vision_transformer.py

"""
Mostly copy-paste from timm library.
https://github.com/rwightman/pytorch-image-models/blob/master/timm/models/vision_transformer.py
"""
import math
from functools import partial

import torch
import torch.nn as nn
from einops import rearrange
from functools import partial, reduce
from operator import mul

from utils import trunc_normal_


def drop_path(x, drop_prob: float = 0., training: bool = False):
    if drop_prob == 0. or not training:
        return x
    keep_prob = 1 - drop_prob
    shape = (x.shape[0],) + (1,) * (x.ndim - 1)  # work with diff dim tensors, not just 2D ConvNets
    random_tensor = keep_prob + torch.rand(shape, dtype=x.dtype, device=x.device)
    random_tensor.floor_()  # binarize
    output = x.div(keep_prob) * random_tensor
    return output


class DropPath(nn.Module):
    """Drop paths (Stochastic Depth) per sample  (when applied in main path of residual blocks).
    """
    def __init__(self, drop_prob=None):
        super(DropPath, self).__init__()
        self.drop_prob = drop_prob

    def forward(self, x):
        return drop_path(x, self.drop_prob, self.training)


class 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_features
        hidden_features = hidden_features or in_features
        self.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 x


class Attention(nn.Module):
    def __init__(self, dim, num_heads=8, qkv_bias=False, qk_scale=None, attn_drop=0., proj_drop=0.):
        super().__init__()
        self.num_heads = num_heads
        head_dim = dim // num_heads
        self.scale = qk_scale or head_dim ** -0.5

        self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)
        self.attn_drop = nn.Dropout(attn_drop)
        self.proj = nn.Linear(dim, dim)
        self.proj_drop = nn.Dropout(proj_drop)

    def forward(self, x):
        B, N, C = x.shape
        qkv = self.qkv(x).reshape(B, N, 3, self.num_heads, C // self.num_heads).permute(2, 0, 3, 1, 4)
        q, k, v = qkv[0], qkv[1], qkv[2]

        attn = (q @ k.transpose(-2, -1)) * self.scale
        attn = attn.softmax(dim=-1)
        attn = self.attn_drop(attn)

        x = (attn @ v).transpose(1, 2).reshape(B, N, C)
        x = self.proj(x)
        x = self.proj_drop(x)
        return x, attn


class Block(nn.Module):
    def __init__(self, dim, num_heads, 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):
        super().__init__()
        self.norm1 = norm_layer(dim)
        self.attn = Attention(
            dim, num_heads=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.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)

    def forward(self, x, return_attention=False):
        y, attn = self.attn(self.norm1(x))
        if return_attention:
            return attn
        x = x + self.drop_path(y)
        x = x + self.drop_path(self.mlp(self.norm2(x)))
        return x


class PatchEmbed(nn.Module):
    """ Image to Patch Embedding
    """
    def __init__(self, img_size=224, patch_size=16, in_chans=3, embed_dim=768):
        super().__init__()
        num_patches = (img_size // patch_size) * (img_size // patch_size)
        self.img_size = img_size
        self.patch_size = (patch_size, patch_size)
        self.grid_size = (img_size/self.patch_size[0], img_size/self.patch_size[1])
        self.num_patches = num_patches

        self.proj = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=patch_size)

    def forward(self, x):
        B, C, H, W = x.shape
        x = self.proj(x).flatten(2).transpose(1, 2)
        return x


class VisionTransformer(nn.Module):
    """ Vision Transformer """
    def __init__(self, img_size=[224], patch_size=16, in_chans=3, num_classes=0, embed_dim=768, depth=12,
                 num_heads=12, mlp_ratio=4., qkv_bias=False, qk_scale=None, drop_rate=0., attn_drop_rate=0.,
                 drop_path_rate=0., norm_layer=nn.LayerNorm, dis_token=False, **kwargs):
        super().__init__()
        self.num_features = self.embed_dim = embed_dim

        self.patch_embed = PatchEmbed(
            img_size=img_size[0], patch_size=patch_size, in_chans=in_chans, embed_dim=embed_dim)
        num_patches = self.patch_embed.num_patches

        self.cls_token = nn.Parameter(torch.zeros(1, 1, embed_dim))
        self.dis_token = None
        if dis_token:
            self.dis_token = nn.Parameter(torch.zeros(1, 1, embed_dim))
            nn.init.normal_(self.dis_token, std=1e-6)
            print("Use distillation token!!")
        self.build_2d_sincos_position_embedding()

        self.pos_drop = nn.Dropout(p=drop_rate)

        dpr = [x.item() for x in torch.linspace(0, drop_path_rate, depth)]  # stochastic depth decay rule
        self.blocks = nn.ModuleList([
            Block(
                dim=embed_dim, num_heads=num_heads, 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)
            for i in range(depth)])
        self.norm = norm_layer(embed_dim)

        # Classifier head
        self.head = nn.Linear(embed_dim, num_classes) if num_classes > 0 else nn.Identity()

        # weight initialization
        for name, m in self.named_modules():
            if isinstance(m, nn.Linear):
                if 'qkv' in name:
                    # treat the weights of Q, K, V separately
                    val = math.sqrt(6. / float(m.weight.shape[0] // 3 + m.weight.shape[1]))
                    nn.init.uniform_(m.weight, -val, val)
                else:
                    nn.init.xavier_uniform_(m.weight)
                nn.init.zeros_(m.bias)
            elif isinstance(m, (nn.LayerNorm, nn.GroupNorm, nn.BatchNorm2d)):
                nn.init.zeros_(m.bias)
                nn.init.ones_(m.weight)
        nn.init.normal_(self.cls_token, std=1e-6)

        if isinstance(self.patch_embed, PatchEmbed):
            # xavier_uniform initialization
            val = math.sqrt(6. / float(3 * reduce(mul, self.patch_embed.patch_size, 1) + self.embed_dim))
            nn.init.uniform_(self.patch_embed.proj.weight, -val, val)
            nn.init.zeros_(self.patch_embed.proj.bias)

            self.patch_embed.proj.weight.requires_grad = False
            self.patch_embed.proj.bias.requires_grad = False


    def build_2d_sincos_position_embedding(self, temperature=10000.):
        h, w = self.patch_embed.grid_size
        grid_w = torch.arange(w, dtype=torch.float32)
        grid_h = torch.arange(h, dtype=torch.float32)
        grid_w, grid_h = torch.meshgrid(grid_w, grid_h)
        assert self.embed_dim % 4 == 0, 'Embed dimension must be divisible by 4 for 2D sin-cos position embedding'
        pos_dim = self.embed_dim // 4
        omega = torch.arange(pos_dim, dtype=torch.float32) / pos_dim
        omega = 1. / (temperature ** omega)
        out_w = torch.einsum('m,d->md', [grid_w.flatten(), omega])
        out_h = torch.einsum('m,d->md', [grid_h.flatten(), omega])
        pos_emb = torch.cat([torch.sin(out_w), torch.cos(out_w), torch.sin(out_h), torch.cos(out_h)], dim=1)[None, :, :]

        # assert self.num_tokens == 1, 'Assuming one and only one token, [cls]'
        num_token = 2 if self.dis_token is not None else 1
        pe_token = torch.zeros([1, num_token, self.embed_dim], dtype=torch.float32)
        self.pos_embed = nn.Parameter(torch.cat([pe_token, pos_emb], dim=1))
        # if self.dis_token is not None:
        #     di_token = torch.zeros([1, 1, self.embed_dim], dtype=torch.float32)
        #     self.pos_embed = nn.Parameter(torch.cat([pe_token, pos_emb, di_token], dim=1))
        self.pos_embed.requires_grad = False

    def prepare_tokens(self, x, dino=False):
        B, nc, w, h = x.shape
        x = self.patch_embed(x)  # patch linear embedding
        if dino is False:
            x = x.detach()

        # add the [CLS] token to the embed patch tokens
        cls_tokens = self.cls_token.expand(B, -1, -1)
        if self.dis_token is not None:
            dis_tokens = self.dis_token.expand(B, -1, -1)
            x = torch.cat((cls_tokens, dis_tokens, x), dim=1)
        else:
            x = torch.cat((cls_tokens, x), dim=1)

        # add positional encoding to each token
        x = x + self.pos_embed

        return self.pos_drop(x)

    def forward(self, x, dino=False):
        x = self.prepare_tokens(x, dino)
        for blk in self.blocks:
            x = blk(x)
        x = self.norm(x)
        return x[:, 0]

    def get_last_selfattention(self, x):
        x = self.prepare_tokens(x)
        for i, blk in enumerate(self.blocks):
            if i < len(self.blocks) - 1:
                x = blk(x)
            else:
                # return attention of the last block
                return blk(x, return_attention=True)

    def get_intermediate_layers(self, x, n=1, dino=False):
        x = self.prepare_tokens(x, dino)
        # we return the output tokens from the `n` last blocks
        output = []
        for i, blk in enumerate(self.blocks):
            x = blk(x)
            if len(self.blocks) - i <= n:
                if i != 11 and self.dis_token is not None:
                    output.append(self.norm(x)[:, 1])
                else:
                    output.append(self.norm(x)[:, 0])

        return torch.cat(output, dim=0)


def vit_tiny(patch_size=16, **kwargs):
    model = VisionTransformer(
        patch_size=patch_size, embed_dim=192, depth=12, num_heads=6, mlp_ratio=4,
        qkv_bias=True, norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs)
    return model


def vit_small(patch_size=16, **kwargs):
    model = VisionTransformer(
        patch_size=patch_size, embed_dim=384, depth=12, num_heads=12, mlp_ratio=4,
        qkv_bias=True, norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs)
    return model


def vit_base(patch_size=16, **kwargs):
    model = VisionTransformer(
        patch_size=patch_size, embed_dim=768, depth=12, num_heads=12, mlp_ratio=4,
        qkv_bias=True, norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs)
    return model