Spaces:
Running
on
Zero
Running
on
Zero
File size: 5,658 Bytes
4583098 |
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 |
import torch
import torch.nn as nn
import torch.nn.functional as F
from diffusers.models.normalization import RMSNorm
from typing import Optional
class FluxAttnProcessor(nn.Module):
def __init__(self, hidden_size, cross_attention_dim=None, scale=1.0, num_tokens=4):
super().__init__()
self.hidden_size = hidden_size
self.cross_attention_dim = cross_attention_dim
self.scale = scale
self.num_tokens = num_tokens
self.to_k_ip = nn.Linear(cross_attention_dim or hidden_size, hidden_size, bias=False)
self.to_v_ip = nn.Linear(cross_attention_dim or hidden_size, hidden_size, bias=False)
self.norm_added_k = RMSNorm(128, eps=1e-5, elementwise_affine=False)
def __call__(
self,
attn,
hidden_states: torch.FloatTensor,
image_emb: torch.FloatTensor,
encoder_hidden_states: torch.FloatTensor = None,
attention_mask: Optional[torch.FloatTensor] = None,
image_rotary_emb: Optional[torch.Tensor] = None,
) -> torch.FloatTensor:
batch_size, _, _ = hidden_states.shape if encoder_hidden_states is None else encoder_hidden_states.shape
query = attn.to_q(hidden_states)
key = attn.to_k(hidden_states)
value = attn.to_v(hidden_states)
inner_dim = key.shape[-1]
head_dim = inner_dim // attn.heads
query = query.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2)
key = key.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2)
value = value.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2)
if attn.norm_q is not None:
query = attn.norm_q(query)
if attn.norm_k is not None:
key = attn.norm_k(key)
if image_emb is not None:
ip_hidden_states = image_emb
ip_hidden_states_key_proj = self.to_k_ip(ip_hidden_states)
ip_hidden_states_value_proj = self.to_v_ip(ip_hidden_states)
ip_hidden_states_key_proj = ip_hidden_states_key_proj.view(
batch_size, -1, attn.heads, head_dim
).transpose(1, 2)
ip_hidden_states_value_proj = ip_hidden_states_value_proj.view(
batch_size, -1, attn.heads, head_dim
).transpose(1, 2)
ip_hidden_states_key_proj = self.norm_added_k(ip_hidden_states_key_proj)
ip_hidden_states = F.scaled_dot_product_attention(query,
ip_hidden_states_key_proj,
ip_hidden_states_value_proj,
dropout_p=0.0, is_causal=False)
ip_hidden_states = ip_hidden_states.transpose(1, 2).reshape(batch_size, -1, attn.heads * head_dim)
ip_hidden_states = ip_hidden_states.to(query.dtype)
if encoder_hidden_states is not None:
encoder_hidden_states_query_proj = attn.add_q_proj(encoder_hidden_states)
encoder_hidden_states_key_proj = attn.add_k_proj(encoder_hidden_states)
encoder_hidden_states_value_proj = attn.add_v_proj(encoder_hidden_states)
encoder_hidden_states_query_proj = encoder_hidden_states_query_proj.view(
batch_size, -1, attn.heads, head_dim
).transpose(1, 2)
encoder_hidden_states_key_proj = encoder_hidden_states_key_proj.view(
batch_size, -1, attn.heads, head_dim
).transpose(1, 2)
encoder_hidden_states_value_proj = encoder_hidden_states_value_proj.view(
batch_size, -1, attn.heads, head_dim
).transpose(1, 2)
if attn.norm_added_q is not None:
encoder_hidden_states_query_proj = attn.norm_added_q(encoder_hidden_states_query_proj)
if attn.norm_added_k is not None:
encoder_hidden_states_key_proj = attn.norm_added_k(encoder_hidden_states_key_proj)
query = torch.cat([encoder_hidden_states_query_proj, query], dim=2)
key = torch.cat([encoder_hidden_states_key_proj, key], dim=2)
value = torch.cat([encoder_hidden_states_value_proj, value], dim=2)
if image_rotary_emb is not None:
from diffusers.models.embeddings import apply_rotary_emb
query = apply_rotary_emb(query, image_rotary_emb)
key = apply_rotary_emb(key, image_rotary_emb)
hidden_states = F.scaled_dot_product_attention(query, key, value, dropout_p=0.0, is_causal=False)
hidden_states = hidden_states.transpose(1, 2).reshape(batch_size, -1, attn.heads * head_dim)
hidden_states = hidden_states.to(query.dtype)
if encoder_hidden_states is not None:
encoder_hidden_states, hidden_states = (
hidden_states[:, : encoder_hidden_states.shape[1]],
hidden_states[:, encoder_hidden_states.shape[1]:],
)
if image_emb is not None:
hidden_states = hidden_states + self.scale * ip_hidden_states
hidden_states = attn.to_out[0](hidden_states)
hidden_states = attn.to_out[1](hidden_states)
encoder_hidden_states = attn.to_add_out(encoder_hidden_states)
return hidden_states, encoder_hidden_states
else:
if image_emb is not None:
hidden_states = hidden_states + self.scale * ip_hidden_states
return hidden_states
|