deepinv 0.1.0.dev0__py3-none-any.whl

deepinv/models/diffunet.py

@@ -0,0 +1,997 @@
+# This file is a concatenation of DiffPIR codes available here: https://github.com/yuanzhi-zhu/DiffPIR/tree/main
+# This code is taken (with minor modifications) from https://github.com/yuanzhi-zhu/DiffPIR/tree/main
+
+import torch
+from .utils import get_weights_url
+from abc import abstractmethod
+import numpy as np
+import torch.nn as nn
+import torch.nn.functional as F
+
+
+class DiffUNet(nn.Module):
+    r"""
+    Diffusion UNet model.
+
+    This is the model with attention and timestep embeddings from `Ho et al. <https://arxiv.org/abs/2108.02938>`_;
+    code is adapted from https://github.com/jychoi118/ilvr_adm.
+
+    It is possible to choose the `standard model <https://arxiv.org/abs/2108.02938>`_
+    with 128 hidden channels per layer (trained on FFHQ)
+    and a `larger model <https://arxiv.org/abs/2105.05233>`_ with 256 hidden channels per layer (trained on ImageNet128)
+
+    A pretrained network for (in_channels=out_channels=3)
+    can be downloaded via setting ``pretrained='download'``.
+
+    The network can handle images of size :math:`2^{n_1}\times 2^{n_2}` with :math:`n_1,n_2 \geq 5`.
+
+
+    :param int in_channels: channels in the input Tensor.
+    :param int out_channels: channels in the output Tensor.
+    :param bool large_model: if True, use the large model with 256 hidden channels per layer trained on ImageNet128
+        (weights size: 2.1 GB).
+        Otherwise, use a smaller model with 128 hidden channels per layer trained on FFHQ (weights size: 357 MB).
+    :param str, None pretrained: use a pretrained network. If ``pretrained=None``, the weights will be initialized at
+        random using Pytorch's default initialization.
+        If ``pretrained='download'``, the weights will be downloaded from an online repository
+        (only available for 3 input and output channels).
+        Finally, ``pretrained`` can also be set as a path to the user's own pretrained weights.
+        See :ref:`pretrained-weights <pretrained-weights>` for more details.
+    """
+
+    def __init__(
+        self,
+        in_channels=3,
+        out_channels=3,
+        large_model=False,
+        use_fp16=False,
+        pretrained="download",
+    ):
+        super().__init__()
+
+        if large_model:
+            model_channels = 256
+            num_res_blocks = 2
+            attention_resolutions = "8,16,32"
+        else:
+            model_channels = 128
+            num_res_blocks = 1
+            attention_resolutions = "16"
+
+        dropout = 0.1
+        conv_resample = True
+        dims = 2
+        num_classes = None
+        use_checkpoint = False
+        num_heads = 4
+        num_head_channels = 64
+        num_heads_upsample = -1
+        use_scale_shift_norm = True
+        resblock_updown = True
+        use_new_attention_order = False
+
+        out_channels = 6 if out_channels == 3 else out_channels
+        channel_mult = (1, 1, 2, 2, 4, 4)
+
+        image_size = 256
+        attention_ds = []
+        for res in attention_resolutions.split(","):
+            attention_ds.append(image_size // int(res))
+        attention_resolutions = tuple(attention_ds)
+
+        if num_heads_upsample == -1:
+            num_heads_upsample = num_heads
+
+        self.image_size = image_size
+        self.in_channels = in_channels
+        self.model_channels = model_channels
+        self.out_channels = out_channels
+        self.num_res_blocks = num_res_blocks
+        self.attention_resolutions = attention_resolutions
+        self.dropout = dropout
+        self.channel_mult = channel_mult
+        self.conv_resample = conv_resample
+        self.num_classes = num_classes
+        self.use_checkpoint = use_checkpoint
+        self.dtype = th.float16 if use_fp16 else th.float32
+        self.num_heads = num_heads
+        self.num_head_channels = num_head_channels
+        self.num_heads_upsample = num_heads_upsample
+
+        time_embed_dim = model_channels * 4
+        self.time_embed = nn.Sequential(
+            linear(model_channels, time_embed_dim),
+            nn.SiLU(),
+            linear(time_embed_dim, time_embed_dim),
+        )
+
+        if self.num_classes is not None:
+            self.label_emb = nn.Embedding(num_classes, time_embed_dim)
+
+        ch = input_ch = int(channel_mult[0] * model_channels)
+        self.input_blocks = nn.ModuleList(
+            [TimestepEmbedSequential(conv_nd(dims, in_channels, ch, 3, padding=1))]
+        )
+        self._feature_size = ch
+        input_block_chans = [ch]
+        ds = 1
+        for level, mult in enumerate(channel_mult):
+            for _ in range(num_res_blocks):
+                layers = [
+                    ResBlock(
+                        ch,
+                        time_embed_dim,
+                        dropout,
+                        out_channels=int(mult * model_channels),
+                        dims=dims,
+                        use_checkpoint=use_checkpoint,
+                        use_scale_shift_norm=use_scale_shift_norm,
+                    )
+                ]
+                ch = int(mult * model_channels)
+                if ds in attention_resolutions:
+                    layers.append(
+                        AttentionBlock(
+                            ch,
+                            use_checkpoint=use_checkpoint,
+                            num_heads=num_heads,
+                            num_head_channels=num_head_channels,
+                            use_new_attention_order=use_new_attention_order,
+                        )
+                    )
+                self.input_blocks.append(TimestepEmbedSequential(*layers))
+                self._feature_size += ch
+                input_block_chans.append(ch)
+            if level != len(channel_mult) - 1:
+                out_ch = ch
+                self.input_blocks.append(
+                    TimestepEmbedSequential(
+                        ResBlock(
+                            ch,
+                            time_embed_dim,
+                            dropout,
+                            out_channels=out_ch,
+                            dims=dims,
+                            use_checkpoint=use_checkpoint,
+                            use_scale_shift_norm=use_scale_shift_norm,
+                            down=True,
+                        )
+                        if resblock_updown
+                        else Downsample(
+                            ch, conv_resample, dims=dims, out_channels=out_ch
+                        )
+                    )
+                )
+                ch = out_ch
+                input_block_chans.append(ch)
+                ds *= 2
+                self._feature_size += ch
+
+        self.middle_block = TimestepEmbedSequential(
+            ResBlock(
+                ch,
+                time_embed_dim,
+                dropout,
+                dims=dims,
+                use_checkpoint=use_checkpoint,
+                use_scale_shift_norm=use_scale_shift_norm,
+            ),
+            AttentionBlock(
+                ch,
+                use_checkpoint=use_checkpoint,
+                num_heads=num_heads,
+                num_head_channels=num_head_channels,
+                use_new_attention_order=use_new_attention_order,
+            ),
+            ResBlock(
+                ch,
+                time_embed_dim,
+                dropout,
+                dims=dims,
+                use_checkpoint=use_checkpoint,
+                use_scale_shift_norm=use_scale_shift_norm,
+            ),
+        )
+        self._feature_size += ch
+
+        self.output_blocks = nn.ModuleList([])
+        for level, mult in list(enumerate(channel_mult))[::-1]:
+            for i in range(num_res_blocks + 1):
+                ich = input_block_chans.pop()
+                layers = [
+                    ResBlock(
+                        ch + ich,
+                        time_embed_dim,
+                        dropout,
+                        out_channels=int(model_channels * mult),
+                        dims=dims,
+                        use_checkpoint=use_checkpoint,
+                        use_scale_shift_norm=use_scale_shift_norm,
+                    )
+                ]
+                ch = int(model_channels * mult)
+                if ds in attention_resolutions:
+                    layers.append(
+                        AttentionBlock(
+                            ch,
+                            use_checkpoint=use_checkpoint,
+                            num_heads=num_heads_upsample,
+                            num_head_channels=num_head_channels,
+                            use_new_attention_order=use_new_attention_order,
+                        )
+                    )
+                if level and i == num_res_blocks:
+                    out_ch = ch
+                    layers.append(
+                        ResBlock(
+                            ch,
+                            time_embed_dim,
+                            dropout,
+                            out_channels=out_ch,
+                            dims=dims,
+                            use_checkpoint=use_checkpoint,
+                            use_scale_shift_norm=use_scale_shift_norm,
+                            up=True,
+                        )
+                        if resblock_updown
+                        else Upsample(ch, conv_resample, dims=dims, out_channels=out_ch)
+                    )
+                    ds //= 2
+                self.output_blocks.append(TimestepEmbedSequential(*layers))
+                self._feature_size += ch
+
+        self.out = nn.Sequential(
+            normalization(ch),
+            nn.SiLU(),
+            zero_module(conv_nd(dims, input_ch, out_channels, 3, padding=1)),
+        )
+
+        if pretrained is not None:
+            if pretrained == "download":
+                if in_channels == 3 and out_channels == 6 and not large_model:
+                    name = "diffusion_ffhq_10m.pt"
+                elif in_channels == 3 and out_channels == 6 and large_model:
+                    name = "diffusion_openai.pt"
+                else:
+                    raise ValueError(
+                        "no existing pretrained model matches the requested configuration"
+                    )
+                url = get_weights_url(model_name="diffunet", file_name=name)
+                ckpt = torch.hub.load_state_dict_from_url(
+                    url, map_location=lambda storage, loc: storage, file_name=name
+                )
+            else:
+                ckpt = torch.load(pretrained, map_location=lambda storage, loc: storage)
+
+            self.load_state_dict(ckpt, strict=True)
+
+    def convert_to_fp16(self):
+        """
+        Convert the torso of the model to float16.
+        """
+        self.input_blocks.apply(convert_module_to_f16)
+        self.middle_block.apply(convert_module_to_f16)
+        self.output_blocks.apply(convert_module_to_f16)
+
+    def convert_to_fp32(self):
+        """
+        Convert the torso of the model to float32.
+        """
+        self.input_blocks.apply(convert_module_to_f32)
+        self.middle_block.apply(convert_module_to_f32)
+        self.output_blocks.apply(convert_module_to_f32)
+
+    def forward(self, x, t, y=None, type_t="noise_level"):
+        r"""
+        Apply the model to an input batch.
+
+        This function takes a noisy image and either a timestep or a noise level as input. Depending on the nature of
+        ``t``, the model returns either a noise map (if ``type_t='timestep'``) or a denoised image (if
+        ``type_t='noise_level'``).
+
+        :param x: an [N x C x ...] Tensor of inputs.
+        :param t: a 1-D batch of timesteps or noise levels.
+        :param y: an [N] Tensor of labels, if class-conditional. Default=None.
+        :param type_t: Nature of the embedding `t`. In traditional diffusion model, and in the authors' code, `t` is
+                       a timestep linked to a noise level; in this case, set ``type_t='timestep'``. We can also choose
+                       ``t`` to be a noise level directly and use the model as a denoiser; in this case, set
+                       ``type_t='noise_level'``. Default: ``'timestep'``.
+        :return: an [N x C x ...] Tensor of outputs. Either a noise map (if ``type_t='timestep'``) or a denoised image
+                    (if ``type_t='noise_level'``).
+        """
+        if type_t == "timestep":
+            return self.forward_diffusion(x, t, y=y)
+        elif type_t == "noise_level":
+            return self.forward_denoise(x, t, y=y)
+        else:
+            raise ValueError('type_t must be either "timestep" or "noise_level"')
+
+    def forward_diffusion(self, x, timesteps, y=None):
+        r"""
+        Apply the model to an input batch.
+
+        This function takes a noisy image and a timestep as input (and not a noise level) and estimates the noise map
+        in the input image.
+        The image is assumed to be in range [-1, 1] and to have dimensions with width and height divisible by a
+        power of 2.
+
+        :param x: an [N x C x ...] Tensor of inputs.
+        :param timesteps: a 1-D batch of timesteps.
+        :param y: an [N] Tensor of labels, if class-conditional. Default=None.
+        :return: an [N x C x ...] Tensor of outputs.
+        """
+        assert (y is not None) == (
+            self.num_classes is not None
+        ), "must specify y if and only if the model is class-conditional"
+
+        hs = []
+        emb = self.time_embed(timestep_embedding(timesteps, self.model_channels))
+
+        if self.num_classes is not None:
+            assert y.shape == (x.shape[0],)
+            emb = emb + self.label_emb(y)
+
+        h = x.type(self.dtype)
+        for module in self.input_blocks:
+            h = module(h, emb)
+            hs.append(h)
+        h = self.middle_block(h, emb)
+        for module in self.output_blocks:
+            h = th.cat([h, hs.pop()], dim=1)
+            h = module(h, emb)
+        h = h.type(x.dtype)
+        return self.out(h)
+
+    def get_alpha_prod(
+        self, beta_start=0.1 / 1000, beta_end=20 / 1000, num_train_timesteps=1000
+    ):
+        """
+        Get the alpha sequences; this is necessary for mapping noise levels to timesteps when performing pure denoising.
+        """
+        betas = np.linspace(beta_start, beta_end, num_train_timesteps, dtype=np.float32)
+        betas = torch.from_numpy(
+            betas
+        )  # .to(self.device) Removing this for now, can be done outside
+        alphas = 1.0 - betas
+        alphas_cumprod = np.cumprod(alphas.cpu(), axis=0)  # This is \overline{\alpha}_t
+
+        # Useful sequences deriving from alphas_cumprod
+        sqrt_alphas_cumprod = torch.sqrt(alphas_cumprod)
+        sqrt_1m_alphas_cumprod = torch.sqrt(1.0 - alphas_cumprod)
+        reduced_alpha_cumprod = torch.div(sqrt_1m_alphas_cumprod, sqrt_alphas_cumprod)
+        sqrt_recip_alphas_cumprod = torch.sqrt(1.0 / alphas_cumprod)
+        sqrt_recipm1_alphas_cumprod = torch.sqrt(1.0 / alphas_cumprod - 1)
+        return (
+            reduced_alpha_cumprod,
+            sqrt_recip_alphas_cumprod,
+            sqrt_recipm1_alphas_cumprod,
+        )
+
+    def find_nearest(self, array, value):
+        """
+        Find the argmin of the nearest value in an array.
+        """
+        array = np.asarray(array)
+        if isinstance(value, torch.Tensor):
+            value = np.asarray(value.cpu())
+        idx = (np.abs(array - value)).argmin()
+        return idx
+
+    def forward_denoise(self, x, sigma, y=None):
+        r"""
+        Apply the denoising model to an input batch.
+
+        This function takes a noisy image and a noise level as input (and not a timestep) and estimates the noiseless
+        underlying image in the input image.
+        The input image is assumed to be in range [0, 1] (up to noise) and to have dimensions with width and height
+        divisible by a power of 2.
+
+        :param x: an [N x C x ...] Tensor of inputs.
+        :param sigma: a 1-D batch of noise levels.
+        :param y: an [N] Tensor of labels, if class-conditional. Default=None.
+        :return: an [N x C x ...] Tensor of outputs.
+        """
+        x = 2.0 * x - 1.0
+        (
+            reduced_alpha_cumprod,
+            sqrt_recip_alphas_cumprod,
+            sqrt_recipm1_alphas_cumprod,
+        ) = self.get_alpha_prod()
+        timesteps = self.find_nearest(
+            reduced_alpha_cumprod, sigma * 2
+        )  # Factor 2 because image rescaled in [-1, 1]
+
+        noise_est_sample_var = self.forward_diffusion(
+            x, torch.tensor([timesteps]).to(x.device), y=y
+        )
+        noise_est = noise_est_sample_var[:, :3, ...]
+        denoised = (
+            sqrt_recip_alphas_cumprod[timesteps] * x
+            - sqrt_recipm1_alphas_cumprod[timesteps] * noise_est
+        )
+        denoised = denoised.clamp(-1, 1)
+        return denoised / 2.0 + 0.5
+
+
+class AttentionPool2d(nn.Module):
+    """
+    Adapted from CLIP: https://github.com/openai/CLIP/blob/main/clip/model.py
+    """
+
+    def __init__(
+        self,
+        spacial_dim: int,
+        embed_dim: int,
+        num_heads_channels: int,
+        output_dim: int = None,
+    ):
+        super().__init__()
+        self.positional_embedding = nn.Parameter(
+            th.randn(embed_dim, spacial_dim**2 + 1) / embed_dim**0.5
+        )
+        self.qkv_proj = conv_nd(1, embed_dim, 3 * embed_dim, 1)
+        self.c_proj = conv_nd(1, embed_dim, output_dim or embed_dim, 1)
+        self.num_heads = embed_dim // num_heads_channels
+        self.attention = QKVAttention(self.num_heads)
+
+    def forward(self, x):
+        b, c, *_spatial = x.shape
+        x = x.reshape(b, c, -1)  # NC(HW)
+        x = th.cat([x.mean(dim=-1, keepdim=True), x], dim=-1)  # NC(HW+1)
+        x = x + self.positional_embedding[None, :, :].to(x.dtype)  # NC(HW+1)
+        x = self.qkv_proj(x)
+        x = self.attention(x)
+        x = self.c_proj(x)
+        return x[:, :, 0]
+
+
+class TimestepBlock(nn.Module):
+    """
+    Any module where forward() takes timestep embeddings as a second argument.
+    """
+
+    @abstractmethod
+    def forward(self, x, emb):
+        """
+        Apply the module to `x` given `emb` timestep embeddings.
+        """
+
+
+class TimestepEmbedSequential(nn.Sequential, TimestepBlock):
+    """
+    A sequential module that passes timestep embeddings to the children that
+    support it as an extra input.
+    """
+
+    def forward(self, x, emb):
+        for layer in self:
+            if isinstance(layer, TimestepBlock):
+                x = layer(x, emb)
+            else:
+                x = layer(x)
+        return x
+
+
+class Upsample(nn.Module):
+    """
+    An upsampling layer with an optional convolution.
+
+    :param channels: channels in the inputs and outputs.
+    :param use_conv: a bool determining if a convolution is applied.
+    :param dims: determines if the signal is 1D, 2D, or 3D. If 3D, then
+                 upsampling occurs in the inner-two dimensions.
+    """
+
+    def __init__(self, channels, use_conv, dims=2, out_channels=None):
+        super().__init__()
+        self.channels = channels
+        self.out_channels = out_channels or channels
+        self.use_conv = use_conv
+        self.dims = dims
+        if use_conv:
+            self.conv = conv_nd(dims, self.channels, self.out_channels, 3, padding=1)
+
+    def forward(self, x):
+        assert x.shape[1] == self.channels
+        if self.dims == 3:
+            x = F.interpolate(
+                x, (x.shape[2], x.shape[3] * 2, x.shape[4] * 2), mode="nearest"
+            )
+        else:
+            x = F.interpolate(x, scale_factor=2, mode="nearest")
+        if self.use_conv:
+            x = self.conv(x)
+        return x
+
+
+class Downsample(nn.Module):
+    """
+    A downsampling layer with an optional convolution.
+
+    :param channels: channels in the inputs and outputs.
+    :param use_conv: a bool determining if a convolution is applied.
+    :param dims: determines if the signal is 1D, 2D, or 3D. If 3D, then
+                 downsampling occurs in the inner-two dimensions.
+    """
+
+    def __init__(self, channels, use_conv, dims=2, out_channels=None):
+        super().__init__()
+        self.channels = channels
+        self.out_channels = out_channels or channels
+        self.use_conv = use_conv
+        self.dims = dims
+        stride = 2 if dims != 3 else (1, 2, 2)
+        if use_conv:
+            self.op = conv_nd(
+                dims, self.channels, self.out_channels, 3, stride=stride, padding=1
+            )
+        else:
+            assert self.channels == self.out_channels
+            self.op = avg_pool_nd(dims, kernel_size=stride, stride=stride)
+
+    def forward(self, x):
+        assert x.shape[1] == self.channels
+        return self.op(x)
+
+
+class ResBlock(TimestepBlock):
+    """
+    A residual block that can optionally change the number of channels.
+
+    :param channels: the number of input channels.
+    :param emb_channels: the number of timestep embedding channels.
+    :param dropout: the rate of dropout.
+    :param out_channels: if specified, the number of out channels.
+    :param use_conv: if True and out_channels is specified, use a spatial
+        convolution instead of a smaller 1x1 convolution to change the
+        channels in the skip connection.
+    :param dims: determines if the signal is 1D, 2D, or 3D.
+    :param use_checkpoint: if True, use gradient checkpointing on this module.
+    :param up: if True, use this block for upsampling.
+    :param down: if True, use this block for downsampling.
+    """
+
+    def __init__(
+        self,
+        channels,
+        emb_channels,
+        dropout,
+        out_channels=None,
+        use_conv=False,
+        use_scale_shift_norm=False,
+        dims=2,
+        use_checkpoint=False,
+        up=False,
+        down=False,
+    ):
+        super().__init__()
+        self.channels = channels
+        self.emb_channels = emb_channels
+        self.dropout = dropout
+        self.out_channels = out_channels or channels
+        self.use_conv = use_conv
+        self.use_checkpoint = use_checkpoint
+        self.use_scale_shift_norm = use_scale_shift_norm
+
+        self.in_layers = nn.Sequential(
+            normalization(channels),
+            nn.SiLU(),
+            conv_nd(dims, channels, self.out_channels, 3, padding=1),
+        )
+
+        self.updown = up or down
+
+        if up:
+            self.h_upd = Upsample(channels, False, dims)
+            self.x_upd = Upsample(channels, False, dims)
+        elif down:
+            self.h_upd = Downsample(channels, False, dims)
+            self.x_upd = Downsample(channels, False, dims)
+        else:
+            self.h_upd = self.x_upd = nn.Identity()
+
+        self.emb_layers = nn.Sequential(
+            nn.SiLU(),
+            linear(
+                emb_channels,
+                2 * self.out_channels if use_scale_shift_norm else self.out_channels,
+            ),
+        )
+        self.out_layers = nn.Sequential(
+            normalization(self.out_channels),
+            nn.SiLU(),
+            nn.Dropout(p=dropout),
+            zero_module(
+                conv_nd(dims, self.out_channels, self.out_channels, 3, padding=1)
+            ),
+        )
+
+        if self.out_channels == channels:
+            self.skip_connection = nn.Identity()
+        elif use_conv:
+            self.skip_connection = conv_nd(
+                dims, channels, self.out_channels, 3, padding=1
+            )
+        else:
+            self.skip_connection = conv_nd(dims, channels, self.out_channels, 1)
+
+    def forward(self, x, emb):
+        """
+        Apply the block to a Tensor, conditioned on a timestep embedding.
+
+        :param x: an [N x C x ...] Tensor of features.
+        :param emb: an [N x emb_channels] Tensor of timestep embeddings.
+        :return: an [N x C x ...] Tensor of outputs.
+        """
+        return checkpoint(
+            self._forward, (x, emb), self.parameters(), self.use_checkpoint
+        )
+
+    def _forward(self, x, emb):
+        if self.updown:
+            in_rest, in_conv = self.in_layers[:-1], self.in_layers[-1]
+            h = in_rest(x)
+            h = self.h_upd(h)
+            x = self.x_upd(x)
+            h = in_conv(h)
+        else:
+            h = self.in_layers(x)
+        emb_out = self.emb_layers(emb).type(h.dtype)
+        while len(emb_out.shape) < len(h.shape):
+            emb_out = emb_out[..., None]
+        if self.use_scale_shift_norm:
+            out_norm, out_rest = self.out_layers[0], self.out_layers[1:]
+            scale, shift = th.chunk(emb_out, 2, dim=1)
+            h = out_norm(h) * (1 + scale) + shift
+            h = out_rest(h)
+        else:
+            h = h + emb_out
+            h = self.out_layers(h)
+        return self.skip_connection(x) + h
+
+
+class AttentionBlock(nn.Module):
+    """
+    An attention block that allows spatial positions to attend to each other.
+
+    Originally ported from here, but adapted to the N-d case.
+    https://github.com/hojonathanho/diffusion/blob/1e0dceb3b3495bbe19116a5e1b3596cd0706c543/diffusion_tf/models/unet.py#L66.
+    """
+
+    def __init__(
+        self,
+        channels,
+        num_heads=1,
+        num_head_channels=-1,
+        use_checkpoint=False,
+        use_new_attention_order=False,
+    ):
+        super().__init__()
+        self.channels = channels
+        if num_head_channels == -1:
+            self.num_heads = num_heads
+        else:
+            assert (
+                channels % num_head_channels == 0
+            ), f"q,k,v channels {channels} is not divisible by num_head_channels {num_head_channels}"
+            self.num_heads = channels // num_head_channels
+        self.use_checkpoint = use_checkpoint
+        self.norm = normalization(channels)
+        self.qkv = conv_nd(1, channels, channels * 3, 1)
+        if use_new_attention_order:
+            # split qkv before split heads
+            self.attention = QKVAttention(self.num_heads)
+        else:
+            # split heads before split qkv
+            self.attention = QKVAttentionLegacy(self.num_heads)
+
+        self.proj_out = zero_module(conv_nd(1, channels, channels, 1))
+
+    def forward(self, x):
+        return checkpoint(self._forward, (x,), self.parameters(), True)
+
+    def _forward(self, x):
+        b, c, *spatial = x.shape
+        x = x.reshape(b, c, -1)
+        qkv = self.qkv(self.norm(x))
+        h = self.attention(qkv)
+        h = self.proj_out(h)
+        return (x + h).reshape(b, c, *spatial)
+
+
+def count_flops_attn(model, _x, y):
+    """
+    A counter for the `thop` package to count the operations in an
+    attention operation.
+    Meant to be used like:
+        macs, params = thop.profile(
+            model,
+            inputs=(inputs, timestamps),
+            custom_ops={QKVAttention: QKVAttention.count_flops},
+        )
+    """
+    b, c, *spatial = y[0].shape
+    num_spatial = int(np.prod(spatial))
+    # We perform two matmuls with the same number of ops.
+    # The first computes the weight matrix, the second computes
+    # the combination of the value vectors.
+    matmul_ops = 2 * b * (num_spatial**2) * c
+    model.total_ops += th.DoubleTensor([matmul_ops])
+
+
+class QKVAttentionLegacy(nn.Module):
+    """
+    A module which performs QKV attention. Matches legacy QKVAttention + input/ouput heads shaping
+    """
+
+    def __init__(self, n_heads):
+        super().__init__()
+        self.n_heads = n_heads
+
+    def forward(self, qkv):
+        """
+        Apply QKV attention.
+
+        :param qkv: an [N x (H * 3 * C) x T] tensor of Qs, Ks, and Vs.
+        :return: an [N x (H * C) x T] tensor after attention.
+        """
+        bs, width, length = qkv.shape
+        assert width % (3 * self.n_heads) == 0
+        ch = width // (3 * self.n_heads)
+        q, k, v = qkv.reshape(bs * self.n_heads, ch * 3, length).split(ch, dim=1)
+        scale = 1 / math.sqrt(math.sqrt(ch))
+        weight = th.einsum(
+            "bct,bcs->bts", q * scale, k * scale
+        )  # More stable with f16 than dividing afterwards
+        weight = th.softmax(weight.float(), dim=-1).type(weight.dtype)
+        a = th.einsum("bts,bcs->bct", weight, v)
+        return a.reshape(bs, -1, length)
+
+    @staticmethod
+    def count_flops(model, _x, y):
+        return count_flops_attn(model, _x, y)
+
+
+class QKVAttention(nn.Module):
+    """
+    A module which performs QKV attention and splits in a different order.
+    """
+
+    def __init__(self, n_heads):
+        super().__init__()
+        self.n_heads = n_heads
+
+    def forward(self, qkv):
+        """
+        Apply QKV attention.
+
+        :param qkv: an [N x (3 * H * C) x T] tensor of Qs, Ks, and Vs.
+        :return: an [N x (H * C) x T] tensor after attention.
+        """
+        bs, width, length = qkv.shape
+        assert width % (3 * self.n_heads) == 0
+        ch = width // (3 * self.n_heads)
+        q, k, v = qkv.chunk(3, dim=1)
+        scale = 1 / math.sqrt(math.sqrt(ch))
+        weight = th.einsum(
+            "bct,bcs->bts",
+            (q * scale).view(bs * self.n_heads, ch, length),
+            (k * scale).view(bs * self.n_heads, ch, length),
+        )  # More stable with f16 than dividing afterwards
+        weight = th.softmax(weight.float(), dim=-1).type(weight.dtype)
+        a = th.einsum("bts,bcs->bct", weight, v.reshape(bs * self.n_heads, ch, length))
+        return a.reshape(bs, -1, length)
+
+    @staticmethod
+    def count_flops(model, _x, y):
+        return count_flops_attn(model, _x, y)
+
+
+def checkpoint(func, inputs, params, flag):
+    """
+    Evaluate a function without caching intermediate activations, allowing for
+    reduced memory at the expense of extra compute in the backward pass.
+
+    :param func: the function to evaluate.
+    :param inputs: the argument sequence to pass to `func`.
+    :param params: a sequence of parameters `func` depends on but does not
+                   explicitly take as arguments.
+    :param flag: if False, disable gradient checkpointing.
+    """
+    if flag:
+        args = tuple(inputs) + tuple(params)
+        return CheckpointFunction.apply(func, len(inputs), *args)
+    else:
+        return func(*inputs)
+
+
+"""
+Various utilities for neural networks.
+"""
+
+import math
+
+import torch as th
+import torch.nn as nn
+
+
+# PyTorch 1.7 has SiLU, but we support PyTorch 1.5.
+class SiLU(nn.Module):
+    def forward(self, x):
+        return x * th.sigmoid(x)
+
+
+class GroupNorm32(nn.GroupNorm):
+    def forward(self, x):
+        return super().forward(x.float()).type(x.dtype)
+
+
+def conv_nd(dims, *args, **kwargs):
+    """
+    Create a 1D, 2D, or 3D convolution module.
+    """
+    if dims == 1:
+        return nn.Conv1d(*args, **kwargs)
+    elif dims == 2:
+        return nn.Conv2d(*args, **kwargs)
+    elif dims == 3:
+        return nn.Conv3d(*args, **kwargs)
+    raise ValueError(f"unsupported dimensions: {dims}")
+
+
+def linear(*args, **kwargs):
+    """
+    Create a linear module.
+    """
+    return nn.Linear(*args, **kwargs)
+
+
+def avg_pool_nd(dims, *args, **kwargs):
+    """
+    Create a 1D, 2D, or 3D average pooling module.
+    """
+    if dims == 1:
+        return nn.AvgPool1d(*args, **kwargs)
+    elif dims == 2:
+        return nn.AvgPool2d(*args, **kwargs)
+    elif dims == 3:
+        return nn.AvgPool3d(*args, **kwargs)
+    raise ValueError(f"unsupported dimensions: {dims}")
+
+
+def update_ema(target_params, source_params, rate=0.99):
+    """
+    Update target parameters to be closer to those of source parameters using
+    an exponential moving average.
+
+    :param target_params: the target parameter sequence.
+    :param source_params: the source parameter sequence.
+    :param rate: the EMA rate (closer to 1 means slower).
+    """
+    for targ, src in zip(target_params, source_params):
+        targ.detach().mul_(rate).add_(src, alpha=1 - rate)
+
+
+def zero_module(module):
+    """
+    Zero out the parameters of a module and return it.
+    """
+    for p in module.parameters():
+        p.detach().zero_()
+    return module
+
+
+def scale_module(module, scale):
+    """
+    Scale the parameters of a module and return it.
+    """
+    for p in module.parameters():
+        p.detach().mul_(scale)
+    return module
+
+
+def mean_flat(tensor):
+    """
+    Take the mean over all non-batch dimensions.
+    """
+    return tensor.mean(dim=list(range(1, len(tensor.shape))))
+
+
+def normalization(channels):
+    """
+    Make a standard normalization layer.
+
+    :param channels: number of input channels.
+    :return: an nn.Module for normalization.
+    """
+    return GroupNorm32(32, channels)
+
+
+def timestep_embedding(timesteps, dim, max_period=10000):
+    """
+    Create sinusoidal timestep embeddings.
+
+    :param timesteps: a 1-D Tensor of N indices, one per batch element.
+                      These may be fractional.
+    :param dim: the dimension of the output.
+    :param max_period: controls the minimum frequency of the embeddings.
+    :return: an [N x dim] Tensor of positional embeddings.
+    """
+    half = dim // 2
+    freqs = th.exp(
+        -math.log(max_period) * th.arange(start=0, end=half, dtype=th.float32) / half
+    ).to(device=timesteps.device)
+    args = timesteps[:, None].float() * freqs[None]
+    embedding = th.cat([th.cos(args), th.sin(args)], dim=-1)
+    if dim % 2:
+        embedding = th.cat([embedding, th.zeros_like(embedding[:, :1])], dim=-1)
+    return embedding
+
+
+def checkpoint(func, inputs, params, flag):
+    """
+    Evaluate a function without caching intermediate activations, allowing for
+    reduced memory at the expense of extra compute in the backward pass.
+
+    :param func: the function to evaluate.
+    :param inputs: the argument sequence to pass to `func`.
+    :param params: a sequence of parameters `func` depends on but does not
+                   explicitly take as arguments.
+    :param flag: if False, disable gradient checkpointing.
+    """
+    if flag:
+        args = tuple(inputs) + tuple(params)
+        return CheckpointFunction.apply(func, len(inputs), *args)
+    else:
+        return func(*inputs)
+
+
+class CheckpointFunction(th.autograd.Function):
+    @staticmethod
+    def forward(ctx, run_function, length, *args):
+        ctx.run_function = run_function
+        ctx.input_tensors = list(args[:length])
+        ctx.input_params = list(args[length:])
+        with th.no_grad():
+            output_tensors = ctx.run_function(*ctx.input_tensors)
+        return output_tensors
+
+    @staticmethod
+    def backward(ctx, *output_grads):
+        ctx.input_tensors = [x.detach().requires_grad_(True) for x in ctx.input_tensors]
+        with th.enable_grad():
+            # Fixes a bug where the first op in run_function modifies the
+            # Tensor storage in place, which is not allowed for detach()'d
+            # Tensors.
+            shallow_copies = [x.view_as(x) for x in ctx.input_tensors]
+            output_tensors = ctx.run_function(*shallow_copies)
+        input_grads = th.autograd.grad(
+            output_tensors,
+            ctx.input_tensors + ctx.input_params,
+            output_grads,
+            allow_unused=True,
+        )
+        del ctx.input_tensors
+        del ctx.input_params
+        del output_tensors
+        return (None, None) + input_grads
+
+
+def convert_module_to_f16(l):
+    """
+    Convert primitive modules to float16.
+    """
+    if isinstance(l, (nn.Conv1d, nn.Conv2d, nn.Conv3d)):
+        l.weight.data = l.weight.data.half()
+        if l.bias is not None:
+            l.bias.data = l.bias.data.half()
+
+
+def convert_module_to_f32(l):
+    """
+    Convert primitive modules to float32, undoing convert_module_to_f16().
+    """
+    if isinstance(l, (nn.Conv1d, nn.Conv2d, nn.Conv3d)):
+        l.weight.data = l.weight.data.float()
+        if l.bias is not None:
+            l.bias.data = l.bias.data.float()