Delete attention_processor.py
Browse files- attention_processor.py +0 -70
attention_processor.py
DELETED
@@ -1,70 +0,0 @@
|
|
1 |
-
from typing import Optional
|
2 |
-
|
3 |
-
import torch
|
4 |
-
import torch.nn as nn
|
5 |
-
from diffusers.models import UNet2DConditionModel
|
6 |
-
from diffusers.models.attention import Attention
|
7 |
-
from diffusers.models.attention_processor import AttnProcessor2_0
|
8 |
-
|
9 |
-
|
10 |
-
|
11 |
-
def add_imagedream_attn_processor(unet: UNet2DConditionModel) -> nn.Module:
|
12 |
-
attn_procs = {}
|
13 |
-
for key, attn_processor in unet.attn_processors.items():
|
14 |
-
if "attn1" in key:
|
15 |
-
attn_procs[key] = ImageDreamAttnProcessor2_0()
|
16 |
-
else:
|
17 |
-
attn_procs[key] = attn_processor
|
18 |
-
unet.set_attn_processor(attn_procs)
|
19 |
-
return unet
|
20 |
-
|
21 |
-
|
22 |
-
class ImageDreamAttnProcessor2_0(AttnProcessor2_0):
|
23 |
-
def __call__(
|
24 |
-
self,
|
25 |
-
attn: Attention,
|
26 |
-
hidden_states: torch.Tensor,
|
27 |
-
encoder_hidden_states: Optional[torch.Tensor] = None,
|
28 |
-
attention_mask: Optional[torch.Tensor] = None,
|
29 |
-
temb: Optional[torch.Tensor] = None,
|
30 |
-
num_views: int = 1,
|
31 |
-
*args,
|
32 |
-
**kwargs,
|
33 |
-
):
|
34 |
-
if num_views == 1:
|
35 |
-
return super().__call__(
|
36 |
-
attn=attn,
|
37 |
-
hidden_states=hidden_states,
|
38 |
-
encoder_hidden_states=encoder_hidden_states,
|
39 |
-
attention_mask=attention_mask,
|
40 |
-
temb=temb,
|
41 |
-
*args,
|
42 |
-
**kwargs,
|
43 |
-
)
|
44 |
-
|
45 |
-
input_ndim = hidden_states.ndim
|
46 |
-
B = hidden_states.size(0)
|
47 |
-
if B % num_views:
|
48 |
-
raise ValueError(
|
49 |
-
f"`batch_size`(got {B}) must be a multiple of `num_views`(got {num_views})."
|
50 |
-
)
|
51 |
-
real_B = B // num_views
|
52 |
-
if input_ndim == 4:
|
53 |
-
H, W = hidden_states.shape[2:]
|
54 |
-
hidden_states = hidden_states.reshape(real_B, -1, H, W).transpose(1, 2)
|
55 |
-
else:
|
56 |
-
hidden_states = hidden_states.reshape(real_B, -1, hidden_states.size(-1))
|
57 |
-
hidden_states = super().__call__(
|
58 |
-
attn=attn,
|
59 |
-
hidden_states=hidden_states,
|
60 |
-
encoder_hidden_states=encoder_hidden_states,
|
61 |
-
attention_mask=attention_mask,
|
62 |
-
temb=temb,
|
63 |
-
*args,
|
64 |
-
**kwargs,
|
65 |
-
)
|
66 |
-
if input_ndim == 4:
|
67 |
-
hidden_states = hidden_states.transpose(-1, -2).reshape(B, -1, H, W)
|
68 |
-
else:
|
69 |
-
hidden_states = hidden_states.reshape(B, -1, hidden_states.size(-1))
|
70 |
-
return hidden_states
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|