import random import torch import torch.nn as nn import torch.nn.functional as F from torch import autocast from x_transformers import Encoder from models.diffusion.nn import timestep_embedding, normalization, zero_module, conv_nd, linear from models.diffusion.unet_diffusion import AttentionBlock, TimestepEmbedSequential, \ Downsample, Upsample, TimestepBlock from models.audio.tts.mini_encoder import AudioMiniEncoder from models.audio.tts.unet_diffusion_tts7 import CheckpointedXTransformerEncoder from scripts.audio.gen.use_diffuse_tts import ceil_multiple from trainer.networks import register_model from utils.util import checkpoint def is_sequence(t): return t.dtype == torch.long class ResBlock(TimestepBlock): def __init__( self, channels, emb_channels, dropout, out_channels=None, dims=2, kernel_size=3, efficient_config=True, use_scale_shift_norm=False, ): super().__init__() self.channels = channels self.emb_channels = emb_channels self.dropout = dropout self.out_channels = out_channels or channels self.use_scale_shift_norm = use_scale_shift_norm padding = {1: 0, 3: 1, 5: 2}[kernel_size] eff_kernel = 1 if efficient_config else 3 eff_padding = 0 if efficient_config else 1 self.in_layers = nn.Sequential( normalization(channels), nn.SiLU(), conv_nd(dims, channels, self.out_channels, eff_kernel, padding=eff_padding), ) 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, kernel_size, padding=padding) ), ) if self.out_channels == channels: self.skip_connection = nn.Identity() else: self.skip_connection = conv_nd(dims, channels, self.out_channels, eff_kernel, padding=eff_padding) 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 ) def _forward(self, x, emb): 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 = torch.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 DiffusionWaveformGen(nn.Module): """ The full UNet model with attention and timestep embedding. Customized to be conditioned on an aligned prior derived from a autoregressive GPT-style model. :param in_channels: channels in the input Tensor. :param in_latent_channels: channels from the input latent. :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_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, model_channels, in_channels=1, in_latent_channels=1024, in_mel_channels=120, conditioning_dim_factor=8, conditioning_expansion=4, out_channels=2, # mean and variance dropout=0, # res 1, 2, 4, 8,16,32,64,128,256,512, 1K, 2K channel_mult= (1,1.5,2, 3, 4, 6, 8, 12, 16, 24, 32, 48), num_res_blocks=(1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2), # spec_cond: 1, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0) # attn: 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1 token_conditioning_resolutions=(1,16,), attention_resolutions=(512,1024,2048), conv_resample=True, dims=1, use_fp16=False, num_heads=1, num_head_channels=-1, num_heads_upsample=-1, kernel_size=3, scale_factor=2, time_embed_dim_multiplier=4, freeze_main_net=False, efficient_convs=True, # Uses kernels with width of 1 in several places rather than 3. use_scale_shift_norm=True, # Parameters for regularization. unconditioned_percentage=.1, # This implements a mechanism similar to what is used in classifier-free training. # Parameters for super-sampling. super_sampling=False, super_sampling_max_noising_factor=.1, ): super().__init__() if num_heads_upsample == -1: num_heads_upsample = num_heads if super_sampling: in_channels *= 2 # In super-sampling mode, the LR input is concatenated directly onto the input. self.in_channels = in_channels self.model_channels = model_channels self.out_channels = out_channels self.attention_resolutions = attention_resolutions self.dropout = dropout self.channel_mult = channel_mult self.conv_resample = conv_resample self.num_heads = num_heads self.num_head_channels = num_head_channels self.num_heads_upsample = num_heads_upsample self.dims = dims self.super_sampling_enabled = super_sampling self.super_sampling_max_noising_factor = super_sampling_max_noising_factor self.unconditioned_percentage = unconditioned_percentage self.enable_fp16 = use_fp16 self.alignment_size = 2 ** (len(channel_mult)+1) self.freeze_main_net = freeze_main_net self.in_mel_channels = in_mel_channels padding = 1 if kernel_size == 3 else 2 down_kernel = 1 if efficient_convs else 3 time_embed_dim = model_channels * time_embed_dim_multiplier self.time_embed = nn.Sequential( linear(model_channels, time_embed_dim), nn.SiLU(), linear(time_embed_dim, time_embed_dim), ) conditioning_dim = model_channels * conditioning_dim_factor # Either code_converter or latent_converter is used, depending on what type of conditioning data is fed. # This model is meant to be able to be trained on both for efficiency purposes - it is far less computationally # complex to generate tokens, while generating latents will normally mean propagating through a deep autoregressive # transformer network. self.mel_converter = nn.Sequential( nn.Conv1d(in_mel_channels, conditioning_dim, 3, padding=1), CheckpointedXTransformerEncoder( needs_permute=True, max_seq_len=-1, attn_layers=Encoder( dim=conditioning_dim, depth=3, heads=num_heads, ff_dropout=dropout, attn_dropout=dropout, use_rmsnorm=True, ff_glu=True, rotary_pos_emb=True, ) )) self.latent_converter = nn.Conv1d(in_latent_channels, conditioning_dim, 1) self.aligned_latent_padding_embedding = nn.Parameter(torch.randn(1,in_latent_channels,1)) self.unconditioned_embedding = nn.Parameter(torch.randn(1,conditioning_dim,1)) self.conditioning_timestep_integrator = TimestepEmbedSequential( ResBlock(conditioning_dim, time_embed_dim, dropout, out_channels=conditioning_dim, dims=dims, kernel_size=1, use_scale_shift_norm=use_scale_shift_norm), AttentionBlock(conditioning_dim, num_heads=num_heads, num_head_channels=num_head_channels), ResBlock(conditioning_dim, time_embed_dim, dropout, out_channels=conditioning_dim, dims=dims, kernel_size=1, use_scale_shift_norm=use_scale_shift_norm), AttentionBlock(conditioning_dim, num_heads=num_heads, num_head_channels=num_head_channels), ResBlock(conditioning_dim, time_embed_dim, dropout, out_channels=conditioning_dim, dims=dims, kernel_size=1, use_scale_shift_norm=use_scale_shift_norm), ) self.conditioning_expansion = conditioning_expansion self.input_blocks = nn.ModuleList( [ TimestepEmbedSequential( conv_nd(dims, in_channels, model_channels, kernel_size, padding=padding) ) ] ) token_conditioning_blocks = [] self._feature_size = model_channels input_block_chans = [model_channels] ch = model_channels ds = 1 for level, (mult, num_blocks) in enumerate(zip(channel_mult, num_res_blocks)): if ds in token_conditioning_resolutions: token_conditioning_block = nn.Conv1d(conditioning_dim, ch, 1) token_conditioning_block.weight.data *= .02 self.input_blocks.append(token_conditioning_block) token_conditioning_blocks.append(token_conditioning_block) for _ in range(num_blocks): layers = [ ResBlock( ch, time_embed_dim, dropout, out_channels=int(mult * model_channels), dims=dims, kernel_size=kernel_size, efficient_config=efficient_convs, use_scale_shift_norm=use_scale_shift_norm, ) ] ch = int(mult * model_channels) if ds in attention_resolutions: layers.append( AttentionBlock( ch, num_heads=num_heads, num_head_channels=num_head_channels, ) ) 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( Downsample( ch, conv_resample, dims=dims, out_channels=out_ch, factor=scale_factor, ksize=down_kernel, pad=0 if down_kernel == 1 else 1 ) ) ) 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, kernel_size=kernel_size, efficient_config=efficient_convs, use_scale_shift_norm=use_scale_shift_norm, ), AttentionBlock( ch, num_heads=num_heads, num_head_channels=num_head_channels, ), ResBlock( ch, time_embed_dim, dropout, dims=dims, kernel_size=kernel_size, efficient_config=efficient_convs, use_scale_shift_norm=use_scale_shift_norm, ), ) self._feature_size += ch self.output_blocks = nn.ModuleList([]) for level, (mult, num_blocks) in list(enumerate(zip(channel_mult, num_res_blocks)))[::-1]: for i in range(num_blocks + 1): ich = input_block_chans.pop() layers = [ ResBlock( ch + ich, time_embed_dim, dropout, out_channels=int(model_channels * mult), dims=dims, kernel_size=kernel_size, efficient_config=efficient_convs, use_scale_shift_norm=use_scale_shift_norm, ) ] ch = int(model_channels * mult) if ds in attention_resolutions: layers.append( AttentionBlock( ch, num_heads=num_heads_upsample, num_head_channels=num_head_channels, ) ) if level and i == num_blocks: out_ch = ch layers.append( Upsample(ch, conv_resample, dims=dims, out_channels=out_ch, factor=scale_factor) ) 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, kernel_size, padding=padding)), ) if self.freeze_main_net: mains = [self.time_embed, self.contextual_embedder, self.unconditioned_embedding, self.conditioning_timestep_integrator, self.input_blocks, self.middle_block, self.output_blocks, self.out] for m in mains: for p in m.parameters(): p.requires_grad = False p.DO_NOT_TRAIN = True def get_grad_norm_parameter_groups(self): if self.freeze_main_net: return {} groups = { 'input_blocks': list(self.input_blocks.parameters()), 'output_blocks': list(self.output_blocks.parameters()), 'middle_transformer': list(self.middle_block.parameters()), } return groups def is_latent(self, t): return t.shape[1] != self.in_mel_channels def fix_alignment(self, x, aligned_conditioning): """ The UNet requires that the input is a certain multiple of 2, defined by the UNet depth. Enforce this by padding both and before forward propagation and removing the padding before returning. """ cm = ceil_multiple(x.shape[-1], self.alignment_size) if cm != 0: pc = (cm-x.shape[-1])/x.shape[-1] x = F.pad(x, (0,cm-x.shape[-1])) # Also fix aligned_latent, which is aligned to x. if self.is_latent(aligned_conditioning): aligned_conditioning = torch.cat([aligned_conditioning, self.aligned_latent_padding_embedding.repeat(x.shape[0], 1, int(pc * aligned_conditioning.shape[-1]))], dim=-1) else: aligned_conditioning = F.pad(aligned_conditioning, (0,int(pc*aligned_conditioning.shape[-1]))) return x, aligned_conditioning def forward(self, x, timesteps, aligned_conditioning, conditioning_free=False): """ 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 aligned_conditioning: an aligned latent or sequence of tokens providing useful data about the sample to be produced. :param conditioning_free: When set, all conditioning inputs (including tokens and conditioning_input) will not be considered. :return: an [N x C x ...] Tensor of outputs. """ # Shuffle aligned_latent to BxCxS format if self.is_latent(aligned_conditioning): aligned_conditioning = aligned_conditioning.permute(0, 2, 1) # Fix input size to the proper multiple of 2 so we don't get alignment errors going down and back up the U-net. orig_x_shape = x.shape[-1] x, aligned_conditioning = self.fix_alignment(x, aligned_conditioning) with autocast(x.device.type, enabled=self.enable_fp16): hs = [] time_emb = self.time_embed(timestep_embedding(timesteps, self.model_channels)) # Note: this block does not need to repeated on inference, since it is not timestep-dependent. if conditioning_free: code_emb = self.unconditioned_embedding.repeat(x.shape[0], 1, 1) else: if self.is_latent(aligned_conditioning): code_emb = self.latent_converter(aligned_conditioning) else: code_emb = self.mel_converter(aligned_conditioning) # Everything after this comment is timestep dependent. code_emb = torch.repeat_interleave(code_emb, self.conditioning_expansion, dim=-1) code_emb = self.conditioning_timestep_integrator(code_emb, time_emb) first = True time_emb = time_emb.float() h = x for k, module in enumerate(self.input_blocks): if isinstance(module, nn.Conv1d): h_tok = F.interpolate(module(code_emb), size=(h.shape[-1]), mode='nearest') h = h + h_tok else: with autocast(x.device.type, enabled=self.enable_fp16 and not first): # First block has autocast disabled to allow a high precision signal to be properly vectorized. h = module(h, time_emb) hs.append(h) first = False h = self.middle_block(h, time_emb) for module in self.output_blocks: h = torch.cat([h, hs.pop()], dim=1) h = module(h, time_emb) # Last block also has autocast disabled for high-precision outputs. h = h.float() out = self.out(h) # Involve probabilistic or possibly unused parameters in loss so we don't get DDP errors. extraneous_addition = 0 params = [self.aligned_latent_padding_embedding, self.unconditioned_embedding] + list(self.latent_converter.parameters()) for p in params: extraneous_addition = extraneous_addition + p.mean() out = out + extraneous_addition * 0 return out[:, :, :orig_x_shape] @register_model def register_unet_diffusion_waveform_gen(opt_net, opt): return DiffusionWaveformGen(**opt_net['kwargs']) if __name__ == '__main__': clip = torch.randn(2, 1, 32868) aligned_latent = torch.randn(2,388,1024) aligned_sequence = torch.randn(2,120,220) ts = torch.LongTensor([600, 600]) model = DiffusionWaveformGen(128, channel_mult=[1,1.5,2, 3, 4, 6, 8], num_res_blocks=[2, 2, 2, 2, 2, 2, 1], token_conditioning_resolutions=[1,4,16,64], attention_resolutions=[], num_heads=8, kernel_size=3, scale_factor=2, time_embed_dim_multiplier=4, super_sampling=False, efficient_convs=False) # Test with latent aligned conditioning o = model(clip, ts, aligned_latent) # Test with sequence aligned conditioning o = model(clip, ts, aligned_sequence)