restoreformer_arch.py

"""Modified from https://github.com/wzhouxiff/RestoreFormer
"""
import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F


class VectorQuantizer(nn.Module):
    """
    see https://github.com/MishaLaskin/vqvae/blob/d761a999e2267766400dc646d82d3ac3657771d4/models/quantizer.py
    ____________________________________________
    Discretization bottleneck part of the VQ-VAE.
    Inputs:
    - n_e : number of embeddings
    - e_dim : dimension of embedding
    - beta : commitment cost used in loss term, beta * ||z_e(x)-sg[e]||^2
    _____________________________________________
    """

    def __init__(self, n_e, e_dim, beta):
        super(VectorQuantizer, self).__init__()
        self.n_e = n_e
        self.e_dim = e_dim
        self.beta = beta

        self.embedding = nn.Embedding(self.n_e, self.e_dim)
        self.embedding.weight.data.uniform_(-1.0 / self.n_e, 1.0 / self.n_e)

    def forward(self, z):
        """
        Inputs the output of the encoder network z and maps it to a discrete
        one-hot vector that is the index of the closest embedding vector e_j
        z (continuous) -> z_q (discrete)
        z.shape = (batch, channel, height, width)
        quantization pipeline:
            1. get encoder input (B,C,H,W)
            2. flatten input to (B*H*W,C)
        """
        # reshape z -> (batch, height, width, channel) and flatten
        z = z.permute(0, 2, 3, 1).contiguous()
        z_flattened = z.view(-1, self.e_dim)
        # distances from z to embeddings e_j (z - e)^2 = z^2 + e^2 - 2 e * z

        d = torch.sum(z_flattened ** 2, dim=1, keepdim=True) + \
            torch.sum(self.embedding.weight**2, dim=1) - 2 * \
            torch.matmul(z_flattened, self.embedding.weight.t())

        # could possible replace this here
        # #\start...
        # find closest encodings

        min_value, min_encoding_indices = torch.min(d, dim=1)

        min_encoding_indices = min_encoding_indices.unsqueeze(1)

        min_encodings = torch.zeros(min_encoding_indices.shape[0], self.n_e).to(z)
        min_encodings.scatter_(1, min_encoding_indices, 1)

        # dtype min encodings: torch.float32
        # min_encodings shape: torch.Size([2048, 512])
        # min_encoding_indices.shape: torch.Size([2048, 1])

        # get quantized latent vectors
        z_q = torch.matmul(min_encodings, self.embedding.weight).view(z.shape)
        # .........\end

        # with:
        # .........\start
        # min_encoding_indices = torch.argmin(d, dim=1)
        # z_q = self.embedding(min_encoding_indices)
        # ......\end......... (TODO)

        # compute loss for embedding
        loss = torch.mean((z_q.detach() - z)**2) + self.beta * torch.mean((z_q - z.detach())**2)

        # preserve gradients
        z_q = z + (z_q - z).detach()

        # perplexity

        e_mean = torch.mean(min_encodings, dim=0)
        perplexity = torch.exp(-torch.sum(e_mean * torch.log(e_mean + 1e-10)))

        # reshape back to match original input shape
        z_q = z_q.permute(0, 3, 1, 2).contiguous()

        return z_q, loss, (perplexity, min_encodings, min_encoding_indices, d)

    def get_codebook_entry(self, indices, shape):
        # shape specifying (batch, height, width, channel)
        # TODO: check for more easy handling with nn.Embedding
        min_encodings = torch.zeros(indices.shape[0], self.n_e).to(indices)
        min_encodings.scatter_(1, indices[:, None], 1)

        # get quantized latent vectors
        z_q = torch.matmul(min_encodings.float(), self.embedding.weight)

        if shape is not None:
            z_q = z_q.view(shape)

            # reshape back to match original input shape
            z_q = z_q.permute(0, 3, 1, 2).contiguous()

        return z_q


# pytorch_diffusion + derived encoder decoder
def nonlinearity(x):
    # swish
    return x * torch.sigmoid(x)


def Normalize(in_channels):
    return torch.nn.GroupNorm(num_groups=32, num_channels=in_channels, eps=1e-6, affine=True)


class Upsample(nn.Module):

    def __init__(self, in_channels, with_conv):
        super().__init__()
        self.with_conv = with_conv
        if self.with_conv:
            self.conv = torch.nn.Conv2d(in_channels, in_channels, kernel_size=3, stride=1, padding=1)

    def forward(self, x):
        x = torch.nn.functional.interpolate(x, scale_factor=2.0, mode='nearest')
        if self.with_conv:
            x = self.conv(x)
        return x


class Downsample(nn.Module):

    def __init__(self, in_channels, with_conv):
        super().__init__()
        self.with_conv = with_conv
        if self.with_conv:
            # no asymmetric padding in torch conv, must do it ourselves
            self.conv = torch.nn.Conv2d(in_channels, in_channels, kernel_size=3, stride=2, padding=0)

    def forward(self, x):
        if self.with_conv:
            pad = (0, 1, 0, 1)
            x = torch.nn.functional.pad(x, pad, mode='constant', value=0)
            x = self.conv(x)
        else:
            x = torch.nn.functional.avg_pool2d(x, kernel_size=2, stride=2)
        return x


class ResnetBlock(nn.Module):

    def __init__(self, *, in_channels, out_channels=None, conv_shortcut=False, dropout, temb_channels=512):
        super().__init__()
        self.in_channels = in_channels
        out_channels = in_channels if out_channels is None else out_channels
        self.out_channels = out_channels
        self.use_conv_shortcut = conv_shortcut

        self.norm1 = Normalize(in_channels)
        self.conv1 = torch.nn.Conv2d(in_channels, out_channels, kernel_size=3, stride=1, padding=1)
        if temb_channels > 0:
            self.temb_proj = torch.nn.Linear(temb_channels, out_channels)
        self.norm2 = Normalize(out_channels)
        self.dropout = torch.nn.Dropout(dropout)
        self.conv2 = torch.nn.Conv2d(out_channels, out_channels, kernel_size=3, stride=1, padding=1)
        if self.in_channels != self.out_channels:
            if self.use_conv_shortcut:
                self.conv_shortcut = torch.nn.Conv2d(in_channels, out_channels, kernel_size=3, stride=1, padding=1)
            else:
                self.nin_shortcut = torch.nn.Conv2d(in_channels, out_channels, kernel_size=1, stride=1, padding=0)

    def forward(self, x, temb):
        h = x
        h = self.norm1(h)
        h = nonlinearity(h)
        h = self.conv1(h)

        if temb is not None:
            h = h + self.temb_proj(nonlinearity(temb))[:, :, None, None]

        h = self.norm2(h)
        h = nonlinearity(h)
        h = self.dropout(h)
        h = self.conv2(h)

        if self.in_channels != self.out_channels:
            if self.use_conv_shortcut:
                x = self.conv_shortcut(x)
            else:
                x = self.nin_shortcut(x)

        return x + h


class MultiHeadAttnBlock(nn.Module):

    def __init__(self, in_channels, head_size=1):
        super().__init__()
        self.in_channels = in_channels
        self.head_size = head_size
        self.att_size = in_channels // head_size
        assert (in_channels % head_size == 0), 'The size of head should be divided by the number of channels.'

        self.norm1 = Normalize(in_channels)
        self.norm2 = Normalize(in_channels)

        self.q = torch.nn.Conv2d(in_channels, in_channels, kernel_size=1, stride=1, padding=0)
        self.k = torch.nn.Conv2d(in_channels, in_channels, kernel_size=1, stride=1, padding=0)
        self.v = torch.nn.Conv2d(in_channels, in_channels, kernel_size=1, stride=1, padding=0)
        self.proj_out = torch.nn.Conv2d(in_channels, in_channels, kernel_size=1, stride=1, padding=0)
        self.num = 0

    def forward(self, x, y=None):
        h_ = x
        h_ = self.norm1(h_)
        if y is None:
            y = h_
        else:
            y = self.norm2(y)

        q = self.q(y)
        k = self.k(h_)
        v = self.v(h_)

        # compute attention
        b, c, h, w = q.shape
        q = q.reshape(b, self.head_size, self.att_size, h * w)
        q = q.permute(0, 3, 1, 2)  # b, hw, head, att

        k = k.reshape(b, self.head_size, self.att_size, h * w)
        k = k.permute(0, 3, 1, 2)

        v = v.reshape(b, self.head_size, self.att_size, h * w)
        v = v.permute(0, 3, 1, 2)

        q = q.transpose(1, 2)
        v = v.transpose(1, 2)
        k = k.transpose(1, 2).transpose(2, 3)

        scale = int(self.att_size)**(-0.5)
        q.mul_(scale)
        w_ = torch.matmul(q, k)
        w_ = F.softmax(w_, dim=3)

        w_ = w_.matmul(v)

        w_ = w_.transpose(1, 2).contiguous()  # [b, h*w, head, att]
        w_ = w_.view(b, h, w, -1)
        w_ = w_.permute(0, 3, 1, 2)

        w_ = self.proj_out(w_)

        return x + w_


class MultiHeadEncoder(nn.Module):

    def __init__(self,
                 ch,
                 out_ch,
                 ch_mult=(1, 2, 4, 8),
                 num_res_blocks=2,
                 attn_resolutions=(16, ),
                 dropout=0.0,
                 resamp_with_conv=True,
                 in_channels=3,
                 resolution=512,
                 z_channels=256,
                 double_z=True,
                 enable_mid=True,
                 head_size=1,
                 **ignore_kwargs):
        super().__init__()
        self.ch = ch
        self.temb_ch = 0
        self.num_resolutions = len(ch_mult)
        self.num_res_blocks = num_res_blocks
        self.resolution = resolution
        self.in_channels = in_channels
        self.enable_mid = enable_mid

        # downsampling
        self.conv_in = torch.nn.Conv2d(in_channels, self.ch, kernel_size=3, stride=1, padding=1)

        curr_res = resolution
        in_ch_mult = (1, ) + tuple(ch_mult)
        self.down = nn.ModuleList()
        for i_level in range(self.num_resolutions):
            block = nn.ModuleList()
            attn = nn.ModuleList()
            block_in = ch * in_ch_mult[i_level]
            block_out = ch * ch_mult[i_level]
            for i_block in range(self.num_res_blocks):
                block.append(
                    ResnetBlock(
                        in_channels=block_in, out_channels=block_out, temb_channels=self.temb_ch, dropout=dropout))
                block_in = block_out
                if curr_res in attn_resolutions:
                    attn.append(MultiHeadAttnBlock(block_in, head_size))
            down = nn.Module()
            down.block = block
            down.attn = attn
            if i_level != self.num_resolutions - 1:
                down.downsample = Downsample(block_in, resamp_with_conv)
                curr_res = curr_res // 2
            self.down.append(down)

        # middle
        if self.enable_mid:
            self.mid = nn.Module()
            self.mid.block_1 = ResnetBlock(
                in_channels=block_in, out_channels=block_in, temb_channels=self.temb_ch, dropout=dropout)
            self.mid.attn_1 = MultiHeadAttnBlock(block_in, head_size)
            self.mid.block_2 = ResnetBlock(
                in_channels=block_in, out_channels=block_in, temb_channels=self.temb_ch, dropout=dropout)

        # end
        self.norm_out = Normalize(block_in)
        self.conv_out = torch.nn.Conv2d(
            block_in, 2 * z_channels if double_z else z_channels, kernel_size=3, stride=1, padding=1)

    def forward(self, x):
        hs = {}
        # timestep embedding
        temb = None

        # downsampling
        h = self.conv_in(x)
        hs['in'] = h
        for i_level in range(self.num_resolutions):
            for i_block in range(self.num_res_blocks):
                h = self.down[i_level].block[i_block](h, temb)
                if len(self.down[i_level].attn) > 0:
                    h = self.down[i_level].attn[i_block](h)

            if i_level != self.num_resolutions - 1:
                # hs.append(h)
                hs['block_' + str(i_level)] = h
                h = self.down[i_level].downsample(h)

        # middle
        # h = hs[-1]
        if self.enable_mid:
            h = self.mid.block_1(h, temb)
            hs['block_' + str(i_level) + '_atten'] = h
            h = self.mid.attn_1(h)
            h = self.mid.block_2(h, temb)
            hs['mid_atten'] = h

        # end
        h = self.norm_out(h)
        h = nonlinearity(h)
        h = self.conv_out(h)
        # hs.append(h)
        hs['out'] = h

        return hs


class MultiHeadDecoder(nn.Module):

    def __init__(self,
                 ch,
                 out_ch,
                 ch_mult=(1, 2, 4, 8),
                 num_res_blocks=2,
                 attn_resolutions=(16, ),
                 dropout=0.0,
                 resamp_with_conv=True,
                 in_channels=3,
                 resolution=512,
                 z_channels=256,
                 give_pre_end=False,
                 enable_mid=True,
                 head_size=1,
                 **ignorekwargs):
        super().__init__()
        self.ch = ch
        self.temb_ch = 0
        self.num_resolutions = len(ch_mult)
        self.num_res_blocks = num_res_blocks
        self.resolution = resolution
        self.in_channels = in_channels
        self.give_pre_end = give_pre_end
        self.enable_mid = enable_mid

        # compute in_ch_mult, block_in and curr_res at lowest res
        block_in = ch * ch_mult[self.num_resolutions - 1]
        curr_res = resolution // 2**(self.num_resolutions - 1)
        self.z_shape = (1, z_channels, curr_res, curr_res)
        print('Working with z of shape {} = {} dimensions.'.format(self.z_shape, np.prod(self.z_shape)))

        # z to block_in
        self.conv_in = torch.nn.Conv2d(z_channels, block_in, kernel_size=3, stride=1, padding=1)

        # middle
        if self.enable_mid:
            self.mid = nn.Module()
            self.mid.block_1 = ResnetBlock(
                in_channels=block_in, out_channels=block_in, temb_channels=self.temb_ch, dropout=dropout)
            self.mid.attn_1 = MultiHeadAttnBlock(block_in, head_size)
            self.mid.block_2 = ResnetBlock(
                in_channels=block_in, out_channels=block_in, temb_channels=self.temb_ch, dropout=dropout)

        # upsampling
        self.up = nn.ModuleList()
        for i_level in reversed(range(self.num_resolutions)):
            block = nn.ModuleList()
            attn = nn.ModuleList()
            block_out = ch * ch_mult[i_level]
            for i_block in range(self.num_res_blocks + 1):
                block.append(
                    ResnetBlock(
                        in_channels=block_in, out_channels=block_out, temb_channels=self.temb_ch, dropout=dropout))
                block_in = block_out
                if curr_res in attn_resolutions:
                    attn.append(MultiHeadAttnBlock(block_in, head_size))
            up = nn.Module()
            up.block = block
            up.attn = attn
            if i_level != 0:
                up.upsample = Upsample(block_in, resamp_with_conv)
                curr_res = curr_res * 2
            self.up.insert(0, up)  # prepend to get consistent order

        # end
        self.norm_out = Normalize(block_in)
        self.conv_out = torch.nn.Conv2d(block_in, out_ch, kernel_size=3, stride=1, padding=1)

    def forward(self, z):
        # assert z.shape[1:] == self.z_shape[1:]
        self.last_z_shape = z.shape

        # timestep embedding
        temb = None

        # z to block_in
        h = self.conv_in(z)

        # middle
        if self.enable_mid:
            h = self.mid.block_1(h, temb)
            h = self.mid.attn_1(h)
            h = self.mid.block_2(h, temb)

        # upsampling
        for i_level in reversed(range(self.num_resolutions)):
            for i_block in range(self.num_res_blocks + 1):
                h = self.up[i_level].block[i_block](h, temb)
                if len(self.up[i_level].attn) > 0:
                    h = self.up[i_level].attn[i_block](h)
            if i_level != 0:
                h = self.up[i_level].upsample(h)

        # end
        if self.give_pre_end:
            return h

        h = self.norm_out(h)
        h = nonlinearity(h)
        h = self.conv_out(h)
        return h


class MultiHeadDecoderTransformer(nn.Module):

    def __init__(self,
                 ch,
                 out_ch,
                 ch_mult=(1, 2, 4, 8),
                 num_res_blocks=2,
                 attn_resolutions=(16, ),
                 dropout=0.0,
                 resamp_with_conv=True,
                 in_channels=3,
                 resolution=512,
                 z_channels=256,
                 give_pre_end=False,
                 enable_mid=True,
                 head_size=1,
                 **ignorekwargs):
        super().__init__()
        self.ch = ch
        self.temb_ch = 0
        self.num_resolutions = len(ch_mult)
        self.num_res_blocks = num_res_blocks
        self.resolution = resolution
        self.in_channels = in_channels
        self.give_pre_end = give_pre_end
        self.enable_mid = enable_mid

        # compute in_ch_mult, block_in and curr_res at lowest res
        block_in = ch * ch_mult[self.num_resolutions - 1]
        curr_res = resolution // 2**(self.num_resolutions - 1)
        self.z_shape = (1, z_channels, curr_res, curr_res)
        print('Working with z of shape {} = {} dimensions.'.format(self.z_shape, np.prod(self.z_shape)))

        # z to block_in
        self.conv_in = torch.nn.Conv2d(z_channels, block_in, kernel_size=3, stride=1, padding=1)

        # middle
        if self.enable_mid:
            self.mid = nn.Module()
            self.mid.block_1 = ResnetBlock(
                in_channels=block_in, out_channels=block_in, temb_channels=self.temb_ch, dropout=dropout)
            self.mid.attn_1 = MultiHeadAttnBlock(block_in, head_size)
            self.mid.block_2 = ResnetBlock(
                in_channels=block_in, out_channels=block_in, temb_channels=self.temb_ch, dropout=dropout)

        # upsampling
        self.up = nn.ModuleList()
        for i_level in reversed(range(self.num_resolutions)):
            block = nn.ModuleList()
            attn = nn.ModuleList()
            block_out = ch * ch_mult[i_level]
            for i_block in range(self.num_res_blocks + 1):
                block.append(
                    ResnetBlock(
                        in_channels=block_in, out_channels=block_out, temb_channels=self.temb_ch, dropout=dropout))
                block_in = block_out
                if curr_res in attn_resolutions:
                    attn.append(MultiHeadAttnBlock(block_in, head_size))
            up = nn.Module()
            up.block = block
            up.attn = attn
            if i_level != 0:
                up.upsample = Upsample(block_in, resamp_with_conv)
                curr_res = curr_res * 2
            self.up.insert(0, up)  # prepend to get consistent order

        # end
        self.norm_out = Normalize(block_in)
        self.conv_out = torch.nn.Conv2d(block_in, out_ch, kernel_size=3, stride=1, padding=1)

    def forward(self, z, hs):
        # assert z.shape[1:] == self.z_shape[1:]
        # self.last_z_shape = z.shape

        # timestep embedding
        temb = None

        # z to block_in
        h = self.conv_in(z)

        # middle
        if self.enable_mid:
            h = self.mid.block_1(h, temb)
            h = self.mid.attn_1(h, hs['mid_atten'])
            h = self.mid.block_2(h, temb)

        # upsampling
        for i_level in reversed(range(self.num_resolutions)):
            for i_block in range(self.num_res_blocks + 1):
                h = self.up[i_level].block[i_block](h, temb)
                if len(self.up[i_level].attn) > 0:
                    h = self.up[i_level].attn[i_block](h, hs['block_' + str(i_level) + '_atten'])
                    # hfeature = h.clone()
            if i_level != 0:
                h = self.up[i_level].upsample(h)

        # end
        if self.give_pre_end:
            return h

        h = self.norm_out(h)
        h = nonlinearity(h)
        h = self.conv_out(h)
        return h


class RestoreFormer(nn.Module):

    def __init__(self,
                 n_embed=1024,
                 embed_dim=256,
                 ch=64,
                 out_ch=3,
                 ch_mult=(1, 2, 2, 4, 4, 8),
                 num_res_blocks=2,
                 attn_resolutions=(16, ),
                 dropout=0.0,
                 in_channels=3,
                 resolution=512,
                 z_channels=256,
                 double_z=False,
                 enable_mid=True,
                 fix_decoder=False,
                 fix_codebook=True,
                 fix_encoder=False,
                 head_size=8):
        super(RestoreFormer, self).__init__()

        self.encoder = MultiHeadEncoder(
            ch=ch,
            out_ch=out_ch,
            ch_mult=ch_mult,
            num_res_blocks=num_res_blocks,
            attn_resolutions=attn_resolutions,
            dropout=dropout,
            in_channels=in_channels,
            resolution=resolution,
            z_channels=z_channels,
            double_z=double_z,
            enable_mid=enable_mid,
            head_size=head_size)
        self.decoder = MultiHeadDecoderTransformer(
            ch=ch,
            out_ch=out_ch,
            ch_mult=ch_mult,
            num_res_blocks=num_res_blocks,
            attn_resolutions=attn_resolutions,
            dropout=dropout,
            in_channels=in_channels,
            resolution=resolution,
            z_channels=z_channels,
            enable_mid=enable_mid,
            head_size=head_size)

        self.quantize = VectorQuantizer(n_embed, embed_dim, beta=0.25)

        self.quant_conv = torch.nn.Conv2d(z_channels, embed_dim, 1)
        self.post_quant_conv = torch.nn.Conv2d(embed_dim, z_channels, 1)

        if fix_decoder:
            for _, param in self.decoder.named_parameters():
                param.requires_grad = False
            for _, param in self.post_quant_conv.named_parameters():
                param.requires_grad = False
            for _, param in self.quantize.named_parameters():
                param.requires_grad = False
        elif fix_codebook:
            for _, param in self.quantize.named_parameters():
                param.requires_grad = False

        if fix_encoder:
            for _, param in self.encoder.named_parameters():
                param.requires_grad = False

    def encode(self, x):

        hs = self.encoder(x)
        h = self.quant_conv(hs['out'])
        quant, emb_loss, info = self.quantize(h)
        return quant, emb_loss, info, hs

    def decode(self, quant, hs):
        quant = self.post_quant_conv(quant)
        dec = self.decoder(quant, hs)

        return dec

    def forward(self, input, **kwargs):
        quant, diff, info, hs = self.encode(input)
        dec = self.decode(quant, hs)

        return dec, None