File size: 6,929 Bytes
6ec18a1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import torch
from torch import nn, einsum
import torch.nn.functional as F

from einops import rearrange, repeat
from einops_exts import rearrange_many, repeat_many
import pdb


def exists(val):
    return val is not None

def FeedForward(dim, mult = 4):
    inner_dim = int(dim * mult)
    return nn.Sequential(
        nn.LayerNorm(dim),
        nn.Linear(dim, inner_dim, bias = False),
        nn.GELU(),
        nn.Linear(inner_dim, dim, bias = False)
    )

class PerceiverAttention(nn.Module):
    def __init__(
        self,
        *,
        dim,
        dim_head = 64,
        heads = 8
    ):
        super().__init__()
        self.scale = dim_head ** -0.5
        self.heads = heads
        inner_dim = dim_head * heads

        self.norm_media = nn.LayerNorm(dim)
        self.norm_latents = nn.LayerNorm(dim)

        self.to_q = nn.Linear(dim, inner_dim, bias = False)
        self.to_kv = nn.Linear(dim, inner_dim * 2, bias = False)
        self.to_out = nn.Linear(inner_dim, dim, bias = False)

    def forward(self, x, latents):
        """
        einstein notation
        b - batch
        t - time
        n - sequence
        d - dimension
        """
        x = self.norm_media(x)
        latents = self.norm_latents(latents)

        b, m, h = *x.shape[:2], self.heads

        q = self.to_q(latents)

        # the paper differs from Perceiver in which they also concat the key / values derived from the latents to be attended to
        kv_input = torch.cat((x, latents), dim = -2)
        k, v = self.to_kv(kv_input).chunk(2, dim = -1)

        q, k, v = rearrange_many((q, k, v), 'b t n (h d) -> b h t n d', h = h)

        q = q * self.scale

        # attention

        sim = einsum('... i d, ... j d  -> ... i j', q, k)

        sim = sim - sim.amax(dim = -1, keepdim = True).detach()
        attn = sim.softmax(dim = -1)

        out = einsum('... i j, ... j d -> ... i d', attn, v)
        out = rearrange(out, 'b h t n d -> b t n (h d)', h = h)
        return self.to_out(out)

class PerceiverResampler(nn.Module):
    def __init__(
        self,
        *,
        dim,
        depth,
        dim_head = 64,
        heads = 8,
        num_latents = 64,
        num_time_embeds = 4,
        ff_mult = 4,
        inp_dim=None,
    ):
        super().__init__()
        self.latents = nn.Parameter(torch.randn(num_latents, dim))
        self.time_pos_emb = nn.Parameter(torch.randn(num_time_embeds, 1, dim))
        if inp_dim is not None:
            self.inp_linear = nn.Linear(inp_dim, dim, bias=False)
        else:
            self.inp_linear = None

        self.layers = nn.ModuleList([])
        for _ in range(depth):
            self.layers.append(nn.ModuleList([
                PerceiverAttention(dim = dim, dim_head = dim_head, heads = heads),
                FeedForward(dim = dim, mult = ff_mult)
            ]))

        self.norm = nn.LayerNorm(dim)

    def forward(self, x):
        if x.ndim == 3:
            x = rearrange(x, 'b n d -> b 1 n d')

        if self.inp_linear is not None:
            x = self.inp_linear(x)

        times = x.shape[1]
        x = x + self.time_pos_emb[:times]

        latents = repeat(self.latents, 'n d -> b m n d', b = x.shape[0], m = x.shape[1])

        for attn, ff in self.layers:
            latents = attn(x, latents) + latents
            latents = ff(latents) + latents

        return self.norm(latents)

# gated cross attention

class MaskedCrossAttention(nn.Module):
    def __init__(
        self,
        *,
        dim,
        dim_head = 64,
        heads = 8,
        only_attend_immediate_media = True
    ):
        super().__init__()
        self.scale = dim_head ** -0.5
        self.heads = heads
        inner_dim = dim_head * heads

        self.norm = nn.LayerNorm(dim)

        self.to_q = nn.Linear(dim, inner_dim, bias = False)
        self.to_kv = nn.Linear(dim, inner_dim * 2, bias = False)
        self.to_out = nn.Linear(inner_dim, dim, bias = False)

        # whether for text to only attend to immediate preceding image, or all images

        self.only_attend_immediate_media = only_attend_immediate_media

    def forward(
        self,
        x,
        media,
        media_locations = None
    ):
        b, t, m = media.shape[:3]
        h = self.heads

        x = self.norm(x)

        q = self.to_q(x)
        media = rearrange(media, 'b t n d -> b (t n) d')

        k, v = self.to_kv(media).chunk(2, dim = -1)
        q, k, v = rearrange_many((q, k, v), 'b n (h d) -> b h n d', h = h)

        q = q * self.scale

        sim = einsum('... i d, ... j d -> ... i j', q, k)

        if exists(media_locations):
            text_time = media_locations.cumsum(dim = -1) # at each boolean of True, increment the time counter (relative to media time)
            media_time = torch.arange(t, device = x.device) + 1

            # text time must equal media time if only attending to most immediate image
            # otherwise, as long as text time is greater than media time (if attending to all previous images / media)
            mask_op = torch.eq if self.only_attend_immediate_media else torch.ge

            text_to_media_mask = mask_op(rearrange(text_time, 'b i -> b 1 i 1'), repeat(media_time, 'j -> 1 1 1 (j m)', m = m))
            sim = sim.masked_fill(~text_to_media_mask, -torch.finfo(sim.dtype).max)

        sim = sim - sim.amax(dim = -1, keepdim = True).detach()
        attn = sim.softmax(dim = -1)

        if exists(media_locations) and self.only_attend_immediate_media:
            # any text without a preceding media needs to have attention zeroed out
            text_without_media_mask = text_time == 0
            text_without_media_mask = rearrange(text_without_media_mask, 'b i -> b 1 i 1')
            attn.masked_fill(text_without_media_mask, 0.)

        out = einsum('... i j, ... j d -> ... i d', attn, v)
        out = rearrange(out, 'b h n d -> b n (h d)')
        return self.to_out(out)

class GatedCrossAttentionBlock(nn.Module):
    def __init__(
        self,
        *,
        dim,
        dim_head = 64,
        heads = 8,
        ff_mult = 4,
        only_attend_immediate_media = True
    ):
        super().__init__()
        self.attn = MaskedCrossAttention(dim = dim, dim_head = dim_head, heads = heads, only_attend_immediate_media = only_attend_immediate_media)
        self.attn_gate = nn.Parameter(torch.tensor([0.]))

        self.ff = FeedForward(dim, mult = ff_mult)
        self.ff_gate = nn.Parameter(torch.tensor([0.]))

    def forward(
        self,
        x,
        media,                  # media tensor, encoded by perceiver resample - (batch, time, latents, dim)
        media_locations = None  # boolean tensor indicating positions of media - (batch, sequence)
    ):
        x = self.attn(x, media, media_locations = media_locations) * self.attn_gate.tanh() + x
        x = self.ff(x) * self.ff_gate.tanh()  + x
        return x