state)
|
|
encoder_hidden_states = new_encoder_hidden_states
|
|
T5_encoder_hidden_states = self.caption_projection[-1](T5_encoder_hidden_states)
|
|
T5_encoder_hidden_states = T5_encoder_hidden_states.view(batch_size, -1, hidden_states.shape[-1])
|
|
encoder_hidden_states.append(T5_encoder_hidden_states)
|
|
|
|
txt_ids = torch.zeros(
|
|
batch_size,
|
|
encoder_hidden_states[-1].shape[1] + encoder_hidden_states[-2].shape[1] + encoder_hidden_states[0].shape[1],
|
|
3,
|
|
device=img_ids.device, dtype=img_ids.dtype
|
|
)
|
|
ids = torch.cat((img_ids, txt_ids), dim=1)
|
|
rope = self.pe_embedder(ids)
|
|
|
|
# 2. Blocks
|
|
block_id = 0
|
|
initial_encoder_hidden_states = torch.cat([encoder_hidden_states[-1], encoder_hidden_states[-2]], dim=1)
|
|
initial_encoder_hidden_states_seq_len = initial_encoder_hidden_states.shape[1]
|
|
for bid, block in enumerate(self.double_stream_blocks):
|
|
cur_llama31_encoder_hidden_states = encoder_hidden_states[block_id]
|
|
cur_encoder_hidden_states = torch.cat([initial_encoder_hidden_states, cur_llama31_encoder_hidden_states], dim=1)
|
|
hidden_states, initial_encoder_hidden_states = block(
|
|
image_tokens = hidden_states,
|
|
image_tokens_masks = image_tokens_masks,
|
|
text_tokens = cur_encoder_hidden_states,
|
|
adaln_input = adaln_input,
|
|
rope = rope,
|
|
)
|
|
initial_encoder_hidden_states = initial_encoder_hidden_states[:, :initial_encoder_hidden_states_seq_len]
|
|
block_id += 1
|
|
|
|
image_tokens_seq_len = hidden_states.shape[1]
|
|
hidden_states = torch.cat([hidden_states, initial_encoder_hidden_states], dim=1)
|
|
hidden_states_seq_len = hidden_states.shape[1]
|
|
if image_tokens_masks is not None:
|
|
encoder_attention_mask_ones = torch.ones(
|
|
(batch_size, initial_encoder_hidden_states.shape[1] + cur_llama31_encoder_hidden_states.shape[1]),
|
|
device=image_tokens_masks.device, dtype=image_tokens_masks.dtype
|
|
)
|
|
image_tokens_masks = torch.cat([image_tokens_masks, encoder_attention_mask_ones], dim=1)
|
|
|
|
for bid, block in enumerate(self.single_stream_blocks):
|
|
cur_llama31_encoder_hidden_states = encoder_hidden_states[block_id]
|
|
hidden_states = torch.cat([hidden_states, cur_llama31_encoder_hidden_states], dim=1)
|
|
hidden_states = block(
|
|
image_tokens=hidden_states,
|
|
image_tokens_masks=image_tokens_masks,
|
|
text_tokens=None,
|
|
adaln_input=adaln_input,
|
|
rope=rope,
|
|
)
|
|
hidden_states = hidden_states[:, :hidden_states_seq_len]
|
|
block_id += 1
|
|
|
|
hidden_states = hidden_states[:, :image_tokens_seq_len, ...]
|
|
output = self.final_layer(hidden_states, adaln_input)
|
|
output = self.unpatchify(output, img_sizes)
|
|
return -output
|