from abc import abstractmethod import math import numpy as np import torch as th import torch.nn as nn import torch.nn.functional as F from pdb import set_trace as st from einops import rearrange, repeat from .fp16_util import convert_module_to_f16, convert_module_to_f32 from .nn import ( checkpoint, conv_nd, linear, avg_pool_nd, zero_module, normalization, timestep_embedding, ) from ldm.modules.attention_compat import SpatialTransformer 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 # from LDM openaimodel.py 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, context=None): for layer in self: if isinstance(layer, TimestepBlock): x = layer(x, emb) elif isinstance(layer, SpatialTransformer): x = layer(x, context) 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 # ) return self._forward(x, emb) 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) # ! disable checkpoint here since it is incompatible with torch.amp 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) class UNetModel(nn.Module): """ The full UNet model with attention and timestep embedding. :param in_channels: channels in the input Tensor. :param model_channels: base channel count for the model. :param out_channels: channels in the output Tensor. :param num_res_blocks: number of residual blocks per downsample. :param attention_resolutions: a collection of downsample rates at which attention will take place. May be a set, list, or tuple. For example, if this contains 4, then at 4x downsampling, attention will be used. :param dropout: the dropout probability. :param channel_mult: channel multiplier for each level of the UNet. :param conv_resample: if True, use learned convolutions for upsampling and downsampling. :param dims: determines if the signal is 1D, 2D, or 3D. :param num_classes: if specified (as an int), then this model will be class-conditional with `num_classes` classes. :param use_checkpoint: use gradient checkpointing to reduce memory usage. :param num_heads: the number of attention heads in each attention layer. :param num_heads_channels: if specified, ignore num_heads and instead use a fixed channel width per attention head. :param num_heads_upsample: works with num_heads to set a different number of heads for upsampling. Deprecated. :param use_scale_shift_norm: use a FiLM-like conditioning mechanism. :param resblock_updown: use residual blocks for up/downsampling. :param use_new_attention_order: use a different attention pattern for potentially increased efficiency. """ def __init__( self, image_size, in_channels, model_channels, out_channels, num_res_blocks, attention_resolutions, dropout=0, channel_mult=(1, 2, 4, 8), conv_resample=True, dims=2, num_classes=None, use_checkpoint=False, use_fp16=False, num_heads=-1, num_head_channels=-1, num_heads_upsample=-1, use_scale_shift_norm=False, resblock_updown=False, use_new_attention_order=False, mixed_prediction=False, use_spatial_transformer=False, # custom transformer support transformer_depth=1, # custom transformer support context_dim=-1, # custom transformer support n_embed=None, # custom support for prediction of discrete ids into codebook of first stage vq model legacy=True, mixing_logit_init=-6, roll_out=False,**kwargs ): super().__init__() self.roll_out = roll_out if context_dim == -1: context_dim = None if use_spatial_transformer: assert context_dim is not None, 'Fool!! You forgot to include the dimension of your cross-attention conditioning...' if context_dim is not None: assert use_spatial_transformer, 'Fool!! You forgot to use the spatial transformer for your cross-attention conditioning...' # from omegaconf.listconfig import ListConfig # if type(context_dim) == ListConfig: # context_dim = list(context_dim) if num_heads_upsample == -1: num_heads_upsample = num_heads if num_heads == -1: assert num_head_channels != -1, 'Either num_heads or num_head_channels has to be set' if num_head_channels == -1: assert num_heads != -1, 'Either num_heads or num_head_channels has to be set' 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 self.predict_codebook_ids = n_embed is not None # follow LSGM self.mixed_prediction = mixed_prediction # This enables mixed prediction if self.mixed_prediction: if self.roll_out: init = mixing_logit_init * th.ones(size=[1, in_channels*3, 1, 1]) # hard coded for now else: init = mixing_logit_init * th.ones(size=[1, in_channels, 1, 1]) # hard coded for now self.mixing_logit = th.nn.Parameter(init, requires_grad=True) 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) self.input_blocks = nn.ModuleList( [ TimestepEmbedSequential( conv_nd(dims, in_channels, model_channels, 3, padding=1) ) ] ) self._feature_size = model_channels input_block_chans = [model_channels] ch = model_channels ds = 1 for level, mult in enumerate(channel_mult): for _ in range(num_res_blocks): layers = [ ResBlock( ch, time_embed_dim, dropout, out_channels=mult * model_channels, dims=dims, use_checkpoint=use_checkpoint, use_scale_shift_norm=use_scale_shift_norm, ) ] ch = mult * model_channels if ds in attention_resolutions: if num_head_channels == -1: dim_head = ch // num_heads else: num_heads = ch // num_head_channels dim_head = num_head_channels if legacy: #num_heads = 1 dim_head = ch // num_heads if use_spatial_transformer else num_head_channels layers.append( AttentionBlock( ch, use_checkpoint=use_checkpoint, num_heads=num_heads, num_head_channels=dim_head, use_new_attention_order=use_new_attention_order, ) if not use_spatial_transformer else SpatialTransformer( ch, num_heads, dim_head, depth=transformer_depth, context_dim=context_dim ) ) 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 if num_head_channels == -1: dim_head = ch // num_heads else: num_heads = ch // num_head_channels dim_head = num_head_channels if legacy: #num_heads = 1 dim_head = ch // num_heads if use_spatial_transformer else num_head_channels 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=dim_head, use_new_attention_order=use_new_attention_order, ) if not use_spatial_transformer else SpatialTransformer( ch, num_heads, dim_head, depth=transformer_depth, context_dim=context_dim ), 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=model_channels * mult, dims=dims, use_checkpoint=use_checkpoint, use_scale_shift_norm=use_scale_shift_norm, ) ] ch = model_channels * mult if ds in attention_resolutions: if num_head_channels == -1: dim_head = ch // num_heads else: num_heads = ch // num_head_channels dim_head = num_head_channels if legacy: #num_heads = 1 dim_head = ch // num_heads if use_spatial_transformer else num_head_channels layers.append( AttentionBlock( ch, use_checkpoint=use_checkpoint, num_heads=num_heads_upsample, num_head_channels=dim_head, use_new_attention_order=use_new_attention_order, ) if not use_spatial_transformer else SpatialTransformer( ch, num_heads, dim_head, depth=transformer_depth, context_dim=context_dim ) ) 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, model_channels, out_channels, 3, padding=1)), ) if self.predict_codebook_ids: self.id_predictor = nn.Sequential( normalization(ch), conv_nd(dims, model_channels, n_embed, 1), #nn.LogSoftmax(dim=1) # change to cross_entropy and produce non-normalized logits ) 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, timesteps=None, context=None, y=None, get_attr='', **kwargs): """ Apply the model to an input batch. :param x: an [N x C x ...] Tensor of inputs. :param timesteps: a 1-D batch of timesteps. :param context: conditioning plugged in via crossattn :param y: an [N] Tensor of labels, if class-conditional. :return: an [N x C x ...] Tensor of outputs. """ if isinstance(context, dict): context = context['crossattn'] # sgm conditioner compat if get_attr != '': # not breaking the forward hooks return getattr(self, get_attr) # if forward # assert context is not None assert (y is not None) == ( self.num_classes is not None ), "must specify y if and only if the model is class-conditional" hs = [] # st() t_emb = timestep_embedding(timesteps, self.model_channels, repeat_only=False) emb = self.time_embed(t_emb) if self.roll_out: x = rearrange(x, 'b (n c) h w->b c h (n w)', n=3) # torch.Size([84, 4, 32, 96]) 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, context) hs.append(h) h = self.middle_block(h, emb, context) for module in self.output_blocks: h = th.cat([h, hs.pop()], dim=1) h = module(h, emb, context) h = h.type(x.dtype) if self.predict_codebook_ids: return self.id_predictor(h) else: h = self.out(h) if self.roll_out: return rearrange(h, 'b c h (n w) -> b (n c) h w', n=3) return h class SuperResModel(UNetModel): """ A UNetModel that performs super-resolution. Expects an extra kwarg `low_res` to condition on a low-resolution image. """ def __init__(self, image_size, in_channels, *args, **kwargs): super().__init__(image_size, in_channels * 2, *args, **kwargs) def forward(self, x, timesteps, low_res=None, **kwargs): _, _, new_height, new_width = x.shape upsampled = F.interpolate(low_res, (new_height, new_width), mode="bilinear") x = th.cat([x, upsampled], dim=1) return super().forward(x, timesteps, **kwargs) class EncoderUNetModel(nn.Module): """ The half UNet model with attention and timestep embedding. For usage, see UNet. """ def __init__( self, image_size, in_channels, model_channels, out_channels, num_res_blocks, attention_resolutions, dropout=0, channel_mult=(1, 2, 4, 8), conv_resample=True, dims=2, use_checkpoint=False, use_fp16=False, num_heads=1, num_head_channels=-1, num_heads_upsample=-1, use_scale_shift_norm=False, resblock_updown=False, use_new_attention_order=False, pool="adaptive", ): super().__init__() if num_heads_upsample == -1: num_heads_upsample = num_heads 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.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), ) 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.pool = pool if pool == "adaptive": self.out = nn.Sequential( normalization(ch), nn.SiLU(), nn.AdaptiveAvgPool2d((1, 1)), zero_module(conv_nd(dims, ch, out_channels, 1)), nn.Flatten(), ) elif pool == "attention": assert num_head_channels != -1 self.out = nn.Sequential( normalization(ch), nn.SiLU(), AttentionPool2d( (image_size // ds), ch, num_head_channels, out_channels ), ) elif pool == "spatial": self.out = nn.Sequential( nn.Linear(self._feature_size, 2048), nn.ReLU(), nn.Linear(2048, self.out_channels), ) elif pool == "spatial_v2": self.out = nn.Sequential( nn.Linear(self._feature_size, 2048), normalization(2048), nn.SiLU(), nn.Linear(2048, self.out_channels), ) else: raise NotImplementedError(f"Unexpected {pool} pooling") 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) 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) def forward(self, x, timesteps): """ Apply the model to an input batch. :param x: an [N x C x ...] Tensor of inputs. :param timesteps: a 1-D batch of timesteps. :return: an [N x K] Tensor of outputs. """ emb = self.time_embed(timestep_embedding(timesteps, self.model_channels)) results = [] h = x.type(self.dtype) for module in self.input_blocks: h = module(h, emb) if self.pool.startswith("spatial"): results.append(h.type(x.dtype).mean(dim=(2, 3))) h = self.middle_block(h, emb) if self.pool.startswith("spatial"): results.append(h.type(x.dtype).mean(dim=(2, 3))) h = th.cat(results, axis=-1) return self.out(h) else: h = h.type(x.dtype) return self.out(h) class UNetModelWithHint(UNetModel): def __init__(self, image_size, in_channels, model_channels, hint_channels, out_channels, num_res_blocks, attention_resolutions, dropout=0, channel_mult=(1, 2, 4, 8), conv_resample=True, dims=2, num_classes=None, use_checkpoint=False, use_fp16=False, num_heads=-1, num_head_channels=-1, num_heads_upsample=-1, use_scale_shift_norm=False, resblock_updown=False, use_new_attention_order=False, mixed_prediction=False, use_spatial_transformer=False, transformer_depth=1, context_dim=-1, n_embed=None, legacy=True, mixing_logit_init=-6, roll_out=False): super().__init__(image_size, in_channels, model_channels, out_channels, num_res_blocks, attention_resolutions, dropout, channel_mult, conv_resample, dims, num_classes, use_checkpoint, use_fp16, num_heads, num_head_channels, num_heads_upsample, use_scale_shift_norm, resblock_updown, use_new_attention_order, mixed_prediction, use_spatial_transformer, transformer_depth, context_dim, n_embed, legacy, mixing_logit_init, roll_out) # lite encoder, borrowed from ControlNet self.input_hint_block = TimestepEmbedSequential( # f=8 conv_nd(dims, hint_channels, 16, 3, padding=1), nn.SiLU(), conv_nd(dims, 16, 16, 3, padding=1), nn.SiLU(), conv_nd(dims, 16, 32, 3, padding=1, stride=2), nn.SiLU(), conv_nd(dims, 32, 32, 3, padding=1), nn.SiLU(), conv_nd(dims, 32, 96, 3, padding=1, stride=2), nn.SiLU(), conv_nd(dims, 96, 96, 3, padding=1), nn.SiLU(), conv_nd(dims, 96, 256, 3, padding=1, stride=2), nn.SiLU(), zero_module(conv_nd(dims, 256, model_channels, 3, padding=1)) ) def forward(self, x, hint, timesteps=None, context=None, y=None, get_attr='', **kwargs): """ Apply the model to an input batch. :param x: an [N x C x ...] Tensor of inputs. :param timesteps: a 1-D batch of timesteps. :param context: conditioning plugged in via crossattn :param y: an [N] Tensor of labels, if class-conditional. :return: an [N x C x ...] Tensor of outputs. """ # st() # if forward # assert context is not None assert context is not None # assert (y is not None) == ( # self.num_classes is not None # ), "must specify y if and only if the model is class-conditional" hs = [] t_emb = timestep_embedding(timesteps, self.model_channels, repeat_only=False) emb = self.time_embed(t_emb) if self.roll_out: x = rearrange(x, 'b (n c) h w->b c h (n w)', n=3) # torch.Size([84, 4, 32, 96]) # if self.num_classes is not None: # assert y.shape == (x.shape[0],) # emb = emb + self.label_emb(y) guided_hint = self.input_hint_block(hint, emb, context) if self.roll_out: guided_hint = repeat(guided_hint, 'b c h w -> b c h (n w)', n=3) # torch.Size([84, 4, 32, 96]) h = x.type(self.dtype) for module in self.input_blocks: if guided_hint is not None: h = module(h, emb, context) # B, 320, 32, 96 h += guided_hint guided_hint = None else: h = module(h, emb, context) hs.append(h) h = self.middle_block(h, emb, context) for module in self.output_blocks: h = th.cat([h, hs.pop()], dim=1) h = module(h, emb, context) h = h.type(x.dtype) if self.predict_codebook_ids: return self.id_predictor(h) else: h = self.out(h) if self.roll_out: return rearrange(h, 'b c h (n w) -> b (n c) h w', n=3) return h