【Transformer】TNT: Transformer iN Transformer

文章目录

一、背景
二、动机
三、方法
- 3.1 Transformer in Transformer
- 3.2 Network Architecture
四、效果
五、代码

论文链接：https://arxiv.org/pdf/2103.00112.pdf
代码链接：https://github.com/huawei-noah/CV-Backbones/tree/master/tnt_pytorch

一、背景

Transformer 是一种主要基于注意力机制的网络结构，能提取输入数据的特征。计算机视觉中的 Transformer 首先将输入图像均分为多个图像块，然后提取器特征和关系。因为图像数据有很多了细节纹理和颜色等信息，所以目前方法所分的图像块的粒度还不够细，以至于难以挖掘不同尺度和位置特征。

二、动机

计算机视觉使用的是图像作为输入，所以输入和输出在语义上有很大的 gap，所以 ViT 将图像分成了几块组成一个 sequence 作为输入，然后使用 attention 计算不同 patch 之间的关系，来生成图像特征表达。

虽然 ViT 及其变体能够较好的提取图像特征来实现识别的任务，但数据集中的图像是很多样的，虽然将图像分块能够很好的找到不同 patch 之间的关系和相似性，但不同图像块内的小图像块也存在很高的相似性。

作者受此启发，提出了一个更加精细的图像划分方法，来生成 visual sequence 并且提升效果。

三、方法

本文中，作者指出，每个 patch 内部的 attention 对网络效果的影响是非常重要的，所以提出了一种新的结构 TNT。

将 patch（16x16）看做 “visual sentence”
将 patch 再切分成更细的 patch（4x4），看做 “visual words”
会计算 sentence 中每个 words 和其他 words 的 attention
然后把 sentences 和 words 的特征进行聚合来提高网络特征的表达能力
TNT 在 ImageNet 上达到了 81.5% 的 top-1 acc，比当前 SOTA 提升了 1.7%

3.1 Transformer in Transformer

1、输入 2D 图像，切分成 $n$ 个 patches

2、将每个 patch 切分成 $m$ 个 sub-patches， $x^{i,j}$ 是第 i 个 visual sentence 中的第 j 个word

3、对 visual sentence 和 visual word 分别进行处理

inner transformer block 用来对 word 之间建模，outer transformer block 捕捉 sentence 之间的关系。

通过对 TNT block 的 L 次堆叠，得到 transformer-in-transformer network。类似于 ViT，classification token 在这里也作为图像整体特征的表达。

① 对于 word embedding

使用线性映射将 visual word 进行 embedding，得到如下的 $Y$

使用 transformer 来抽取不同 words 之间的关系，L 是 block 的总数，第一个 block $Y_0^i$ 的输入是上面的 $Y^i$ 。

经过 transformer 之后，图像中的所有 word embedding 可以表示为 $Y_l=[Y_l^1, Y_l^2, ..., Y_l^n]$ ，这也可以看成一个内部 transformer block，叫做 $T_{in}$ ，这个过程给每个 sentence 内部的所有 word 两两之间建立了关系特征。

举例来说，就是在一个人脸图分割的patch中，一个眼睛对应的 word 和另外一个眼睛对应的 word 关联更高，而与前额关联更少一些。

② 对于 sentence embedding

作者建立了一个 sentence embedding memories 来存储 sentence 层面的序列表达， $Z_{class}$ 是 class token，用来分类：

在每一层 sentence embedding，会将 word embedding 的序列结果经过线性变换后加到 sentence embedding上， $W$ 和 $b$ 分别为权重和偏置

之后，使用标准的 transformer 来进行特征提取，即使用 outer transformer $T_{out}$ 来对不同的 sentence embedding 进行关系建模。

总之，TNT block 的输入和输出都包括了 word embedding 和 sentence embedding，如图1b所示，所以 TNT 总体表达如下：

3、position encoding

对于 word 和 sentence，分别都要使用位置编码来保持空间信息。作者在这里使用的是可学习的一维位置编码。

对于 sentence，每个 sentence 都被分配了一个位置编码， $Esentence∈R(n+1)×dE_{sentence}\in R^{(n+1)\times d}$ 是 sentence 位置编码。

对于sentence 中的 word，也给每个都加上了位置编码， $E_{word}$ 是 word 位置编码，所有 sentence 内的 word 编码是共享的。

3.2 Network Architecture

patch size：16x16
sun-patch size：m=4x4=16
TNT-S(Small, 23.8M) 和 TNT-B(Base, 65.6M) 的结构

四、效果

消融实验：

可视化：

1、特征图可视化

① DeiT 和 TNT 学习到的特征图可视化如下，第 1、6、12 个block 的特征图如图3a所示，TNT 的位置信息保存的更好。

图3b 使用 t-SNE 可视化了第 12 个 block 的所有 384 个 feature map，可以看出来 TNT 的特征图更加多样，保存了更丰富的信息，这可以归功于 inner transformer 对局部特征的建模能力。

② 除了 patch-level 外，作者还可视化了 pixel-level 的 embedding 如图4，对于每个 patch，作者根据空间位置进行了 reshape，然后把所有通道进行了平均。平均后的特征图大小为 14x14，可以看出在浅层局部特征保存的较好，深层特征越来越抽象。

2、Attention map 的可视化

在 TNT block 里边有两个不同的 self-attention，inner self-attention 和 outer self-attention，图5展示了 inner transformer 的不同 query 的 attention map。

① visual word 可视化

对于一个给定的 visual word，和该 visual word 想外观越相似的 word 的 attention value 越高，这也表示他们的特征将和 query 进行更密切的交互，但 ViT 和 DeiT 就没有这种特性。

② visual sentence 可视化

图6展示了某个 patch 对其他所有 patch 的 attention map，随着层的加深，更多的 patch 会有响应。这是因为在越深的层，patch 之间的信息会更好的被关联起来。

在 block-12，TNT 能够关注到有用的 patch，而 DeiT 仍然会关注到和 panda 无关的 patch 上去。

3、class token 和 patch 之间的 attention

图7可视化了 class token 和图像中的所有 patches 之间的关系，可以看出输出特征会更加注意到和目标位置关联的patch

五、代码

import torch
from tnt import TNT
tnt = TNT()
tnt.eval()
inputs = torch.randn(1, 3, 224, 224)
logits = tnt(inputs)

tnt.py 代码如下：


import torch
import torch.nn as nn
from functools import partial
import mathfrom timm.data import IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD
from timm.models.helpers import load_pretrained
from timm.models.layers import DropPath, to_2tuple, trunc_normal_
from timm.models.resnet import resnet26d, resnet50d
from timm.models.registry import register_modeldef _cfg(url='', **kwargs):return {'url': url,'num_classes': 1000, 'input_size': (3, 224, 224), 'pool_size': None,'crop_pct': .9, 'interpolation': 'bicubic','mean': IMAGENET_DEFAULT_MEAN, 'std': IMAGENET_DEFAULT_STD,'first_conv': 'patch_embed.proj', 'classifier': 'head',**kwargs}default_cfgs = {'tnt_s_patch16_224': _cfg(mean=(0.5, 0.5, 0.5), std=(0.5, 0.5, 0.5),),'tnt_b_patch16_224': _cfg(mean=(0.5, 0.5, 0.5), std=(0.5, 0.5, 0.5),),
}def make_divisible(v, divisor=8, min_value=None):min_value = min_value or divisornew_v = max(min_value, int(v + divisor / 2) // divisor * divisor)# Make sure that round down does not go down by more than 10%.if new_v < 0.9 * v:new_v += divisorreturn new_vclass Mlp(nn.Module):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_featureshidden_features = hidden_features or in_featuresself.fc1 = nn.Linear(in_features, hidden_features)self.act = act_layer()self.fc2 = nn.Linear(hidden_features, out_features)self.drop = nn.Dropout(drop)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 xclass SE(nn.Module):def __init__(self, dim, hidden_ratio=None):super().__init__()hidden_ratio = hidden_ratio or 1self.dim = dimhidden_dim = int(dim * hidden_ratio)self.fc = nn.Sequential(nn.LayerNorm(dim),nn.Linear(dim, hidden_dim),nn.ReLU(inplace=True),nn.Linear(hidden_dim, dim),nn.Tanh())def forward(self, x):a = x.mean(dim=1, keepdim=True) # B, 1, Ca = self.fc(a)x = a * xreturn xclass Attention(nn.Module):def __init__(self, dim, hidden_dim, num_heads=8, qkv_bias=False, qk_scale=None, attn_drop=0., proj_drop=0.):super().__init__()self.hidden_dim = hidden_dimself.num_heads = num_headshead_dim = hidden_dim // num_headsself.head_dim = head_dim# NOTE scale factor was wrong in my original version, can set manually to be compat with prev weightsself.scale = qk_scale or head_dim ** -0.5self.qk = nn.Linear(dim, hidden_dim * 2, bias=qkv_bias)self.v = nn.Linear(dim, dim, bias=qkv_bias)self.attn_drop = nn.Dropout(attn_drop, inplace=True)self.proj = nn.Linear(dim, dim)self.proj_drop = nn.Dropout(proj_drop, inplace=True)def forward(self, x):B, N, C = x.shapeqk = self.qk(x).reshape(B, N, 2, self.num_heads, self.head_dim).permute(2, 0, 3, 1, 4)q, k = qk[0], qk[1]   # make torchscript happy (cannot use tensor as tuple)v = self.v(x).reshape(B, N, self.num_heads, -1).permute(0, 2, 1, 3)attn = (q @ k.transpose(-2, -1)) * self.scaleattn = attn.softmax(dim=-1)attn = self.attn_drop(attn)x = (attn @ v).transpose(1, 2).reshape(B, N, -1)x = self.proj(x)x = self.proj_drop(x)return xclass Block(nn.Module):""" TNT Block"""def __init__(self, outer_dim, inner_dim, outer_num_heads, inner_num_heads, num_words, mlp_ratio=4.,qkv_bias=False, qk_scale=None, drop=0., attn_drop=0., drop_path=0., act_layer=nn.GELU,norm_layer=nn.LayerNorm, se=0):super().__init__()self.has_inner = inner_dim > 0if self.has_inner:# Innerself.inner_norm1 = norm_layer(inner_dim)self.inner_attn = Attention(inner_dim, inner_dim, num_heads=inner_num_heads, qkv_bias=qkv_bias,qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop)self.inner_norm2 = norm_layer(inner_dim)self.inner_mlp = Mlp(in_features=inner_dim, hidden_features=int(inner_dim * mlp_ratio),out_features=inner_dim, act_layer=act_layer, drop=drop)self.proj_norm1 = norm_layer(num_words * inner_dim)self.proj = nn.Linear(num_words * inner_dim, outer_dim, bias=False)self.proj_norm2 = norm_layer(outer_dim)# Outerself.outer_norm1 = norm_layer(outer_dim)self.outer_attn = Attention(outer_dim, outer_dim, num_heads=outer_num_heads, qkv_bias=qkv_bias,qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop)self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()self.outer_norm2 = norm_layer(outer_dim)self.outer_mlp = Mlp(in_features=outer_dim, hidden_features=int(outer_dim * mlp_ratio),out_features=outer_dim, act_layer=act_layer, drop=drop)# SEself.se = seself.se_layer = Noneif self.se > 0:self.se_layer = SE(outer_dim, 0.25)def forward(self, inner_tokens, outer_tokens):if self.has_inner:inner_tokens = inner_tokens + self.drop_path(self.inner_attn(self.inner_norm1(inner_tokens))) # B*N, k*k, cinner_tokens = inner_tokens + self.drop_path(self.inner_mlp(self.inner_norm2(inner_tokens))) # B*N, k*k, cB, N, C = outer_tokens.size()outer_tokens[:,1:] = outer_tokens[:,1:] + self.proj_norm2(self.proj(self.proj_norm1(inner_tokens.reshape(B, N-1, -1)))) # B, N, Cif self.se > 0:outer_tokens = outer_tokens + self.drop_path(self.outer_attn(self.outer_norm1(outer_tokens)))tmp_ = self.outer_mlp(self.outer_norm2(outer_tokens))outer_tokens = outer_tokens + self.drop_path(tmp_ + self.se_layer(tmp_))else:outer_tokens = outer_tokens + self.drop_path(self.outer_attn(self.outer_norm1(outer_tokens)))outer_tokens = outer_tokens + self.drop_path(self.outer_mlp(self.outer_norm2(outer_tokens)))return inner_tokens, outer_tokensclass PatchEmbed(nn.Module):""" Image to Visual Word Embedding"""def __init__(self, img_size=224, patch_size=16, in_chans=3, outer_dim=768, inner_dim=24, inner_stride=4):super().__init__()img_size = to_2tuple(img_size)patch_size = to_2tuple(patch_size)num_patches = (img_size[1] // patch_size[1]) * (img_size[0] // patch_size[0])self.img_size = img_sizeself.patch_size = patch_sizeself.num_patches = num_patchesself.inner_dim = inner_dimself.num_words = math.ceil(patch_size[0] / inner_stride) * math.ceil(patch_size[1] / inner_stride)self.unfold = nn.Unfold(kernel_size=patch_size, stride=patch_size)self.proj = nn.Conv2d(in_chans, inner_dim, kernel_size=7, padding=3, stride=inner_stride)def forward(self, x):B, C, H, W = x.shape# FIXME look at relaxing size constraintsassert 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]})."# self.unfold = Unfold(kernel_size=(16, 16), dilation=1, padding=0, stride=(16, 16))x = self.unfold(x) # B, Ck2, N [1, 768, 196] # 输出行数为卷积核横纵大小相乘，每列为每次卷积核卷过的元素x = x.transpose(1, 2).reshape(B * self.num_patches, C, *self.patch_size) # B*N, C, 16, 16 [196, 3, 16, 16]x = self.proj(x) # [196, 48, 4, 4]x = x.reshape(B * self.num_patches, self.inner_dim, -1).transpose(1, 2) # [196, 16, 48]return xclass TNT(nn.Module):""" TNT (Transformer in Transformer) for computer vision"""def __init__(self, img_size=224, patch_size=16, in_chans=3, num_classes=1000, outer_dim=768, inner_dim=48,depth=12, outer_num_heads=12, inner_num_heads=4, mlp_ratio=4., qkv_bias=False, qk_scale=None,drop_rate=0., attn_drop_rate=0., drop_path_rate=0., norm_layer=nn.LayerNorm, inner_stride=4, se=0):super().__init__()self.num_classes = num_classes # 1000self.num_features = self.outer_dim = outer_dim  # num_features for consistency with other modelsself.patch_embed = PatchEmbed(img_size=img_size, patch_size=patch_size, in_chans=in_chans, outer_dim=outer_dim,inner_dim=inner_dim, inner_stride=inner_stride)self.num_patches = num_patches = self.patch_embed.num_patches # 196num_words = self.patch_embed.num_words # 16self.proj_norm1 = norm_layer(num_words * inner_dim)     # LayerNorm((768,), eps=1e-05, elementwise_affine=True)self.proj = nn.Linear(num_words * inner_dim, outer_dim) # Linear(in_features=768, out_features=768, bias=True)self.proj_norm2 = norm_layer(outer_dim)                 # LayerNorm((768,), eps=1e-05, elementwise_affine=True)self.cls_token = nn.Parameter(torch.zeros(1, 1, outer_dim)) # [1, 1, 768]self.outer_tokens = nn.Parameter(torch.zeros(1, num_patches, outer_dim), requires_grad=False) # [1, 196, 768]self.outer_pos = nn.Parameter(torch.zeros(1, num_patches + 1, outer_dim)) # [1, 196, 768]self.inner_pos = nn.Parameter(torch.zeros(1, num_words, inner_dim))       # [1, 16, 48]self.pos_drop = nn.Dropout(p=drop_rate)dpr = [x.item() for x in torch.linspace(0, drop_path_rate, depth)]  # stochastic depth decay rulevanilla_idxs = []blocks = []for i in range(depth):if i in vanilla_idxs:blocks.append(Block(outer_dim=outer_dim, inner_dim=-1, outer_num_heads=outer_num_heads, inner_num_heads=inner_num_heads,num_words=num_words, mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale, drop=drop_rate,attn_drop=attn_drop_rate, drop_path=dpr[i], norm_layer=norm_layer, se=se))else:blocks.append(Block(outer_dim=outer_dim, inner_dim=inner_dim, outer_num_heads=outer_num_heads, inner_num_heads=inner_num_heads,num_words=num_words, mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale, drop=drop_rate,attn_drop=attn_drop_rate, drop_path=dpr[i], norm_layer=norm_layer, se=se))self.blocks = nn.ModuleList(blocks)self.norm = norm_layer(outer_dim)# NOTE as per official impl, we could have a pre-logits representation dense layer + tanh here#self.repr = nn.Linear(outer_dim, representation_size)#self.repr_act = nn.Tanh()# Classifier headself.head = nn.Linear(outer_dim, num_classes) if num_classes > 0 else nn.Identity() # Linear(in_features=768, out_features=1000, bias=True)trunc_normal_(self.cls_token, std=.02)trunc_normal_(self.outer_pos, std=.02)trunc_normal_(self.inner_pos, std=.02)self.apply(self._init_weights)# import pdb; pdb.set_trace()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)@torch.jit.ignoredef no_weight_decay(self):return {'outer_pos', 'inner_pos', 'cls_token'}def get_classifier(self):return self.headdef reset_classifier(self, num_classes, global_pool=''):self.num_classes = num_classesself.head = nn.Linear(self.outer_dim, num_classes) if num_classes > 0 else nn.Identity()def forward_features(self, x):# import pdb; pdb.set_trace()# x.shape=[1, 3, 224, 224]B = x.shape[0] # 1inner_tokens = self.patch_embed(x) + self.inner_pos # self.patch_embed(x) = [196, 16, 48], self.inner_pos=[1, 16, 48]outer_tokens = self.proj_norm2(self.proj(self.proj_norm1(inner_tokens.reshape(B, self.num_patches, -1))))     # [1, 196, 768]outer_tokens = torch.cat((self.cls_token.expand(B, -1, -1), outer_tokens), dim=1) # [1, 197, 768]outer_tokens = outer_tokens + self.outer_pos # [1, 197, 768]outer_tokens = self.pos_drop(outer_tokens)   # [1, 197, 768]for blk in self.blocks:inner_tokens, outer_tokens = blk(inner_tokens, outer_tokens) # inner_tokens.shape=[196, 16, 48]  outer_tokens.shape=[1, 197, 768]outer_tokens = self.norm(outer_tokens)return outer_tokens[:, 0] # [1, 768]def forward(self, x):x = self.forward_features(x) # [1, 768]x = self.head(x)return xdef _conv_filter(state_dict, patch_size=16):""" convert patch embedding weight from manual patchify + linear proj to conv"""out_dict = {}for k, v in state_dict.items():if 'patch_embed.proj.weight' in k:v = v.reshape((v.shape[0], 3, patch_size, patch_size))out_dict[k] = vreturn out_dict@register_model
def tnt_s_patch16_224(pretrained=False, **kwargs):patch_size = 16inner_stride = 4outer_dim = 384inner_dim = 24outer_num_heads = 6inner_num_heads = 4outer_dim = make_divisible(outer_dim, outer_num_heads)inner_dim = make_divisible(inner_dim, inner_num_heads)model = TNT(img_size=224, patch_size=patch_size, outer_dim=outer_dim, inner_dim=inner_dim, depth=12,outer_num_heads=outer_num_heads, inner_num_heads=inner_num_heads, qkv_bias=False,inner_stride=inner_stride, **kwargs)model.default_cfg = default_cfgs['tnt_s_patch16_224']if pretrained:load_pretrained(model, num_classes=model.num_classes, in_chans=kwargs.get('in_chans', 3), filter_fn=_conv_filter)return model@register_model
def tnt_b_patch16_224(pretrained=False, **kwargs):patch_size = 16inner_stride = 4outer_dim = 640inner_dim = 40outer_num_heads = 10inner_num_heads = 4outer_dim = make_divisible(outer_dim, outer_num_heads)inner_dim = make_divisible(inner_dim, inner_num_heads)model = TNT(img_size=224, patch_size=patch_size, outer_dim=outer_dim, inner_dim=inner_dim, depth=12,outer_num_heads=outer_num_heads, inner_num_heads=inner_num_heads, qkv_bias=False,inner_stride=inner_stride, **kwargs)model.default_cfg = default_cfgs['tnt_b_patch16_224']if pretrained:load_pretrained(model, num_classes=model.num_classes, in_chans=kwargs.get('in_chans', 3), filter_fn=_conv_filter)return model

self.blocks

ModuleList((0): Block((inner_norm1): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_attn): Attention((qk): Linear(in_features=48, out_features=96, bias=False)(v): Linear(in_features=48, out_features=48, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=48, out_features=48, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(inner_norm2): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_mlp): Mlp((fc1): Linear(in_features=48, out_features=192, bias=True)(act): GELU()(fc2): Linear(in_features=192, out_features=48, bias=True)(drop): Dropout(p=0.0, inplace=False))(proj_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(proj): Linear(in_features=768, out_features=768, bias=False)(proj_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_attn): Attention((qk): Linear(in_features=768, out_features=1536, bias=False)(v): Linear(in_features=768, out_features=768, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=768, out_features=768, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(drop_path): Identity()(outer_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_mlp): Mlp((fc1): Linear(in_features=768, out_features=3072, bias=True)(act): GELU()(fc2): Linear(in_features=3072, out_features=768, bias=True)(drop): Dropout(p=0.0, inplace=False)))(1): Block((inner_norm1): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_attn): Attention((qk): Linear(in_features=48, out_features=96, bias=False)(v): Linear(in_features=48, out_features=48, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=48, out_features=48, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(inner_norm2): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_mlp): Mlp((fc1): Linear(in_features=48, out_features=192, bias=True)(act): GELU()(fc2): Linear(in_features=192, out_features=48, bias=True)(drop): Dropout(p=0.0, inplace=False))(proj_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(proj): Linear(in_features=768, out_features=768, bias=False)(proj_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_attn): Attention((qk): Linear(in_features=768, out_features=1536, bias=False)(v): Linear(in_features=768, out_features=768, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=768, out_features=768, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(drop_path): Identity()(outer_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_mlp): Mlp((fc1): Linear(in_features=768, out_features=3072, bias=True)(act): GELU()(fc2): Linear(in_features=3072, out_features=768, bias=True)(drop): Dropout(p=0.0, inplace=False)))(2): Block((inner_norm1): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_attn): Attention((qk): Linear(in_features=48, out_features=96, bias=False)(v): Linear(in_features=48, out_features=48, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=48, out_features=48, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(inner_norm2): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_mlp): Mlp((fc1): Linear(in_features=48, out_features=192, bias=True)(act): GELU()(fc2): Linear(in_features=192, out_features=48, bias=True)(drop): Dropout(p=0.0, inplace=False))(proj_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(proj): Linear(in_features=768, out_features=768, bias=False)(proj_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_attn): Attention((qk): Linear(in_features=768, out_features=1536, bias=False)(v): Linear(in_features=768, out_features=768, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=768, out_features=768, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(drop_path): Identity()(outer_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_mlp): Mlp((fc1): Linear(in_features=768, out_features=3072, bias=True)(act): GELU()(fc2): Linear(in_features=3072, out_features=768, bias=True)(drop): Dropout(p=0.0, inplace=False)))(3): Block((inner_norm1): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_attn): Attention((qk): Linear(in_features=48, out_features=96, bias=False)(v): Linear(in_features=48, out_features=48, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=48, out_features=48, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(inner_norm2): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_mlp): Mlp((fc1): Linear(in_features=48, out_features=192, bias=True)(act): GELU()(fc2): Linear(in_features=192, out_features=48, bias=True)(drop): Dropout(p=0.0, inplace=False))(proj_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(proj): Linear(in_features=768, out_features=768, bias=False)(proj_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_attn): Attention((qk): Linear(in_features=768, out_features=1536, bias=False)(v): Linear(in_features=768, out_features=768, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=768, out_features=768, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(drop_path): Identity()(outer_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_mlp): Mlp((fc1): Linear(in_features=768, out_features=3072, bias=True)(act): GELU()(fc2): Linear(in_features=3072, out_features=768, bias=True)(drop): Dropout(p=0.0, inplace=False)))(4): Block((inner_norm1): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_attn): Attention((qk): Linear(in_features=48, out_features=96, bias=False)(v): Linear(in_features=48, out_features=48, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=48, out_features=48, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(inner_norm2): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_mlp): Mlp((fc1): Linear(in_features=48, out_features=192, bias=True)(act): GELU()(fc2): Linear(in_features=192, out_features=48, bias=True)(drop): Dropout(p=0.0, inplace=False))(proj_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(proj): Linear(in_features=768, out_features=768, bias=False)(proj_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_attn): Attention((qk): Linear(in_features=768, out_features=1536, bias=False)(v): Linear(in_features=768, out_features=768, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=768, out_features=768, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(drop_path): Identity()(outer_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_mlp): Mlp((fc1): Linear(in_features=768, out_features=3072, bias=True)(act): GELU()(fc2): Linear(in_features=3072, out_features=768, bias=True)(drop): Dropout(p=0.0, inplace=False)))(5): Block((inner_norm1): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_attn): Attention((qk): Linear(in_features=48, out_features=96, bias=False)(v): Linear(in_features=48, out_features=48, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=48, out_features=48, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(inner_norm2): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_mlp): Mlp((fc1): Linear(in_features=48, out_features=192, bias=True)(act): GELU()(fc2): Linear(in_features=192, out_features=48, bias=True)(drop): Dropout(p=0.0, inplace=False))(proj_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(proj): Linear(in_features=768, out_features=768, bias=False)(proj_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_attn): Attention((qk): Linear(in_features=768, out_features=1536, bias=False)(v): Linear(in_features=768, out_features=768, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=768, out_features=768, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(drop_path): Identity()(outer_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_mlp): Mlp((fc1): Linear(in_features=768, out_features=3072, bias=True)(act): GELU()(fc2): Linear(in_features=3072, out_features=768, bias=True)(drop): Dropout(p=0.0, inplace=False)))(6): Block((inner_norm1): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_attn): Attention((qk): Linear(in_features=48, out_features=96, bias=False)(v): Linear(in_features=48, out_features=48, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=48, out_features=48, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(inner_norm2): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_mlp): Mlp((fc1): Linear(in_features=48, out_features=192, bias=True)(act): GELU()(fc2): Linear(in_features=192, out_features=48, bias=True)(drop): Dropout(p=0.0, inplace=False))(proj_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(proj): Linear(in_features=768, out_features=768, bias=False)(proj_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_attn): Attention((qk): Linear(in_features=768, out_features=1536, bias=False)(v): Linear(in_features=768, out_features=768, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=768, out_features=768, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(drop_path): Identity()(outer_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_mlp): Mlp((fc1): Linear(in_features=768, out_features=3072, bias=True)(act): GELU()(fc2): Linear(in_features=3072, out_features=768, bias=True)(drop): Dropout(p=0.0, inplace=False)))(7): Block((inner_norm1): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_attn): Attention((qk): Linear(in_features=48, out_features=96, bias=False)(v): Linear(in_features=48, out_features=48, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=48, out_features=48, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(inner_norm2): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_mlp): Mlp((fc1): Linear(in_features=48, out_features=192, bias=True)(act): GELU()(fc2): Linear(in_features=192, out_features=48, bias=True)(drop): Dropout(p=0.0, inplace=False))(proj_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(proj): Linear(in_features=768, out_features=768, bias=False)(proj_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_attn): Attention((qk): Linear(in_features=768, out_features=1536, bias=False)(v): Linear(in_features=768, out_features=768, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=768, out_features=768, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(drop_path): Identity()(outer_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_mlp): Mlp((fc1): Linear(in_features=768, out_features=3072, bias=True)(act): GELU()(fc2): Linear(in_features=3072, out_features=768, bias=True)(drop): Dropout(p=0.0, inplace=False)))(8): Block((inner_norm1): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_attn): Attention((qk): Linear(in_features=48, out_features=96, bias=False)(v): Linear(in_features=48, out_features=48, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=48, out_features=48, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(inner_norm2): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_mlp): Mlp((fc1): Linear(in_features=48, out_features=192, bias=True)(act): GELU()(fc2): Linear(in_features=192, out_features=48, bias=True)(drop): Dropout(p=0.0, inplace=False))(proj_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(proj): Linear(in_features=768, out_features=768, bias=False)(proj_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_attn): Attention((qk): Linear(in_features=768, out_features=1536, bias=False)(v): Linear(in_features=768, out_features=768, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=768, out_features=768, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(drop_path): Identity()(outer_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_mlp): Mlp((fc1): Linear(in_features=768, out_features=3072, bias=True)(act): GELU()(fc2): Linear(in_features=3072, out_features=768, bias=True)(drop): Dropout(p=0.0, inplace=False)))(9): Block((inner_norm1): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_attn): Attention((qk): Linear(in_features=48, out_features=96, bias=False)(v): Linear(in_features=48, out_features=48, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=48, out_features=48, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(inner_norm2): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_mlp): Mlp((fc1): Linear(in_features=48, out_features=192, bias=True)(act): GELU()(fc2): Linear(in_features=192, out_features=48, bias=True)(drop): Dropout(p=0.0, inplace=False))(proj_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(proj): Linear(in_features=768, out_features=768, bias=False)(proj_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_attn): Attention((qk): Linear(in_features=768, out_features=1536, bias=False)(v): Linear(in_features=768, out_features=768, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=768, out_features=768, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(drop_path): Identity()(outer_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_mlp): Mlp((fc1): Linear(in_features=768, out_features=3072, bias=True)(act): GELU()(fc2): Linear(in_features=3072, out_features=768, bias=True)(drop): Dropout(p=0.0, inplace=False)))(10): Block((inner_norm1): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_attn): Attention((qk): Linear(in_features=48, out_features=96, bias=False)(v): Linear(in_features=48, out_features=48, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=48, out_features=48, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(inner_norm2): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_mlp): Mlp((fc1): Linear(in_features=48, out_features=192, bias=True)(act): GELU()(fc2): Linear(in_features=192, out_features=48, bias=True)(drop): Dropout(p=0.0, inplace=False))(proj_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(proj): Linear(in_features=768, out_features=768, bias=False)(proj_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_attn): Attention((qk): Linear(in_features=768, out_features=1536, bias=False)(v): Linear(in_features=768, out_features=768, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=768, out_features=768, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(drop_path): Identity()(outer_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_mlp): Mlp((fc1): Linear(in_features=768, out_features=3072, bias=True)(act): GELU()(fc2): Linear(in_features=3072, out_features=768, bias=True)(drop): Dropout(p=0.0, inplace=False)))(11): Block((inner_norm1): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_attn): Attention((qk): Linear(in_features=48, out_features=96, bias=False)(v): Linear(in_features=48, out_features=48, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=48, out_features=48, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(inner_norm2): LayerNorm((48,), eps=1e-05, elementwise_affine=True)(inner_mlp): Mlp((fc1): Linear(in_features=48, out_features=192, bias=True)(act): GELU()(fc2): Linear(in_features=192, out_features=48, bias=True)(drop): Dropout(p=0.0, inplace=False))(proj_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(proj): Linear(in_features=768, out_features=768, bias=False)(proj_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_norm1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_attn): Attention((qk): Linear(in_features=768, out_features=1536, bias=False)(v): Linear(in_features=768, out_features=768, bias=False)(attn_drop): Dropout(p=0.0, inplace=True)(proj): Linear(in_features=768, out_features=768, bias=True)(proj_drop): Dropout(p=0.0, inplace=True))(drop_path): Identity()(outer_norm2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)(outer_mlp): Mlp((fc1): Linear(in_features=768, out_features=3072, bias=True)(act): GELU()(fc2): Linear(in_features=3072, out_features=768, bias=True)(drop): Dropout(p=0.0, inplace=False)))
)