unet.py

# Copyright 2022 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and

# limitations under the License.

# helpers functions

import torch
from torch import nn

from ..configuration_utils import ConfigMixin
from ..modeling_utils import ModelMixin
from .attention import AttentionBlock
from .embeddings import get_timestep_embedding
from .resnet import Downsample, ResnetBlock, Upsample


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 UNetModel(ModelMixin, ConfigMixin):
    def __init__(
        self,
        ch=128,
        out_ch=3,
        ch_mult=(1, 1, 2, 2, 4, 4),
        num_res_blocks=2,
        attn_resolutions=(16,),
        dropout=0.0,
        resamp_with_conv=True,
        in_channels=3,
        resolution=256,
    ):
        super().__init__()
        self.register_to_config(
            ch=ch,
            out_ch=out_ch,
            ch_mult=ch_mult,
            num_res_blocks=num_res_blocks,
            attn_resolutions=attn_resolutions,
            dropout=dropout,
            resamp_with_conv=resamp_with_conv,
            in_channels=in_channels,
            resolution=resolution,
        )
        ch_mult = tuple(ch_mult)
        self.ch = ch
        self.temb_ch = self.ch * 4
        self.num_resolutions = len(ch_mult)
        self.num_res_blocks = num_res_blocks
        self.resolution = resolution
        self.in_channels = in_channels

        # timestep embedding
        self.temb = nn.Module()
        self.temb.dense = nn.ModuleList(
            [
                torch.nn.Linear(self.ch, self.temb_ch),
                torch.nn.Linear(self.temb_ch, self.temb_ch),
            ]
        )

        # 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,) + 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(AttentionBlock(block_in, overwrite_qkv=True))
            down = nn.Module()
            down.block = block
            down.attn = attn
            if i_level != self.num_resolutions - 1:
                down.downsample = Downsample(block_in, use_conv=resamp_with_conv, padding=0)
                curr_res = curr_res // 2
            self.down.append(down)

        # middle
        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 = AttentionBlock(block_in, overwrite_qkv=True)
        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]
            skip_in = ch * ch_mult[i_level]
            for i_block in range(self.num_res_blocks + 1):
                if i_block == self.num_res_blocks:
                    skip_in = ch * in_ch_mult[i_level]
                block.append(
                    ResnetBlock(
                        in_channels=block_in + skip_in,
                        out_channels=block_out,
                        temb_channels=self.temb_ch,
                        dropout=dropout,
                    )
                )
                block_in = block_out
                if curr_res in attn_resolutions:
                    attn.append(AttentionBlock(block_in, overwrite_qkv=True))
            up = nn.Module()
            up.block = block
            up.attn = attn
            if i_level != 0:
                up.upsample = Upsample(block_in, use_conv=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, x, timesteps):
        assert x.shape[2] == x.shape[3] == self.resolution

        if not torch.is_tensor(timesteps):
            timesteps = torch.tensor([timesteps], dtype=torch.long, device=x.device)

        # timestep embedding
        temb = get_timestep_embedding(timesteps, self.ch)
        temb = self.temb.dense[0](temb)
        temb = nonlinearity(temb)
        temb = self.temb.dense[1](temb)

        # downsampling
        hs = [self.conv_in(x)]
        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](hs[-1], temb)
                if len(self.down[i_level].attn) > 0:
                    h = self.down[i_level].attn[i_block](h)
                hs.append(h)
            if i_level != self.num_resolutions - 1:
                hs.append(self.down[i_level].downsample(hs[-1]))

        # middle
        h = hs[-1]
        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](torch.cat([h, hs.pop()], dim=1), 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
        h = self.norm_out(h)
        h = nonlinearity(h)
        h = self.conv_out(h)
        return h