File size: 5,614 Bytes
c7a0be3 |
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 |
# Copyright (c) Alibaba Cloud.
#
# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.
from collections import OrderedDict
import math
import requests
from io import BytesIO
from functools import partial
from PIL import Image
from typing import Callable, Optional, Sequence, Tuple, List, Union
import numpy as np
import torch
from torch import nn
from torch.nn import functional as F
from torch.nn.init import trunc_normal_
from torchvision import transforms
from torchvision.transforms import InterpolationMode
def get_abs_pos(abs_pos, tgt_size):
# abs_pos: L, C
# tgt_size: (H, W)
# return: M, C
src_size = int(math.sqrt(abs_pos.size(0)))
dtype = abs_pos.dtype
return F.interpolate(
abs_pos.float().reshape(1, src_size, src_size, -1).permute(0, 3, 1, 2),
size=(tgt_size[0], tgt_size[1]),
mode="bicubic",
align_corners=False,
).permute(0, 2, 3, 1).flatten(0, 2).to(dtype=dtype)
# https://github.com/facebookresearch/mae/blob/efb2a8062c206524e35e47d04501ed4f544c0ae8/util/pos_embed.py#L20
def get_2d_sincos_pos_embed(embed_dim, grid_size, cls_token=False):
"""
grid_size: int of the grid height and width
return:
pos_embed: [grid_size*grid_size, embed_dim] or [1+grid_size*grid_size, embed_dim] (w/ or w/o cls_token)
"""
if isinstance(grid_size, int):
grid_h_size, grid_w_size = grid_size, grid_size
else:
grid_h_size, grid_w_size = grid_size[0], grid_size[1]
grid_h = np.arange(grid_h_size, dtype=np.float32)
grid_w = np.arange(grid_w_size, dtype=np.float32)
grid = np.meshgrid(grid_w, grid_h) # here w goes first
grid = np.stack(grid, axis=0)
grid = grid.reshape([2, 1, grid_h_size, grid_w_size])
pos_embed = get_2d_sincos_pos_embed_from_grid(embed_dim, grid)
if cls_token:
pos_embed = np.concatenate([np.zeros([1, embed_dim]), pos_embed], axis=0)
return pos_embed
def get_2d_sincos_pos_embed_from_grid(embed_dim, grid):
assert embed_dim % 2 == 0
# use half of dimensions to encode grid_h
emb_h = get_1d_sincos_pos_embed_from_grid(embed_dim // 2, grid[0]) # (H*W, D/2)
emb_w = get_1d_sincos_pos_embed_from_grid(embed_dim // 2, grid[1]) # (H*W, D/2)
emb = np.concatenate([emb_h, emb_w], axis=1) # (H*W, D)
return emb
def get_1d_sincos_pos_embed_from_grid(embed_dim, pos):
"""
embed_dim: output dimension for each position
pos: a list of positions to be encoded: size (M,)
out: (M, D)
"""
assert embed_dim % 2 == 0
omega = np.arange(embed_dim // 2, dtype=np.float32)
omega /= embed_dim / 2.
omega = 1. / 10000 ** omega # (D/2,)
pos = pos.reshape(-1) # (M,)
out = np.einsum('m,d->md', pos, omega) # (M, D/2), outer product
emb_sin = np.sin(out) # (M, D/2)
emb_cos = np.cos(out) # (M, D/2)
emb = np.concatenate([emb_sin, emb_cos], axis=1) # (M, D)
return emb
class Resampler(nn.Module):
"""
A 2D perceiver-resampler network with one cross attention layers by
(grid_size**2) learnable queries and 2d sincos pos_emb
Outputs:
A tensor with the shape of (grid_size**2, embed_dim)
"""
def __init__(
self,
grid_size,
embed_dim,
num_heads,
kv_dim=None,
norm_layer=partial(nn.LayerNorm, eps=1e-6),
adaptive=False
):
super().__init__()
self.num_queries = grid_size ** 2
self.embed_dim = embed_dim
self.num_heads = num_heads
self.adaptive = adaptive
self.pos_embed = nn.Parameter(
torch.from_numpy(get_2d_sincos_pos_embed(embed_dim, grid_size)).float()
).requires_grad_(False)
self.query = nn.Parameter(torch.zeros(self.num_queries, embed_dim))
trunc_normal_(self.query, std=.02)
if kv_dim is not None and kv_dim != embed_dim:
self.kv_proj = nn.Linear(kv_dim, embed_dim, bias=False)
else:
self.kv_proj = nn.Identity()
self.attn = nn.MultiheadAttention(embed_dim, num_heads)
self.ln_q = norm_layer(embed_dim)
self.ln_kv = norm_layer(embed_dim)
self.ln_post = norm_layer(embed_dim)
self.proj = nn.Parameter((embed_dim ** -0.5) * torch.randn(embed_dim, embed_dim))
self.apply(self._init_weights)
def _init_weights(self, m):
if isinstance(m, nn.Linear):
trunc_normal_(m.weight, std=.02)
if isinstance(m, nn.Linear) and m.bias is not None:
nn.init.constant_(m.bias, 0)
elif isinstance(m, nn.LayerNorm):
nn.init.constant_(m.bias, 0)
nn.init.constant_(m.weight, 1.0)
def forward(self, x, tgt_size=None, attn_mask=None):
if self.adaptive:
# print("adaptive")
# raise Exception
pos_embed = torch.Tensor(get_2d_sincos_pos_embed(self.embed_dim, tgt_size)).float().to(device=x.device, dtype=x.dtype)
else:
pos_embed = get_abs_pos(self.pos_embed, tgt_size)
x = self.kv_proj(x)
x = self.ln_kv(x).permute(1, 0, 2)
N = x.shape[1]
q = self.ln_q(self.query)
out = self.attn(
self._repeat(q, N) + self.pos_embed.unsqueeze(1),
x + pos_embed.unsqueeze(1),
x,
attn_mask=attn_mask)[0]
x = out.permute(1, 0, 2)
x = self.ln_post(x)
x = x @ self.proj
return x
def _repeat(self, query, N: int):
return query.unsqueeze(1).repeat(1, N, 1)
|