ViT源码讲解
前言
以下学习内容根据我的B站导师 霹雳吧啦Wz 的 ViT源码讲解视频整理而成,文中图片均来自霹雳吧啦Wz,特此记录,方便日后查阅:
源码讲解视频:霹雳吧啦Wz ViT源码讲解
ViT原理视频:霹雳吧啦Wz ViT原理讲解
源码讲解
网络总体架构
ViT中的DropPath
这边不过多介绍,可参看其他博文:droppath
ViT中的Patch Embedding
对应左下角第一个虚线框中的操作。
- 输入image卷积
- 按照卷积后的长宽维度展平
class PatchEmbed(nn.Module):
"""
2D Image to Patch Embedding
"""
def __init__(self, img_size=224, patch_size=16, in_c=3, embed_dim=768, norm_layer=None):
super().__init__()
img_size = (img_size, img_size)
patch_size = (patch_size, patch_size)
self.img_size = img_size
self.patch_size = patch_size
self.grid_size = (img_size[0] // patch_size[0], img_size[1] // patch_size[1]) # 输出特征长宽
self.num_patches = self.grid_size[0] * self.grid_size[1]
self.proj = nn.Conv2d(in_c, embed_dim, kernel_size=patch_size, stride=patch_size)
# 如果传入normlayer则进行normlayer, 否则Identity, 不进行任何操作
self.norm = norm_layer(embed_dim) if norm_layer else nn.Identity()
def forward(self, x):
B, C, H, W = x.shape
# ViT模型的输入图片大小必须是固定的,可能是因为在传入transformer时会加提前训练好的位置编码
assert H == self.img_size[0] and W == self.img_size[1], \
f"Input image size ({H}*{W}) doesn't match model ({self.img_size[0]}*{self.img_size[1]})."
# flatten: [B, C, H, W] -> [B, C, HW]将HW维度合并
# transpose: [B, C, HW] -> [B, HW, C]调整HW和通道维度
x = self.proj(x).flatten(2).transpose(1, 2)
x = self.norm(x)
return x
ViT中的attention模块
对应multi-head attention
class Attention(nn.Module):
def __init__(self,
dim, # 输入token的dim
num_heads=8, # head数量
qkv_bias=False, # 生成qkv过程中是否需要加bias
qk_scale=None,
attn_drop_ratio=0., # Qk后的dropout
proj_drop_ratio=0.):# 多头结果输出后的dropout
super(Attention, self).__init__()
self.num_heads = num_heads
head_dim = dim // num_heads # 每个header的dimension
self.scale = qk_scale or head_dim ** -0.5 # 对应的是softmax中的1/dk**0.5
self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias) # 这边通过将维度*3来一次性生成qkv
self.attn_drop = nn.Dropout(attn_drop_ratio)
self.proj = nn.Linear(dim, dim) #对应的是多头中的 Concat(head1,head2...,headn)*Wo 中的Wo
self.proj_drop = nn.Dropout(proj_drop_ratio)
def forward(self, x):
# [batch_size, num_patches + 1, total_embed_dim] # 1是class token
B, N, C = x.shape
# qkv(): -> [batch_size, num_patches + 1, 3 * total_embed_dim] 最后一个维度扩大三倍 分别对应QKV
# reshape: -> [batch_size, num_patches + 1, 3, num_heads, embed_dim_per_head] 为了将QKV的多头划分出来
# permute: -> [3, batch_size, num_heads, num_patches + 1, embed_dim_per_head]
qkv = self.qkv(x).reshape(B, N, 3, self.num_heads, C // self.num_heads).permute(2, 0, 3, 1, 4)
# [batch_size, num_heads, num_patches + 1, embed_dim_per_head]
q, k, v = qkv[0], qkv[1], qkv[2]
# transpose: -> [batch_size, num_heads, embed_dim_per_head, num_patches + 1]
# @: multiply -> [batch_size, num_heads, num_patches + 1, num_patches + 1]
# @表示矩阵乘法,仅最后两个维度相乘
attn = (q @ k.transpose(-2, -1)) * self.scale
# 对最后一个维度进行softmax处理
attn = attn.softmax(dim=-1)
# 对V的每个权重进行dropout
attn = self.attn_drop(attn)
# @: multiply -> [batch_size, num_heads, num_patches + 1, embed_dim_per_head]
# transpose: -> [batch_size, num_patches + 1, num_heads, embed_dim_per_head]
# reshape: -> [batch_size, num_patches + 1, total_embed_dim]
x = (attn @ v).transpose(1, 2).reshape(B, N, C)
x = self.proj(x)
x = self.proj_drop(x)
return x
ViT中的MLPblock模块
对应图中MLPblock
class Mlp(nn.Module):
"""
MLP as used in Vision Transformer, MLP-Mixer and related networks
"""
def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
super().__init__()
out_features = out_features or in_features
hidden_features = hidden_features or in_features
self.fc1 = nn.Linear(in_features, hidden_features)
self.act = act_layer() # GELU激活函数
self.fc2 = nn.Linear(hidden_features, out_features)
self.drop = nn.Dropout(drop) # dropout可以重复使用
def forward(self, x):
x = self.fc1(x)
x = self.act(x)
x = self.drop(x)
x = self.fc2(x)
x = self.drop(x)
return x
ViT中的Encoder Block模块
class Block(nn.Module):
def __init__(self,
dim, # 对于每个token的dimension
num_heads, # head的个数
mlp_ratio=4., # MLP中第一个全连接层是输入维度的四倍
qkv_bias=False,
qk_scale=None,
drop_ratio=0., # attention中输出以及MLP中dropout比例
attn_drop_ratio=0., # QKT的dropout
drop_path_ratio=0., # Encoder Block中的droppath
act_layer=nn.GELU,
norm_layer=nn.LayerNorm):
super(Block, self).__init__()
self.norm1 = norm_layer(dim) # 如果normlayer层没有可学习参数,就没必要定义两个norm1以及norm2,直接self.norm = norm_layer(dim) 可复用
self.attn = Attention(dim, num_heads=num_heads, qkv_bias=qkv_bias, qk_scale=qk_scale,
attn_drop_ratio=attn_drop_ratio, proj_drop_ratio=drop_ratio)
# NOTE: drop path for stochastic depth, we shall see if this is better than dropout here
self.drop_path = DropPath(drop_path_ratio) if drop_path_ratio > 0. else nn.Identity()
self.norm2 = norm_layer(dim)
mlp_hidden_dim = int(dim * mlp_ratio)
self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop_ratio)
def forward(self, x):
x = x + self.drop_path(self.attn(self.norm1(x))) # 对应原图中attention处的残差结构, 仅encoder block部分使用了droppath
x = x + self.drop_path(self.mlp(self.norm2(x))) # 对应原图中MLPblock处的残差结构
return x
整个ViT模型
class VisionTransformer(nn.Module):
def __init__(self, img_size=224, patch_size=16, in_c=3, num_classes=1000,
embed_dim=768, depth=12, num_heads=12, mlp_ratio=4.0, qkv_bias=True,
qk_scale=None, representation_size=None, distilled=False, drop_ratio=0.,
attn_drop_ratio=0., drop_path_ratio=0., embed_layer=PatchEmbed, norm_layer=None,
act_layer=None):
"""
Args:
img_size (int, tuple): input image size
patch_size (int, tuple): patch size
in_c (int): number of input channels
num_classes (int): number of classes for classification head
embed_dim (int): embedding dimension
depth (int): depth of transformer
num_heads (int): number of attention heads
mlp_ratio (int): ratio of mlp hidden dim to embedding dim
qkv_bias (bool): enable bias for qkv if True
qk_scale (float): override default qk scale of head_dim ** -0.5 if set
representation_size (Optional[int]): enable and set representation layer (pre-logits) to this value if set
distilled (bool): model includes a distillation token and head as in DeiT models 不用管
drop_ratio (float): dropout rate
attn_drop_ratio (float): attention dropout rate
drop_path_ratio (float): stochastic depth rate
embed_layer (nn.Module): patch embedding layer
norm_layer: (nn.Module): normalization layer
"""
super(VisionTransformer, self).__init__()
self.num_classes = num_classes
self.num_features = self.embed_dim = embed_dim # num_features for consistency with other models
self.num_tokens = 2 if distilled else 1 # 1
norm_layer = norm_layer or partial(nn.LayerNorm, eps=1e-6) # partial方法,将eps参数传入LayerNorm
act_layer = act_layer or nn.GELU
self.patch_embed = embed_layer(img_size=img_size, patch_size=patch_size, in_c=in_c, embed_dim=embed_dim)
num_patches = self.patch_embed.num_patches
self.cls_token = nn.Parameter(torch.zeros(1, 1, embed_dim)) # nn.Parameters中的参数在反向传播时也会更新
self.dist_token = nn.Parameter(torch.zeros(1, 1, embed_dim)) if distilled else None # 不用管
self.pos_embed = nn.Parameter(torch.zeros(1, num_patches + self.num_tokens, embed_dim))
self.pos_drop = nn.Dropout(p=drop_ratio)
# 在每个Encoder block中使用的droppath参数是递增的
dpr = [x.item() for x in torch.linspace(0, drop_path_ratio, depth)] # stochastic depth decay rule
self.blocks = nn.Sequential(*[
Block(dim=embed_dim, num_heads=num_heads, mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
drop_ratio=drop_ratio, attn_drop_ratio=attn_drop_ratio, drop_path_ratio=dpr[i],
norm_layer=norm_layer, act_layer=act_layer)
for i in range(depth)
])
self.norm = norm_layer(embed_dim)
# Representation layer 如果传入representation_size则会在MLP head中构建pre logist
# 也就是多加一层隐藏层以及Tanh激活函数
if representation_size and not distilled:
self.has_logits = True
self.num_features = representation_size
self.pre_logits = nn.Sequential(OrderedDict([
("fc", nn.Linear(embed_dim, representation_size)),
("act", nn.Tanh())
]))
else:
self.has_logits = False
self.pre_logits = nn.Identity()
# Classifier head(s)
self.head = nn.Linear(self.num_features, num_classes) if num_classes > 0 else nn.Identity()
self.head_dist = None # 不用看
if distilled: # 不用看
self.head_dist = nn.Linear(self.embed_dim, self.num_classes) if num_classes > 0 else nn.Identity()
# Weight init 权重初始化
# 将self.pos_embed/dist_token/cls_token 取值范围限定在[-2,2]区间内,且取值满足mean=0,std=0.02的正态分布
nn.init.trunc_normal_(self.pos_embed, std=0.02)
if self.dist_token is not None:
nn.init.trunc_normal_(self.dist_token, std=0.02)
nn.init.trunc_normal_(self.cls_token, std=0.02)
for m in self.modules(): # 或者自己定义一个函数然后用self.apply(fn)来初始化参数,apply方法会递归遍历所有module中的权重参数,然后按要求进行初始化
if isinstance(m, nn.Linear):
nn.init.trunc_normal_(m.weight, std=.01)
if m.bias is not None:
nn.init.zeros_(m.bias)
elif isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight, mode="fan_out")
if m.bias is not None:
nn.init.zeros_(m.bias)
elif isinstance(m, nn.LayerNorm):
nn.init.zeros_(m.bias)
nn.init.ones_(m.weight)
def forward_features(self, x):
# [B, C, H, W] -> [B, num_patches, embed_dim]
x = self.patch_embed(x) # [B, 196, 768]
# [1, 1, 768] -> [B, 1, 768] # 将cls_token在batch维度复制B次,以与x进行拼接
cls_token = self.cls_token.expand(x.shape[0], -1, -1)
if self.dist_token is None:
x = torch.cat((cls_token, x), dim=1) # [B, 197, 768]
else: # 忽略
x = torch.cat((cls_token, self.dist_token.expand(x.shape[0], -1, -1), x), dim=1)
x = self.pos_drop(x + self.pos_embed)
x = self.blocks(x)
x = self.norm(x)
if self.dist_token is None:
return self.pre_logits(x[:, 0]) # representation_size为None则为identity否则为一层fc
else: # 忽略
return x[:, 0], x[:, 1]
def forward(self, x):
x = self.forward_features(x)
if self.head_dist is not None: # 忽略
x, x_dist = self.head(x[0]), self.head_dist(x[1])
if self.training and not torch.jit.is_scripting():
# during inference, return the average of both classifier predictions
return x, x_dist
else:
return (x + x_dist) / 2
else:
x = self.head(x)
return x
ViT-base/large/huge
建议使用ViT-base模型,其他模型参数量太大,同时要有预训练模型,不然效果较差。
def vit_base_patch16_224(num_classes: int = 1000):
"""
ViT-Base model (ViT-B/16) from original paper (https://arxiv.org/abs/2010.11929).
ImageNet-1k weights @ 224x224, source https://github.com/google-research/vision_transformer.
weights ported from official Google JAX impl:
链接: https://pan.baidu.com/s/1zqb08naP0RPqqfSXfkB2EA 密码: eu9f
"""
model = VisionTransformer(img_size=224,
patch_size=16,
embed_dim=768,
depth=12,
num_heads=12,
representation_size=None,
num_classes=num_classes)
return model
def vit_base_patch16_224_in21k(num_classes: int = 21843, has_logits: bool = True):
"""
ViT-Base model (ViT-B/16) from original paper (https://arxiv.org/abs/2010.11929).
ImageNet-21k weights @ 224x224, source https://github.com/google-research/vision_transformer.
weights ported from official Google JAX impl:
https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-vitjx/jx_vit_base_patch16_224_in21k-e5005f0a.pth
"""
model = VisionTransformer(img_size=224,
patch_size=16,
embed_dim=768,
depth=12,
num_heads=12,
representation_size=768 if has_logits else None,
num_classes=num_classes)
return model
def vit_base_patch32_224(num_classes: int = 1000):
"""
ViT-Base model (ViT-B/32) from original paper (https://arxiv.org/abs/2010.11929).
ImageNet-1k weights @ 224x224, source https://github.com/google-research/vision_transformer.
weights ported from official Google JAX impl:
链接: https://pan.baidu.com/s/1hCv0U8pQomwAtHBYc4hmZg 密码: s5hl
"""
model = VisionTransformer(img_size=224,
patch_size=32,
embed_dim=768,
depth=12,
num_heads=12,
representation_size=None,
num_classes=num_classes)
return model
def vit_base_patch32_224_in21k(num_classes: int = 21843, has_logits: bool = True):
"""
ViT-Base model (ViT-B/32) from original paper (https://arxiv.org/abs/2010.11929).
ImageNet-21k weights @ 224x224, source https://github.com/google-research/vision_transformer.
weights ported from official Google JAX impl:
https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-vitjx/jx_vit_base_patch32_224_in21k-8db57226.pth
"""
model = VisionTransformer(img_size=224,
patch_size=32,
embed_dim=768,
depth=12,
num_heads=12,
representation_size=768 if has_logits else None,
num_classes=num_classes)
return model
def vit_large_patch16_224(num_classes: int = 1000):
"""
ViT-Large model (ViT-L/16) from original paper (https://arxiv.org/abs/2010.11929).
ImageNet-1k weights @ 224x224, source https://github.com/google-research/vision_transformer.
weights ported from official Google JAX impl:
链接: https://pan.baidu.com/s/1cxBgZJJ6qUWPSBNcE4TdRQ 密码: qqt8
"""
model = VisionTransformer(img_size=224,
patch_size=16,
embed_dim=1024,
depth=24,
num_heads=16,
representation_size=None,
num_classes=num_classes)
return model
def vit_large_patch16_224_in21k(num_classes: int = 21843, has_logits: bool = True):
"""
ViT-Large model (ViT-L/16) from original paper (https://arxiv.org/abs/2010.11929).
ImageNet-21k weights @ 224x224, source https://github.com/google-research/vision_transformer.
weights ported from official Google JAX impl:
https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-vitjx/jx_vit_large_patch16_224_in21k-606da67d.pth
"""
model = VisionTransformer(img_size=224,
patch_size=16,
embed_dim=1024,
depth=24,
num_heads=16,
representation_size=1024 if has_logits else None,
num_classes=num_classes)
return model
def vit_large_patch32_224_in21k(num_classes: int = 21843, has_logits: bool = True):
"""
ViT-Large model (ViT-L/32) from original paper (https://arxiv.org/abs/2010.11929).
ImageNet-21k weights @ 224x224, source https://github.com/google-research/vision_transformer.
weights ported from official Google JAX impl:
https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-vitjx/jx_vit_large_patch32_224_in21k-9046d2e7.pth
"""
model = VisionTransformer(img_size=224,
patch_size=32,
embed_dim=1024,
depth=24,
num_heads=16,
representation_size=1024 if has_logits else None,
num_classes=num_classes)
return model
def vit_huge_patch14_224_in21k(num_classes: int = 21843, has_logits: bool = True):
"""
ViT-Huge model (ViT-H/14) from original paper (https://arxiv.org/abs/2010.11929).
ImageNet-21k weights @ 224x224, source https://github.com/google-research/vision_transformer.
NOTE: converted weights not currently available, too large for github release hosting.
"""
model = VisionTransformer(img_size=224,
patch_size=14,
embed_dim=1280,
depth=32,
num_heads=16,
representation_size=1280 if has_logits else None,
num_classes=num_classes)
return model