PSHuman / lib /net /Transformer.py
fffiloni's picture
Migrated from GitHub
2252f3d verified
raw
history blame
18.2 kB
# ------------------------------------------------------------------------------------
# Enhancing Transformers
# Copyright (c) 2022 Thuan H. Nguyen. All Rights Reserved.
# Licensed under the MIT License [see LICENSE for details]
# ------------------------------------------------------------------------------------
# Modified from ViT-Pytorch (https://github.com/lucidrains/vit-pytorch)
# Copyright (c) 2020 Phil Wang. All Rights Reserved.
# ------------------------------------------------------------------------------------
import math
import numpy as np
from typing import Union, Tuple, List, Optional
from functools import partial
import pytorch_lightning as pl
import torch
import torch.nn as nn
import torch.nn.functional as F
from einops import rearrange, repeat
from einops.layers.torch import Rearrange
def get_2d_sincos_pos_embed(embed_dim, grid_size):
"""
grid_size: int or (int, 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)
"""
grid_size = (grid_size, grid_size) if type(grid_size) != tuple else grid_size
grid_h = np.arange(grid_size[0], dtype=np.float32)
grid_w = np.arange(grid_size[1], 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_size[0], grid_size[1]])
pos_embed = get_2d_sincos_pos_embed_from_grid(embed_dim, grid)
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
def init_weights(m):
if isinstance(m, nn.Linear):
# we use xavier_uniform following official JAX ViT:
torch.nn.init.xavier_uniform_(m.weight)
if 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)
elif isinstance(m, nn.Conv2d) or isinstance(m, nn.ConvTranspose2d):
w = m.weight.data
torch.nn.init.xavier_uniform_(w.view([w.shape[0], -1]))
class PreNorm(nn.Module):
def __init__(self, dim: int, fn: nn.Module) -> None:
super().__init__()
self.norm = nn.LayerNorm(dim)
self.fn = fn
def forward(self, x: torch.FloatTensor, **kwargs) -> torch.FloatTensor:
return self.fn(self.norm(x), **kwargs)
class FeedForward(nn.Module):
def __init__(self, dim: int, hidden_dim: int) -> None:
super().__init__()
self.net = nn.Sequential(
nn.Linear(dim, hidden_dim),
nn.Tanh(),
nn.Linear(hidden_dim, dim)
)
def forward(self, x: torch.FloatTensor) -> torch.FloatTensor:
return self.net(x)
class Attention(nn.Module):
def __init__(self, dim: int, heads: int = 8, dim_head: int = 64) -> None:
super().__init__()
inner_dim = dim_head * heads
project_out = not (heads == 1 and dim_head == dim)
self.heads = heads
self.scale = dim_head ** -0.5
self.attend = nn.Softmax(dim = -1)
self.to_qkv = nn.Linear(dim, inner_dim * 3, bias = False)
self.to_out = nn.Linear(inner_dim, dim) if project_out else nn.Identity()
def forward(self, x: torch.FloatTensor) -> torch.FloatTensor:
qkv = self.to_qkv(x).chunk(3, dim = -1)
q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> b h n d', h = self.heads), qkv)
attn = torch.matmul(q, k.transpose(-1, -2)) * self.scale
attn = self.attend(attn)
out = torch.matmul(attn, v)
out = rearrange(out, 'b h n d -> b n (h d)')
return self.to_out(out)
class CrossAttention(nn.Module):
def __init__(self, dim: int, heads: int = 8, dim_head: int = 64) -> None:
super().__init__()
inner_dim = dim_head * heads
project_out = not (heads == 1 and dim_head == dim)
self.heads = heads
self.scale = dim_head ** -0.5
self.attend = nn.Softmax(dim = -1)
self.to_kv = nn.Linear(dim, inner_dim * 2, bias = False)
self.to_q = nn.Linear(dim, inner_dim, bias = False)
self.norm = nn.LayerNorm(dim)
self.to_out = nn.Linear(inner_dim, dim) if project_out else nn.Identity()
self.multi_head_attention=PreNorm(dim, Attention(dim, heads=heads, dim_head=dim_head))
def forward(self, x: torch.FloatTensor, q_x:torch.FloatTensor) -> torch.FloatTensor:
q_in = self.multi_head_attention(q_x)+q_x
q_in = self.norm(q_in)
q = rearrange(self.to_q(q_in),'b n (h d) -> b h n d', h = self.heads)
kv = self.to_kv(x).chunk(2, dim = -1)
k, v = map(lambda t: rearrange(t, 'b n (h d) -> b h n d', h = self.heads), kv)
attn = torch.matmul(q, k.transpose(-1, -2)) * self.scale
attn = self.attend(attn)
out = torch.matmul(attn, v)
out = rearrange(out, 'b h n d -> b n (h d)')
return self.to_out(out),q_in
class Transformer(nn.Module):
def __init__(self, dim: int, depth: int, heads: int, dim_head: int, mlp_dim: int) -> None:
super().__init__()
self.layers = nn.ModuleList([])
for idx in range(depth):
layer = nn.ModuleList([PreNorm(dim, Attention(dim, heads=heads, dim_head=dim_head)),
PreNorm(dim, FeedForward(dim, mlp_dim))])
self.layers.append(layer)
self.norm = nn.LayerNorm(dim)
def forward(self, x: torch.FloatTensor) -> torch.FloatTensor:
for attn, ff in self.layers:
x = attn(x) + x
x = ff(x) + x
return self.norm(x)
class CrossTransformer(nn.Module):
def __init__(self, dim: int, depth: int, heads: int, dim_head: int, mlp_dim: int) -> None:
super().__init__()
self.layers = nn.ModuleList([])
for idx in range(depth):
layer = nn.ModuleList([CrossAttention(dim, heads=heads, dim_head=dim_head),
PreNorm(dim, FeedForward(dim, mlp_dim))])
self.layers.append(layer)
self.norm = nn.LayerNorm(dim)
def forward(self, x: torch.FloatTensor, q_x:torch.FloatTensor) -> torch.FloatTensor:
encoder_output=x
for attn, ff in self.layers:
x,q_in = attn(encoder_output, q_x)
x = x + q_in
x = ff(x) + x
q_x=x
return self.norm(q_x)
class ViTEncoder(nn.Module):
def __init__(self, image_size: Union[Tuple[int, int], int], patch_size: Union[Tuple[int, int], int],
dim: int, depth: int, heads: int, mlp_dim: int, channels: int = 3, dim_head: int = 64) -> None:
super().__init__()
image_height, image_width = image_size if isinstance(image_size, tuple) \
else (image_size, image_size)
patch_height, patch_width = patch_size if isinstance(patch_size, tuple) \
else (patch_size, patch_size)
assert image_height % patch_height == 0 and image_width % patch_width == 0, 'Image dimensions must be divisible by the patch size.'
en_pos_embedding = get_2d_sincos_pos_embed(dim, (image_height // patch_height, image_width // patch_width))
self.num_patches = (image_height // patch_height) * (image_width // patch_width)
self.patch_dim = channels * patch_height * patch_width
self.to_patch_embedding = nn.Sequential(
nn.Conv2d(channels, dim, kernel_size=patch_size, stride=patch_size),
Rearrange('b c h w -> b (h w) c'),
)
self.en_pos_embedding = nn.Parameter(torch.from_numpy(en_pos_embedding).float().unsqueeze(0), requires_grad=False)
self.transformer = Transformer(dim, depth, heads, dim_head, mlp_dim)
self.apply(init_weights)
def forward(self, img: torch.FloatTensor) -> torch.FloatTensor:
x = self.to_patch_embedding(img)
x = x + self.en_pos_embedding
x = self.transformer(x)
return x
class ViTDecoder(nn.Module):
def __init__(self, image_size: Union[Tuple[int, int], int], patch_size: Union[Tuple[int, int], int],
dim: int, depth: int, heads: int, mlp_dim: int, channels: int = 32, dim_head: int = 64) -> None:
super().__init__()
image_height, image_width = image_size if isinstance(image_size, tuple) \
else (image_size, image_size)
patch_height, patch_width = patch_size if isinstance(patch_size, tuple) \
else (patch_size, patch_size)
assert image_height % patch_height == 0 and image_width % patch_width == 0, 'Image dimensions must be divisible by the patch size.'
de_pos_embedding = get_2d_sincos_pos_embed(dim, (image_height // patch_height, image_width // patch_width))
self.num_patches = (image_height // patch_height) * (image_width // patch_width)
self.patch_dim = channels * patch_height * patch_width
self.transformer = Transformer(dim, depth, heads, dim_head, mlp_dim)
self.de_pos_embedding = nn.Parameter(torch.from_numpy(de_pos_embedding).float().unsqueeze(0), requires_grad=False)
self.to_pixel = nn.Sequential(
Rearrange('b (h w) c -> b c h w', h=image_height // patch_height),
nn.ConvTranspose2d(dim, channels, kernel_size=4, stride=4)
)
self.apply(init_weights)
def forward(self, token: torch.FloatTensor) -> torch.FloatTensor:
x = token + self.de_pos_embedding
x = self.transformer(x)
x = self.to_pixel(x)
return x
def get_last_layer(self) -> nn.Parameter:
return self.to_pixel[-1].weight
class CrossAttDecoder(nn.Module):
def __init__(self, image_size: Union[Tuple[int, int], int], patch_size: Union[Tuple[int, int], int],
dim: int, depth: int, heads: int, mlp_dim: int, channels: int = 32, dim_head: int = 64) -> None:
super().__init__()
image_height, image_width = image_size if isinstance(image_size, tuple) \
else (image_size, image_size)
patch_height, patch_width = patch_size if isinstance(patch_size, tuple) \
else (patch_size, patch_size)
self.to_patch_embedding = nn.Sequential(
nn.Conv2d(3, dim, kernel_size=patch_size, stride=patch_size),
Rearrange('b c h w -> b (h w) c'),
)
assert image_height % patch_height == 0 and image_width % patch_width == 0, 'Image dimensions must be divisible by the patch size.'
de_pos_embedding = get_2d_sincos_pos_embed(dim, (image_height // patch_height, image_width // patch_width))
self.num_patches = (image_height // patch_height) * (image_width // patch_width)
self.patch_dim = channels * patch_height * patch_width
self.transformer = CrossTransformer(dim, depth, heads, dim_head, mlp_dim)
self.de_pos_embedding = nn.Parameter(torch.from_numpy(de_pos_embedding).float().unsqueeze(0), requires_grad=False)
self.to_pixel = nn.Sequential(
Rearrange('b (h w) c -> b c h w', h=image_height // patch_height),
nn.ConvTranspose2d(dim, channels, kernel_size=4, stride=4)
)
self.apply(init_weights)
def forward(self, token: torch.FloatTensor, query_img:torch.FloatTensor) -> torch.FloatTensor:
# batch_size=token.shape[0]
# query=self.query.repeat(batch_size,1,1)+self.de_pos_embedding
query=self.to_patch_embedding(query_img)+self.de_pos_embedding
x = token + self.de_pos_embedding
x = self.transformer(x,query)
x = self.to_pixel(x)
return x
def get_last_layer(self) -> nn.Parameter:
return self.to_pixel[-1].weight
class BaseQuantizer(nn.Module):
def __init__(self, embed_dim: int, n_embed: int, straight_through: bool = True, use_norm: bool = True,
use_residual: bool = False, num_quantizers: Optional[int] = None) -> None:
super().__init__()
self.straight_through = straight_through
self.norm = lambda x: F.normalize(x, dim=-1) if use_norm else x
self.use_residual = use_residual
self.num_quantizers = num_quantizers
self.embed_dim = embed_dim
self.n_embed = n_embed
self.embedding = nn.Embedding(self.n_embed, self.embed_dim)
self.embedding.weight.data.normal_()
def quantize(self, z: torch.FloatTensor) -> Tuple[torch.FloatTensor, torch.FloatTensor, torch.LongTensor]:
pass
def forward(self, z: torch.FloatTensor) -> Tuple[torch.FloatTensor, torch.FloatTensor, torch.LongTensor]:
if not self.use_residual:
z_q, loss, encoding_indices = self.quantize(z)
else:
z_q = torch.zeros_like(z)
residual = z.detach().clone()
losses = []
encoding_indices = []
for _ in range(self.num_quantizers):
z_qi, loss, indices = self.quantize(residual.clone())
residual.sub_(z_qi)
z_q.add_(z_qi)
encoding_indices.append(indices)
losses.append(loss)
losses, encoding_indices = map(partial(torch.stack, dim = -1), (losses, encoding_indices))
loss = losses.mean()
# preserve gradients with straight-through estimator
if self.straight_through:
z_q = z + (z_q - z).detach()
return z_q, loss, encoding_indices
class VectorQuantizer(BaseQuantizer):
def __init__(self, embed_dim: int, n_embed: int, beta: float = 0.25, use_norm: bool = True,
use_residual: bool = False, num_quantizers: Optional[int] = None, **kwargs) -> None:
super().__init__(embed_dim, n_embed, True,
use_norm, use_residual, num_quantizers)
self.beta = beta
def quantize(self, z: torch.FloatTensor) -> Tuple[torch.FloatTensor, torch.FloatTensor, torch.LongTensor]:
z_reshaped_norm = self.norm(z.view(-1, self.embed_dim))
embedding_norm = self.norm(self.embedding.weight)
d = torch.sum(z_reshaped_norm ** 2, dim=1, keepdim=True) + \
torch.sum(embedding_norm ** 2, dim=1) - 2 * \
torch.einsum('b d, n d -> b n', z_reshaped_norm, embedding_norm)
encoding_indices = torch.argmin(d, dim=1).unsqueeze(1)
encoding_indices = encoding_indices.view(*z.shape[:-1])
z_q = self.embedding(encoding_indices).view(z.shape)
z_qnorm, z_norm = self.norm(z_q), self.norm(z)
# compute loss for embedding
loss = self.beta * torch.mean((z_qnorm.detach() - z_norm)**2) + \
torch.mean((z_qnorm - z_norm.detach())**2)
return z_qnorm, loss, encoding_indices
class ViTVQ(pl.LightningModule):
def __init__(self,image_size=512, patch_size=16,channels=3) -> None:
super().__init__()
self.encoder = ViTEncoder(image_size=image_size, patch_size=patch_size, dim=256,depth=8,heads=8,mlp_dim=2048,channels=channels)
self.F_decoder = ViTDecoder(image_size=image_size, patch_size=patch_size, dim=256,depth=3,heads=8,mlp_dim=2048)
self.B_decoder= CrossAttDecoder(image_size=image_size, patch_size=patch_size, dim=256,depth=3,heads=8,mlp_dim=2048)
self.R_decoder= CrossAttDecoder(image_size=image_size, patch_size=patch_size, dim=256,depth=3,heads=8,mlp_dim=2048)
self.L_decoder= CrossAttDecoder(image_size=image_size, patch_size=patch_size, dim=256,depth=3,heads=8,mlp_dim=2048)
# self.quantizer = VectorQuantizer(embed_dim=32,n_embed=8192)
# self.pre_quant = nn.Linear(512, 32)
# self.post_quant = nn.Linear(32, 512)
def forward(self, x: torch.FloatTensor,smpl_normal) -> torch.FloatTensor:
enc_out = self.encode(x)
dec = self.decode(enc_out,smpl_normal)
return dec
def encode(self, x: torch.FloatTensor) -> Tuple[torch.FloatTensor, torch.FloatTensor]:
h = self.encoder(x)
# h = self.pre_quant(h)
# quant, emb_loss, _ = self.quantizer(h)
return h #, emb_loss
def decode(self, enc_out: torch.FloatTensor,smpl_normal) -> torch.FloatTensor:
back_query=smpl_normal['T_normal_B']
right_query=smpl_normal['T_normal_R']
left_query=smpl_normal['T_normal_L']
# quant = self.post_quant(quant)
dec_F = self.F_decoder(enc_out)
dec_B = self.B_decoder(enc_out,back_query)
dec_R = self.R_decoder(enc_out,right_query)
dec_L = self.L_decoder(enc_out,left_query)
return (dec_F,dec_B,dec_R,dec_L)
# def encode_codes(self, x: torch.FloatTensor) -> torch.LongTensor:
# h = self.encoder(x)
# h = self.pre_quant(h)
# _, _, codes = self.quantizer(h)
# return codes
# def decode_codes(self, code: torch.LongTensor) -> torch.FloatTensor:
# quant = self.quantizer.embedding(code)
# quant = self.quantizer.norm(quant)
# if self.quantizer.use_residual:
# quant = quant.sum(-2)
# dec = self.decode(quant)
# return dec