391 lines
16 KiB
Python
391 lines
16 KiB
Python
from models.diffusion.fp16_util import convert_module_to_f32, convert_module_to_f16
|
|
from models.diffusion.nn import timestep_embedding, normalization, zero_module, conv_nd, linear
|
|
from models.diffusion.unet_diffusion import AttentionPool2d, AttentionBlock, ResBlock, TimestepEmbedSequential, \
|
|
Downsample, Upsample
|
|
import torch
|
|
import torch.nn as nn
|
|
|
|
from models.gpt_voice.mini_encoder import AudioMiniEncoder, EmbeddingCombiner
|
|
from trainer.networks import register_model
|
|
from utils.util import get_mask_from_lengths
|
|
|
|
|
|
class DiscreteSpectrogramConditioningBlock(nn.Module):
|
|
def __init__(self, dvae_channels, channels, level):
|
|
super().__init__()
|
|
self.intg = nn.Sequential(nn.Conv1d(dvae_channels, channels, kernel_size=1),
|
|
normalization(channels),
|
|
nn.SiLU(),
|
|
nn.Conv1d(channels, channels, kernel_size=3))
|
|
self.level = level
|
|
|
|
"""
|
|
Embeds the given codes and concatenates them onto x. Return shape is the same as x.shape.
|
|
|
|
:param x: bxcxS waveform latent
|
|
:param codes: bxN discrete codes, N <= S
|
|
"""
|
|
def forward(self, x, dvae_in):
|
|
b, c, S = x.shape
|
|
_, q, N = dvae_in.shape
|
|
emb = self.intg(dvae_in)
|
|
emb = nn.functional.interpolate(emb, size=(S,), mode='nearest')
|
|
return torch.cat([x, emb], dim=1)
|
|
|
|
|
|
class DiffusionVocoderWithRef(nn.Module):
|
|
"""
|
|
The full UNet model with attention and timestep embedding.
|
|
|
|
Customized to be conditioned on a spectrogram prior.
|
|
|
|
:param in_channels: channels in the input Tensor.
|
|
:param spectrogram_channels: channels in the conditioning spectrogram.
|
|
: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,
|
|
out_channels=2, # mean and variance
|
|
discrete_codes=512,
|
|
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
|
|
spectrogram_conditioning_resolutions=(512,),
|
|
attention_resolutions=(512,1024,2048),
|
|
conv_resample=True,
|
|
dims=1,
|
|
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,
|
|
kernel_size=3,
|
|
scale_factor=2,
|
|
conditioning_inputs_provided=True,
|
|
conditioning_input_dim=80,
|
|
time_embed_dim_multiplier=4,
|
|
freeze_layers_below=None, # powers of 2; ex: 1,2,4,8,16,32,etc..
|
|
):
|
|
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.attention_resolutions = attention_resolutions
|
|
self.dropout = dropout
|
|
self.channel_mult = channel_mult
|
|
self.conv_resample = conv_resample
|
|
self.dtype = torch.float16 if use_fp16 else torch.float32
|
|
self.num_heads = num_heads
|
|
self.num_head_channels = num_head_channels
|
|
self.num_heads_upsample = num_heads_upsample
|
|
self.dims = dims
|
|
|
|
padding = 1 if kernel_size == 3 else 2
|
|
|
|
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),
|
|
)
|
|
|
|
self.conditioning_enabled = conditioning_inputs_provided
|
|
if conditioning_inputs_provided:
|
|
self.contextual_embedder = AudioMiniEncoder(in_channels, time_embed_dim, base_channels=32, depth=6, resnet_blocks=1,
|
|
attn_blocks=2, num_attn_heads=2, dropout=dropout, downsample_factor=4, kernel_size=5)
|
|
|
|
seqlyr = TimestepEmbedSequential(
|
|
conv_nd(dims, in_channels, model_channels, kernel_size, padding=padding)
|
|
)
|
|
seqlyr.level = 0
|
|
self.input_blocks = nn.ModuleList([seqlyr])
|
|
spectrogram_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 spectrogram_conditioning_resolutions:
|
|
spec_cond_block = DiscreteSpectrogramConditioningBlock(discrete_codes, ch, 2 ** level)
|
|
self.input_blocks.append(spec_cond_block)
|
|
spectrogram_blocks.append(spec_cond_block)
|
|
ch *= 2
|
|
|
|
for _ in range(num_blocks):
|
|
layers = [
|
|
ResBlock(
|
|
ch,
|
|
time_embed_dim,
|
|
dropout,
|
|
out_channels=int(mult * model_channels),
|
|
dims=dims,
|
|
use_scale_shift_norm=use_scale_shift_norm,
|
|
kernel_size=kernel_size,
|
|
)
|
|
]
|
|
ch = int(mult * model_channels)
|
|
if ds in attention_resolutions:
|
|
layers.append(
|
|
AttentionBlock(
|
|
ch,
|
|
num_heads=num_heads,
|
|
num_head_channels=num_head_channels,
|
|
use_new_attention_order=use_new_attention_order,
|
|
)
|
|
)
|
|
layer = TimestepEmbedSequential(*layers)
|
|
layer.level = 2 ** level
|
|
self.input_blocks.append(layer)
|
|
self._feature_size += ch
|
|
input_block_chans.append(ch)
|
|
if level != len(channel_mult) - 1:
|
|
out_ch = ch
|
|
upblk = TimestepEmbedSequential(
|
|
ResBlock(
|
|
ch,
|
|
time_embed_dim,
|
|
dropout,
|
|
out_channels=out_ch,
|
|
dims=dims,
|
|
use_scale_shift_norm=use_scale_shift_norm,
|
|
down=True,
|
|
kernel_size=kernel_size,
|
|
)
|
|
if resblock_updown
|
|
else Downsample(
|
|
ch, conv_resample, dims=dims, out_channels=out_ch, factor=scale_factor
|
|
)
|
|
)
|
|
upblk.level = 2 ** level
|
|
self.input_blocks.append(upblk)
|
|
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_scale_shift_norm=use_scale_shift_norm,
|
|
kernel_size=kernel_size,
|
|
),
|
|
AttentionBlock(
|
|
ch,
|
|
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_scale_shift_norm=use_scale_shift_norm,
|
|
kernel_size=kernel_size,
|
|
),
|
|
)
|
|
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,
|
|
use_scale_shift_norm=use_scale_shift_norm,
|
|
kernel_size=kernel_size,
|
|
)
|
|
]
|
|
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,
|
|
use_new_attention_order=use_new_attention_order,
|
|
)
|
|
)
|
|
if level and i == num_blocks:
|
|
out_ch = ch
|
|
layers.append(
|
|
ResBlock(
|
|
ch,
|
|
time_embed_dim,
|
|
dropout,
|
|
out_channels=out_ch,
|
|
dims=dims,
|
|
use_scale_shift_norm=use_scale_shift_norm,
|
|
up=True,
|
|
kernel_size=kernel_size,
|
|
)
|
|
if resblock_updown
|
|
else Upsample(ch, conv_resample, dims=dims, out_channels=out_ch, factor=scale_factor)
|
|
)
|
|
ds //= 2
|
|
layer = TimestepEmbedSequential(*layers)
|
|
layer.level = 2 ** level
|
|
self.output_blocks.append(layer)
|
|
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 freeze_layers_below is not None:
|
|
# Freeze all parameters first.
|
|
for p in self.parameters():
|
|
p.DO_NOT_TRAIN = True
|
|
p.requires_grad = False
|
|
# Now un-freeze the modules we actually want to train.
|
|
unfrozen_modules = [self.out]
|
|
for blk in self.input_blocks:
|
|
if blk.level <= freeze_layers_below:
|
|
unfrozen_modules.append(blk)
|
|
last_frozen_output_block = None
|
|
for blk in self.output_blocks:
|
|
if blk.level <= freeze_layers_below:
|
|
unfrozen_modules.append(blk)
|
|
else:
|
|
last_frozen_output_block = blk
|
|
# And finally, the last upsample block in output blocks.
|
|
unfrozen_modules.append(last_frozen_output_block[1])
|
|
unfrozen_params = 0
|
|
for m in unfrozen_modules:
|
|
for p in m.parameters():
|
|
del p.DO_NOT_TRAIN
|
|
p.requires_grad = True
|
|
unfrozen_params += 1
|
|
print(f"freeze_layers_below specified. Training a total of {unfrozen_params} parameters.")
|
|
|
|
def forward(self, x, timesteps, spectrogram, conditioning_input=None):
|
|
"""
|
|
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 y: an [N] Tensor of labels, if class-conditional.
|
|
:return: an [N x C x ...] Tensor of outputs.
|
|
"""
|
|
assert x.shape[-1] % 2048 == 0 # This model operates at base//2048 at it's bottom levels, thus this requirement.
|
|
if self.conditioning_enabled:
|
|
assert conditioning_input is not None
|
|
|
|
hs = []
|
|
emb1 = self.time_embed(timestep_embedding(timesteps, self.model_channels))
|
|
if self.conditioning_enabled:
|
|
emb2 = self.contextual_embedder(conditioning_input)
|
|
emb = emb1 + emb2
|
|
else:
|
|
emb = emb1
|
|
|
|
h = x.type(self.dtype)
|
|
for k, module in enumerate(self.input_blocks):
|
|
if isinstance(module, DiscreteSpectrogramConditioningBlock):
|
|
h = module(h, spectrogram)
|
|
else:
|
|
h = module(h, emb)
|
|
hs.append(h)
|
|
h = self.middle_block(h, emb)
|
|
for module in self.output_blocks:
|
|
h = torch.cat([h, hs.pop()], dim=1)
|
|
h = module(h, emb)
|
|
h = h.type(x.dtype)
|
|
return self.out(h)
|
|
|
|
|
|
def move_all_layers_down(pretrained_path, output_path, layers_to_be_added=3):
|
|
# layers_to_be_added should be=num_res_blocks+1+[1if spectrogram_conditioning_resolutions;else0]
|
|
sd = torch.load(pretrained_path)
|
|
out = sd.copy()
|
|
replaced = []
|
|
for n, p in sd.items():
|
|
if n.startswith('input_blocks.') and not n.startswith('input_blocks.0.'):
|
|
if n not in replaced:
|
|
del out[n]
|
|
components = n.split('.')
|
|
components[1] = str(int(components[1]) + layers_to_be_added)
|
|
new_name = '.'.join(components)
|
|
out[new_name] = p
|
|
replaced.append(new_name)
|
|
torch.save(out, output_path)
|
|
|
|
|
|
@register_model
|
|
def register_unet_diffusion_vocoder_with_ref(opt_net, opt):
|
|
return DiffusionVocoderWithRef(**opt_net['kwargs'])
|
|
|
|
|
|
# Test for ~4 second audio clip at 22050Hz
|
|
if __name__ == '__main__':
|
|
path = 'X:\\dlas\\experiments\\train_diffusion_vocoder_with_cond_new_dvae_full\\models\\6100_generator_ema.pth'
|
|
move_all_layers_down(path, 'diffuse_new_lyr.pth', layers_to_be_added=2)
|
|
|
|
clip = torch.randn(2, 1, 40960)
|
|
spec = torch.randn(2,80,160)
|
|
cond = torch.randn(2, 1, 40960)
|
|
ts = torch.LongTensor([555, 556])
|
|
model = DiffusionVocoderWithRef(model_channels=128, channel_mult=[1,1,1.5,2, 3, 4, 6, 8, 8, 8, 8 ],
|
|
num_res_blocks=[1,2, 2, 2, 2, 2, 2, 2, 2, 1, 1 ], spectrogram_conditioning_resolutions=[2,512],
|
|
dropout=.05, attention_resolutions=[512,1024], num_heads=4, kernel_size=3, scale_factor=2,
|
|
conditioning_inputs_provided=True, conditioning_input_dim=80, time_embed_dim_multiplier=4,
|
|
discrete_codes=80, freeze_layers_below=1)
|
|
loading_errors = model.load_state_dict(torch.load('diffuse_new_lyr.pth'), strict=False)
|
|
new_params = loading_errors.missing_keys
|
|
new_params_trained = []
|
|
existing_params_trained = []
|
|
for n,p in model.named_parameters():
|
|
if not hasattr(p, 'DO_NOT_TRAIN'):
|
|
if n in new_params:
|
|
new_params_trained.append(n)
|
|
else:
|
|
existing_params_trained.append(n)
|
|
for n in new_params:
|
|
if n not in new_params_trained:
|
|
print(f"{n} is a new parameter, but it is not marked as trainable.")
|
|
|
|
print(model(clip, ts, spec, cond).shape)
|