blakelemoine/DL-Art-School
1
0
Fork 0
forked from mrq/DL-Art-School
Code Pull Requests Activity

Make it even better

Browse Source
This commit is contained in:
James Betker 2022-03-21 14:50:59 -06:00
parent e735d8e1fa
commit 723f324eda
1 changed files with 39 additions and 34 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

73
codes/models/audio/tts/unet_diffusion_tts_flat.py
View File

@ -76,28 +76,26 @@ class DiffusionTtsFlat(nn.Module):
) )
) )
self.latent_converter = nn.Conv1d(in_latent_channels, model_channels, 1) self.latent_converter = nn.Conv1d(in_latent_channels, model_channels, 1)
if in_channels > 60: # It's a spectrogram. # The contextual embedder processes a sample MEL that the output should be "like".
self.contextual_embedder = nn.Sequential(nn.Conv1d(in_channels,model_channels,3,padding=1,stride=2), self.contextual_embedder = nn.Sequential(nn.Conv1d(in_channels,model_channels,3,padding=1,stride=2),
CheckpointedXTransformerEncoder( CheckpointedXTransformerEncoder(
needs_permute=True, needs_permute=True,
max_seq_len=-1, max_seq_len=-1,
use_pos_emb=False, use_pos_emb=False,
attn_layers=Encoder( attn_layers=Encoder(
dim=model_channels, dim=model_channels,
depth=4, depth=4,
heads=num_heads, heads=num_heads,
ff_dropout=dropout, ff_dropout=dropout,
attn_dropout=dropout, attn_dropout=dropout,
use_rmsnorm=True, use_rmsnorm=True,
ff_glu=True, ff_glu=True,
rotary_emb_dim=True, rotary_emb_dim=True,
) )
)) ))
else:
self.contextual_embedder = AudioMiniEncoder(1, model_channels, base_channels=32, depth=6, resnet_blocks=1,
attn_blocks=3, num_attn_heads=8, dropout=dropout, downsample_factor=4, kernel_size=5)
self.conditioning_conv = nn.Conv1d(model_channels*2, model_channels, 1) self.conditioning_conv = nn.Conv1d(model_channels*2, model_channels, 1)
self.unconditioned_embedding = nn.Parameter(torch.randn(1,model_channels,1)) self.unconditioned_embedding = nn.Parameter(torch.randn(1,model_channels,1))
# This is a further encoder extension that integrates a timestep signal into the conditioning signal.
self.conditioning_timestep_integrator = CheckpointedXTransformerEncoder( self.conditioning_timestep_integrator = CheckpointedXTransformerEncoder(
needs_permute=True, needs_permute=True,
max_seq_len=-1, max_seq_len=-1,
@ -117,6 +115,7 @@ class DiffusionTtsFlat(nn.Module):
) )
self.integrate_conditioning = nn.Conv1d(model_channels*2, model_channels, 1) self.integrate_conditioning = nn.Conv1d(model_channels*2, model_channels, 1)
# This is the main processing module.
self.layers = CheckpointedXTransformerEncoder( self.layers = CheckpointedXTransformerEncoder(
needs_permute=True, needs_permute=True,
max_seq_len=-1, max_seq_len=-1,
@ -151,18 +150,7 @@ class DiffusionTtsFlat(nn.Module):
} }
return groups return groups
def forward(self, x, timesteps, aligned_conditioning, conditioning_input, conditioning_free=False): def get_conditioning_encodings(self, aligned_conditioning, conditioning_input, conditioning_free, return_unused=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_input: a full-resolution audio clip that is used as a reference to the style you want decoded.
:param lr_input: for super-sampling models, a guidance audio clip at a lower sampling rate.
: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 # Shuffle aligned_latent to BxCxS format
if is_latent(aligned_conditioning): if is_latent(aligned_conditioning):
aligned_conditioning = aligned_conditioning.permute(0, 2, 1) aligned_conditioning = aligned_conditioning.permute(0, 2, 1)
@ -184,14 +172,31 @@ class DiffusionTtsFlat(nn.Module):
unused_params.extend(list(self.latent_converter.parameters())) unused_params.extend(list(self.latent_converter.parameters()))
cond_emb_spread = cond_emb.unsqueeze(-1).repeat(1, 1, code_emb.shape[-1]) cond_emb_spread = cond_emb.unsqueeze(-1).repeat(1, 1, code_emb.shape[-1])
code_emb = self.conditioning_conv(torch.cat([cond_emb_spread, code_emb], dim=1)) code_emb = self.conditioning_conv(torch.cat([cond_emb_spread, code_emb], dim=1))
# Mask out the conditioning branch for whole batch elements, implementing something similar to classifier-free guidance. # Mask out the conditioning branch for whole batch elements, implementing something similar to classifier-free guidance.
if self.training and self.unconditioned_percentage > 0: if self.training and self.unconditioned_percentage > 0:
unconditioned_batches = torch.rand((code_emb.shape[0], 1, 1), unconditioned_batches = torch.rand((code_emb.shape[0], 1, 1),
device=code_emb.device) < self.unconditioned_percentage device=code_emb.device) < self.unconditioned_percentage
code_emb = torch.where(unconditioned_batches, self.unconditioned_embedding.repeat(x.shape[0], 1, 1), code_emb = torch.where(unconditioned_batches, self.unconditioned_embedding.repeat(conditioning_input.shape[0], 1, 1),
code_emb) code_emb)
# Everything after this comment is timestep dependent. if return_unused:
return code_emb, unused_params
return code_emb
def forward(self, x, timesteps, aligned_conditioning, conditioning_input, 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_input: a full-resolution audio clip that is used as a reference to the style you want decoded.
: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.
"""
code_emb, unused_params = self.get_conditioning_encodings(aligned_conditioning, conditioning_input, conditioning_free, return_unused=True)
# Everything after this comment is timestep-dependent.
time_emb = self.time_embed(timestep_embedding(timesteps, self.model_channels)) time_emb = self.time_embed(timestep_embedding(timesteps, self.model_channels))
code_emb = self.conditioning_timestep_integrator(code_emb, time_emb=time_emb) code_emb = self.conditioning_timestep_integrator(code_emb, time_emb=time_emb)
x = self.inp_block(x) x = self.inp_block(x)