ops.py

import torch
import torch.nn as nn
from torch.nn import Module, ModuleList, Linear, Dropout, LayerNorm, Identity, Parameter, init
import torch.nn.functional as F
from stochastic_depth import DropPath


class DownsampleLayer(nn.Module):
    def __init__(self, in_ch, out_ch):
        super(DownsampleLayer, self).__init__()
        self.Conv_BN_ReLU_2 = nn.Sequential(
            nn.Conv2d(in_channels=in_ch, out_channels=out_ch, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(out_ch),
            nn.ReLU(),
            nn.Conv2d(in_channels=out_ch, out_channels=out_ch, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(out_ch),
            nn.ReLU()
        )
        self.downsample=nn.Sequential(
            nn.Conv2d(in_channels=out_ch, out_channels=out_ch, kernel_size=3, stride=2, padding=1),
            nn.BatchNorm2d(out_ch),
            nn.ReLU()
        )

    def forward(self, x):
        out = self.Conv_BN_ReLU_2(x)
        out_2 = self.downsample(out)
        return out, out_2


class UpSampleLayer(nn.Module):
    def __init__(self, in_ch, out_ch):
        super(UpSampleLayer, self).__init__()
        self.Conv_BN_ReLU_2 = nn.Sequential(
            nn.Conv2d(in_channels=in_ch, out_channels=out_ch*2, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(out_ch*2),
            nn.ReLU(),
            nn.Conv2d(in_channels=out_ch*2, out_channels=out_ch*2, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(out_ch*2),
            nn.ReLU()
        )
        self.upsample=nn.Sequential(
            nn.ConvTranspose2d(in_channels=out_ch*2, out_channels=out_ch, kernel_size=3, stride=2,
                               padding=1, output_padding=1),
            nn.BatchNorm2d(out_ch),
            nn.ReLU()
        )

    def forward(self, x, out):
        x_out = self.Conv_BN_ReLU_2(x)
        x_out = self.upsample(x_out)
        cat_out = torch.cat((x_out, out), dim=1)
        return cat_out


class SE(Module):
    def __init__(self, in_chnls, ratio):
        super(SE, self).__init__()
        self.squeeze = nn.AdaptiveAvgPool2d((1, 1))
        self.compress = nn.Conv2d(in_chnls, in_chnls // ratio, 1, 1, 0)
        self.excitation = nn.Conv2d(in_chnls // ratio, in_chnls, 1, 1, 0)

    def forward(self, x):
        out = self.squeeze(x)
        out = self.compress(out)
        out = F.relu(out)
        out = self.excitation(out)
        return F.sigmoid(out)


class BasicBlock(nn.Module):
    def __init__(self, in_channels, out_channels, mode=0):
        super(BasicBlock, self).__init__()
        self.se = SE(in_channels, 4)
        self.bn1 = nn.BatchNorm2d(in_channels)
        self.conv1 = nn.Conv2d(in_channels, in_channels // 4, 1, stride=1, bias=False)
        self.bn2 = nn.BatchNorm2d(in_channels // 4)
        if mode == 0:
            MidChannels = in_channels // 4
        elif mode == 1:
            MidChannels = in_channels // 2
        else:
            MidChannels = in_channels // 4
        if mode == 0:
            self.conv2 = nn.Conv2d(in_channels // 4, MidChannels, 3, stride=2, padding=1, bias=False)
        elif mode == 1:
            self.conv2 = nn.ConvTranspose2d(in_channels // 4, MidChannels, kernel_size=2, stride=2, bias=False)
        else:
            self.conv2 = nn.Conv2d(in_channels // 4, MidChannels, 3, stride=1, padding=1, bias=False)
        self.bn3 = nn.BatchNorm2d(MidChannels)
        self.conv3 = nn.Conv2d(MidChannels, out_channels, 1, stride=1, bias=False)

    def forward(self, x):
        coefficient = self.se(x)
        x = x * coefficient
        out = self.bn1(x)
        out = F.relu(out)
        out = self.conv1(out)
        out = self.bn2(out)
        out = F.relu(out)
        out = self.conv2(out)
        out = self.bn3(out)
        out = F.relu(out)
        out = self.conv3(out)
        return F.relu(out)


class Attention(Module):
    """
    Obtained from timm: github.com:rwightman/pytorch-image-models
    """

    def __init__(self, dim, num_heads=8, attention_dropout=0.1, projection_dropout=0.1):
        super().__init__()
        self.num_heads = num_heads
        head_dim = dim // self.num_heads
        self.scale = head_dim ** -0.5

        self.q = Linear(dim, dim, bias=False)
        self.kv = Linear(dim, dim * 2, bias=False)
        self.attn_drop = Dropout(attention_dropout)
        self.proj = Linear(dim, dim)
        self.proj_drop = Dropout(projection_dropout)

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

        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


class TransformerEncoderLayer(Module):
    """
    Inspired by torch.nn.TransformerEncoderLayer and timm.
    """

    def __init__(self, d_model, nhead, dim_feedforward=2048, dropout=0.1,
                 attention_dropout=0.1, drop_path_rate=0.1, out_dim=None):
        super(TransformerEncoderLayer, self).__init__()
        if out_dim is None:
            out_dim = d_model
        self.q_norm = LayerNorm(d_model)
        self.kv_norm = LayerNorm(d_model)
        self.self_attn = Attention(dim=d_model, num_heads=nhead,
                                   attention_dropout=attention_dropout, projection_dropout=dropout)

        self.linear1 = Linear(d_model, dim_feedforward)
        self.dropout1 = Dropout(dropout)
        self.norm1 = LayerNorm(d_model)
        self.linear2 = Linear(dim_feedforward, out_dim)
        self.dropout2 = Dropout(dropout)

        self.drop_path = DropPath(drop_path_rate) if drop_path_rate > 0 else Identity()

        self.activation = F.gelu

    def forward(self, src_q: torch.Tensor, src_kv, *args, **kwargs) -> torch.Tensor:
        src = src_q + self.drop_path(self.self_attn(self.q_norm(src_q), self.kv_norm(src_kv)))
        src = self.norm1(src)
        src2 = self.linear2(self.dropout1(self.activation(self.linear1(src))))
        src = src + self.drop_path(self.dropout2(src2))
        return src


class TransformerClassifier(Module):
    def __init__(self,
                 seq_pool=True,
                 embedding_dim=768,
                 num_layers=12,
                 num_heads=12,
                 mlp_ratio=4.0,
                 num_classes=1000,
                 dropout=0.1,
                 attention_dropout=0.1,
                 stochastic_depth=0.1,
                 positional_embedding='learnable',
                 sequence_length=None):
        super().__init__()
        positional_embedding = positional_embedding if \
            positional_embedding in ['sine', 'learnable', 'none'] else 'sine'
        dim_feedforward = int(embedding_dim * mlp_ratio)
        self.embedding_dim = embedding_dim
        self.sequence_length = sequence_length
        self.seq_pool = seq_pool
        self.num_tokens = 0

        assert sequence_length is not None or positional_embedding == 'none', \
            f"Positional embedding is set to {positional_embedding} and" \
            f" the sequence length was not specified."

        # print("seq ", seq_pool, 'emb_dim ', embedding_dim, 'pos_emb ', positional_embedding)

        if not seq_pool:
            sequence_length += 1
            self.class_emb = Parameter(torch.zeros(1, 1, self.embedding_dim),
                                       requires_grad=True)
            self.num_tokens = 1
        else:
            self.attention_pool = Linear(self.embedding_dim, 1)

        if positional_embedding != 'none':
            if positional_embedding == 'learnable':
                self.positional_emb = Parameter(torch.zeros(1, sequence_length, embedding_dim),
                                                requires_grad=True)
                init.trunc_normal_(self.positional_emb, std=0.2)
            else:
                self.positional_emb = Parameter(self.sinusoidal_embedding(sequence_length, embedding_dim),
                                                requires_grad=False)
        else:
            self.positional_emb = None

        self.dropout = Dropout(p=dropout)
        dpr = [x.item() for x in torch.linspace(0, stochastic_depth, num_layers)]
        self.blocks = ModuleList([
            TransformerEncoderLayer(d_model=embedding_dim, nhead=num_heads,
                                    dim_feedforward=dim_feedforward, dropout=dropout,
                                    attention_dropout=attention_dropout, drop_path_rate=dpr[i])
            for i in range(num_layers)])
        self.norm = LayerNorm(embedding_dim)

        self.fc = Linear(embedding_dim, num_classes)
        self.apply(self.init_weight)

    def forward(self, x):
        # if self.positional_emb is None and x.size(1) < self.sequence_length:
        #     x = F.pad(x, (0, 0, 0, self.n_channels - x.size(1)), mode='constant', value=0)

        # if not self.seq_pool:
        #     cls_token = self.class_emb.expand(x.shape[0], -1, -1)
        #     x = torch.cat((cls_token, x), dim=1)

        if self.positional_emb is not None:
            x = x + self.positional_emb

        x = self.dropout(x)

        for blk in self.blocks:
            x = blk(x)
        x_seq = self.norm(x)

        # print('before seq pool ', x.shape)
        # if self.seq_pool:
        x = torch.matmul(F.softmax(self.attention_pool(x_seq), dim=1).transpose(-1, -2), x).squeeze(-2)
        # else:
        #     x = x_seq[:, 0]
        # print('aften seq pool ', x.shape)

        x = self.fc(x)
        return x, x_seq

    @staticmethod
    def init_weight(m):
        if isinstance(m, Linear):
            init.trunc_normal_(m.weight, std=.02)
            if isinstance(m, Linear) and m.bias is not None:
                init.constant_(m.bias, 0)
        elif isinstance(m, LayerNorm):
            init.constant_(m.bias, 0)
            init.constant_(m.weight, 1.0)

    @staticmethod
    def sinusoidal_embedding(n_channels, dim):
        pe = torch.FloatTensor([[p / (10000 ** (2 * (i // 2) / dim)) for i in range(dim)]
                                for p in range(n_channels)])
        pe[:, 0::2] = torch.sin(pe[:, 0::2])
        pe[:, 1::2] = torch.cos(pe[:, 1::2])
        return pe.unsqueeze(0)