fix unified_voice
This commit is contained in:
James Betker
parent
136744dc1d
commit
91f28580e2
|
|
@ -62,7 +62,7 @@ class MelEncoder(nn.Module):
|
|||
|
||||
class UnifiedVoice(nn.Module):
|
||||
def __init__(self, layers=8, model_dim=512, heads=8, max_text_tokens=120, max_mel_tokens=250, max_conditioning_inputs=1,
|
||||
max_conditioning_length=60, shuffle_conditioning=True, mel_length_compression=1024, number_text_tokens=256,
|
||||
mel_length_compression=1024, number_text_tokens=256,
|
||||
start_text_token=255, stop_text_token=0, number_mel_codes=8194, start_mel_token=8192,
|
||||
stop_mel_token=8193, train_solo_embeddings=False, use_mel_codes_as_input=True,
|
||||
checkpointing=True):
|
||||
|
|
@ -74,8 +74,6 @@ class UnifiedVoice(nn.Module):
|
|||
max_text_tokens: Maximum number of text tokens that will be encountered by model.
|
||||
max_mel_tokens: Maximum number of MEL tokens that will be encountered by model.
|
||||
max_conditioning_inputs: Maximum number of conditioning inputs provided to the model. If (1), conditioning input can be of format (b,80,s), otherwise (b,n,80,s).
|
||||
max_conditioning_length: Maximum length of conditioning input. Only needed if shuffle_conditioning=True
|
||||
shuffle_conditioning: Whether or not the conditioning inputs will be shuffled across the sequence dimension. Useful if you want to provide the same input as conditioning and mel_codes.
|
||||
mel_length_compression: The factor between <number_input_samples> and <mel_tokens>. Used to compute MEL code padding given wav input length.
|
||||
number_text_tokens:
|
||||
start_text_token:
|
||||
|
|
@ -95,7 +93,6 @@ class UnifiedVoice(nn.Module):
|
|||
self.number_mel_codes = number_mel_codes
|
||||
self.start_mel_token = start_mel_token
|
||||
self.stop_mel_token = stop_mel_token
|
||||
self.shuffle_conditioning = shuffle_conditioning
|
||||
self.layers = layers
|
||||
self.heads = heads
|
||||
self.max_mel_tokens = max_mel_tokens
|
||||
|
|
@ -110,7 +107,7 @@ class UnifiedVoice(nn.Module):
|
|||
else:
|
||||
self.mel_embedding = MelEncoder(model_dim, resblocks_per_reduction=1)
|
||||
self.gpt, self.mel_pos_embedding, self.text_pos_embedding, self.mel_layer_pos_embedding, self.text_layer_pos_embedding = \
|
||||
build_hf_gpt_transformer(layers, model_dim, heads, self.max_text_tokens+2, self.max_mel_tokens+3, checkpointing)
|
||||
build_hf_gpt_transformer(layers, model_dim, heads, self.max_mel_tokens+2+self.max_conditioning_inputs, self.max_text_tokens+2, checkpointing)
|
||||
if train_solo_embeddings:
|
||||
self.mel_solo_embedding = nn.Parameter(torch.randn(1, 1, model_dim) * .02, requires_grad=True)
|
||||
self.text_solo_embedding = nn.Parameter(torch.randn(1, 1, model_dim) * .02, requires_grad=True)
|
||||
|
|
@ -121,7 +118,6 @@ class UnifiedVoice(nn.Module):
|
|||
self.final_norm = nn.LayerNorm(model_dim)
|
||||
self.text_head = nn.Linear(model_dim, self.number_text_tokens)
|
||||
self.mel_head = nn.Linear(model_dim, self.number_mel_codes)
|
||||
self.max_conditioning_length = max_conditioning_length
|
||||
|
||||
# Initialize the embeddings per the GPT-2 scheme
|
||||
embeddings = [self.text_embedding]
|
||||
|
|
@ -149,23 +145,11 @@ class UnifiedVoice(nn.Module):
|
|||
mel_input_tokens[b, actual_end:] = self.stop_mel_token
|
||||
return mel_input_tokens
|
||||
|
||||
def randomly_permute_conditioning_input(self, speech_conditioning_input):
|
||||
"""
|
||||
Randomly permute the conditioning spectrogram, to destroy any structure present. Note that since the
|
||||
conditioning input is derived from a discrete spectrogram, it does actually retain structure, but only a little
|
||||
bit (actually: exactly how much we want; enough to discriminate different vocal qualities, but nothing about
|
||||
what is being said).
|
||||
"""
|
||||
cond_input = speech_conditioning_input[:,:,torch.randperm(speech_conditioning_input.shape[-1])]
|
||||
if cond_input.shape[-1] > self.max_conditioning_length:
|
||||
cond_input = cond_input[:,:,:self.max_conditioning_length]
|
||||
return cond_input
|
||||
|
||||
def get_logits(self, speech_conditioning_input, first_inputs, first_head, second_inputs=None, second_head=None, get_attns=False):
|
||||
def get_logits(self, speech_conditioning_inputs, first_inputs, first_head, second_inputs=None, second_head=None, get_attns=False):
|
||||
if second_inputs is not None:
|
||||
emb = torch.cat([speech_conditioning_input, first_inputs, second_inputs], dim=1)
|
||||
emb = torch.cat([speech_conditioning_inputs, first_inputs, second_inputs], dim=1)
|
||||
else:
|
||||
emb = torch.cat([speech_conditioning_input, first_inputs], dim=1)
|
||||
emb = torch.cat([speech_conditioning_inputs, first_inputs], dim=1)
|
||||
|
||||
gpt_out = self.gpt(inputs_embeds=emb, return_dict=True, output_attentions=get_attns)
|
||||
if get_attns:
|
||||
|
|
@ -209,9 +193,11 @@ class UnifiedVoice(nn.Module):
|
|||
raw_mels = raw_mels[:, :, :max_mel_len*4]
|
||||
mel_codes = self.set_mel_padding(mel_codes, wav_lengths)
|
||||
|
||||
if self.shuffle_conditioning:
|
||||
speech_conditioning_input = self.randomly_permute_conditioning_input(speech_conditioning_input)
|
||||
speech_conditioning_input = self.conditioning_encoder(speech_conditioning_input).unsqueeze(1)
|
||||
speech_conditioning_input = speech_conditioning_input.unsqueeze(1) if len(speech_conditioning_input.shape) == 3 else speech_conditioning_input
|
||||
conds = []
|
||||
for j in range(speech_conditioning_input.shape[1]):
|
||||
conds.append(self.conditioning_encoder(speech_conditioning_input[:, j]))
|
||||
conds = torch.stack(conds, dim=1)
|
||||
|
||||
text_inputs, text_targets = self.build_aligned_inputs_and_targets(text_inputs, self.start_text_token, self.stop_text_token)
|
||||
text_emb = self.text_embedding(text_inputs) + self.text_pos_embedding(text_inputs)
|
||||
|
|
@ -223,9 +209,9 @@ class UnifiedVoice(nn.Module):
|
|||
mel_emb = self.mel_embedding(mel_inp)
|
||||
mel_emb = mel_emb + self.mel_pos_embedding(mel_codes)
|
||||
if text_first:
|
||||
text_logits, mel_logits = self.get_logits(speech_conditioning_input, text_emb, self.text_head, mel_emb, self.mel_head, get_attns=return_attentions)
|
||||
text_logits, mel_logits = self.get_logits(conds, text_emb, self.text_head, mel_emb, self.mel_head, get_attns=return_attentions)
|
||||
else:
|
||||
mel_logits, text_logits = self.get_logits(speech_conditioning_input, mel_emb, self.mel_head, text_emb, self.text_head, get_attns=return_attentions)
|
||||
mel_logits, text_logits = self.get_logits(conds, mel_emb, self.mel_head, text_emb, self.text_head, get_attns=return_attentions)
|
||||
|
||||
if return_attentions:
|
||||
return mel_logits
|
||||
|
|
@ -245,13 +231,15 @@ class UnifiedVoice(nn.Module):
|
|||
max_text_len = text_lengths.max()
|
||||
text_inputs = F.pad(text_inputs[:, :max_text_len], (0,1), value=self.stop_text_token)
|
||||
|
||||
if self.shuffle_conditioning:
|
||||
speech_conditioning_input = self.randomly_permute_conditioning_input(speech_conditioning_input)
|
||||
speech_conditioning_input = self.conditioning_encoder(speech_conditioning_input).unsqueeze(1)
|
||||
speech_conditioning_input = speech_conditioning_input.unsqueeze(1) if len(speech_conditioning_input.shape) == 3 else speech_conditioning_input
|
||||
conds = []
|
||||
for j in range(speech_conditioning_input.shape[1]):
|
||||
conds.append(self.conditioning_encoder(speech_conditioning_input[:, j]))
|
||||
conds = torch.stack(conds, dim=1)
|
||||
|
||||
text_inputs, text_targets = self.build_aligned_inputs_and_targets(text_inputs, self.start_text_token, self.stop_text_token)
|
||||
text_emb = self.text_embedding(text_inputs) + self.text_pos_embedding(text_inputs) + self.text_solo_embedding
|
||||
text_logits = self.get_logits(speech_conditioning_input, text_emb, self.text_head)
|
||||
text_logits = self.get_logits(conds, text_emb, self.text_head)
|
||||
loss_text = F.cross_entropy(text_logits, text_targets.long())
|
||||
return loss_text.mean()
|
||||
|
||||
|
|
@ -269,9 +257,11 @@ class UnifiedVoice(nn.Module):
|
|||
if raw_mels is not None:
|
||||
raw_mels = raw_mels[:, :, :max_mel_len*4]
|
||||
|
||||
if self.shuffle_conditioning:
|
||||
speech_conditioning_input = self.randomly_permute_conditioning_input(speech_conditioning_input)
|
||||
speech_conditioning_input = self.conditioning_encoder(speech_conditioning_input).unsqueeze(1)
|
||||
speech_conditioning_input = speech_conditioning_input.unsqueeze(1) if len(speech_conditioning_input.shape) == 3 else speech_conditioning_input
|
||||
conds = []
|
||||
for j in range(speech_conditioning_input.shape[1]):
|
||||
conds.append(self.conditioning_encoder(speech_conditioning_input[:, j]))
|
||||
conds = torch.stack(conds, dim=1)
|
||||
|
||||
mel_codes, mel_targets = self.build_aligned_inputs_and_targets(mel_codes, self.start_mel_token, self.stop_mel_token)
|
||||
if raw_mels is not None:
|
||||
|
|
@ -280,7 +270,7 @@ class UnifiedVoice(nn.Module):
|
|||
mel_inp = mel_codes
|
||||
mel_emb = self.mel_embedding(mel_inp)
|
||||
mel_emb = mel_emb + self.mel_pos_embedding(mel_codes) + self.mel_solo_embedding
|
||||
mel_logits = self.get_logits(speech_conditioning_input, mel_emb, self.mel_head)
|
||||
mel_logits = self.get_logits(conds, mel_emb, self.mel_head)
|
||||
loss_mel = F.cross_entropy(mel_logits, mel_targets.long())
|
||||
return loss_mel.mean()
|
||||
|
||||
|
|
@ -302,12 +292,13 @@ class UnifiedVoice(nn.Module):
|
|||
text_inputs, text_targets = self.build_aligned_inputs_and_targets(text_inputs, self.start_text_token, self.stop_text_token)
|
||||
text_emb = self.text_embedding(text_inputs) + self.text_pos_embedding(text_inputs)
|
||||
|
||||
if self.shuffle_conditioning:
|
||||
# Randomly permute the conditioning spectrogram, to destroy any structure present.
|
||||
speech_conditioning_input = self.randomly_permute_conditioning_input(speech_conditioning_input)
|
||||
cond = self.conditioning_encoder(speech_conditioning_input).unsqueeze(1)
|
||||
speech_conditioning_input = speech_conditioning_input.unsqueeze(1) if len(speech_conditioning_input.shape) == 3 else speech_conditioning_input
|
||||
conds = []
|
||||
for j in range(speech_conditioning_input.shape[1]):
|
||||
conds.append(self.conditioning_encoder(speech_conditioning_input[:, j]))
|
||||
conds = torch.stack(conds, dim=1)
|
||||
|
||||
emb = torch.cat([cond, text_emb], dim=1)
|
||||
emb = torch.cat([conds, text_emb], dim=1)
|
||||
self.inference_model.store_mel_emb(emb)
|
||||
|
||||
fake_inputs = torch.full((emb.shape[0], emb.shape[1]+1,), fill_value=1, dtype=torch.long, device=text_inputs.device)
|
||||
|
|
@ -324,10 +315,10 @@ def register_unified_voice2(opt_net, opt):
|
|||
|
||||
|
||||
if __name__ == '__main__':
|
||||
gpt = UnifiedVoice(model_dim=256, heads=4, train_solo_embeddings=True, use_mel_codes_as_input=True)
|
||||
l = gpt(torch.randn(2, 80, 800),
|
||||
torch.randint(high=len(symbols), size=(2,80)),
|
||||
torch.tensor([32, 80]),
|
||||
gpt = UnifiedVoice(model_dim=256, heads=4, train_solo_embeddings=True, use_mel_codes_as_input=True, max_conditioning_inputs=4)
|
||||
l = gpt(torch.randn(2, 3, 80, 800),
|
||||
torch.randint(high=len(symbols), size=(2,120)),
|
||||
torch.tensor([32, 120]),
|
||||
torch.randint(high=8192, size=(2,250)),
|
||||
torch.tensor([150*256,195*256]))
|
||||
torch.tensor([250*256,195*256]))
|
||||
gpt.text_forward(torch.randn(2,80,800), torch.randint(high=50, size=(2,80)), torch.tensor([32, 80]))
|
||||
|
|
|
|||
|
|
@ -124,7 +124,7 @@ if __name__ == '__main__':
|
|||
text = F.pad(text, (0,1)) # This may not be necessary.
|
||||
|
||||
cond_path = args.cond_path if args.cond_preset is None else preselected_cond_voices[args.cond_preset]
|
||||
conds, cond_wav = load_conditioning(cond_path)
|
||||
conds, cond_wav = load_conditioning(cond_path, cond_length=88000)
|
||||
|
||||
with torch.no_grad():
|
||||
print("Performing GPT inference..")
|
||||
|
|
|
|||
Loading…
Reference in New Issue
Block a user