ecker/vall-e
5
9
Fork 0
Code Issues 8 Pull Requests Packages Projects Releases Wiki Activity

work on an interleaved AR (spoiler: it does not work)

Browse Source
This commit is contained in:
mrq 2023-09-03 21:27:58 -05:00
parent 8a6c203277
commit c56ce033d9
2 changed files with 601 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
vall_e/ext/interleaver.py Normal file
View File

@ -0,0 +1,2 @@
# From: https://github.com/facebookresearch/audiocraft/blob/main/audiocraft/modules/codebooks_patterns.py
# audiocraft has heavy dependencies, so it doesn't make sense to depend on it just for this file.

599
vall_e/models/interleaved_ar.py Normal file
View File

@ -0,0 +1,599 @@
import math
import torch
import torch.nn.functional as F
import traceback
from typing import Literal, overload
from functools import partial
from einops import rearrange
from torch import Tensor, einsum, nn
from torch.distributions import Categorical
from torch.nn.utils.rnn import pad_sequence
from torch.utils.checkpoint import checkpoint
from torchmetrics.classification import BinaryAccuracy, MulticlassAccuracy, MulticlassPrecision
from .retnet import RetNetDecoder, RetNetConfig
from .transformer import SinusoidalEmbedding, Block as TransformerBlock
from ..ext.interleaver import (
CodebooksPatternProvider,
DelayedPatternProvider,
MusicLMPattern,
ParallelPatternProvider,
UnrolledPatternProvider,
VALLEPattern,
)
from ..config import cfg
def _get_pattern_provider( name ):
return {
'parallel': ParallelPatternProvider,
'delay': DelayedPatternProvider,
'unroll': UnrolledPatternProvider,
'valle': VALLEPattern,
'musiclm': MusicLMPattern,
}[name]
def _create_mask(l, device):
"""1 is valid region and 0 is invalid."""
seq = torch.arange(max(l), device=device).unsqueeze(0) # (1 t)
stop = torch.tensor(l, device=device).unsqueeze(1) # (b 1)
return (seq < stop).float() # (b t)
def _join(x: tuple[Tensor], sep: Tensor):
"""
Args:
x: (k t d)
sep: (d)
"""
ret = x[0]
for i in range(1, len(x)):
ret = torch.cat((ret, sep[None], x[i]), dim=0)
return ret
def list_to_tensor(x_list: list[Tensor], pattern="t b c -> b t c"):
"""
Args:
x_list: [(t d)]
Returns:
x: (? ? ?)
m: (? ? ?), same as x
"""
l = list(map(len, x_list))
x = rearrange(pad_sequence(x_list), pattern)
m = _create_mask(l, x_list[0].device)
m = m.t().unsqueeze(-1) # (t b 1)
m = rearrange(m, pattern)
m = m.to(x)
return x, m
class Embedding(nn.Embedding):
def forward(self, x_list: list[Tensor]) -> list[Tensor]:
if len(x_list) == 0:
return []
return super().forward(torch.cat(x_list)).split([*map(len, x_list)])
class MultiEmbedding(nn.Embedding):
"""
This embedding sums embeddings on different levels.
"""
def __init__(self, max_n_levels, n_tokens, token_dim):
super().__init__(max_n_levels, token_dim)
self.max_n_levels = max_n_levels
self.n_tokens = n_tokens
self.weight = nn.Parameter(torch.randn(max_n_levels, n_tokens, token_dim))
def forward(self, x_list: list[Tensor]) -> list[Tensor]:
if len(x_list) == 0:
return []
w = self.weight
padded_x_list = []
for xi in x_list:
xi = F.one_hot(xi.to(torch.int64), num_classes=self.n_tokens) # t l' k
xi = F.pad(xi, (0, 0, 0, w.shape[0] - xi.shape[1])) # t l k
padded_x_list.append(xi.to(w))
x = torch.cat(padded_x_list) # n l k
x = einsum("l k d, n l k -> n d", w, x)
x_list = x.split([*map(len, x_list)])
return x_list
class Base(nn.Module):
@property
def causal(self):
return True
@property
def use_stop_token(self):
return True
@property
def norm_type(self):
return "ln"
@property
def arch_type(self) -> str:
return "retnet"
@property
def n_prom_levels(self) -> int:
return 4
@property
def n_resp_levels(self) -> int:
return 4
@property
def n_max_levels(self) -> int:
return 4
@property
def n_tasks(self) -> int:
return 16
@property
def resp_loss_only(self) -> bool:
return False
@property
def recurrent_chunk_size(self) -> int:
return 0
@property
def interleave_pattern(self) -> str | None:
return "musiclm"
@property
def stop_token(self):
return self.n_tokens + 0
@property
def interleaved_token(self):
return self.n_tokens + 1
@property
def ignore_index(self):
return -100 # self.interleaved_token
def _prune(self, l: Tensor):
indices = (l == self.stop_token).nonzero()
if len(indices) == 0:
return l
return l[: indices.min().item()]
@staticmethod
def _unsqueeze_list(x_list, axis=-1):
return [x.unsqueeze(dim=axis) for x in x_list]
@staticmethod
def _samplewise_merge_tensors(*l, sep: Tensor | None):
if sep is None:
cat = torch.cat
else:
cat = partial(_join, sep=sep)
return [*map(cat, zip(*l))]
def _interleave( self, codes ):
if not self.interleave_pattern:
return codes
return codes.flatten()
"""
pattern_provider = _get_pattern_provider( self.interleave_pattern )( self.n_resp_levels )
pattern = pattern_provider.get_pattern( codes.shape[0] )
res, _, _ = pattern.build_pattern_sequence( codes.t()[None, :, :], self.interleaved_token, keep_only_valid_steps=True )
return res[0].t().flatten()
"""
def _deinterleave( self, codes ):
if not self.interleave_pattern:
return codes
return torch.unflatten( codes[:codes.shape[0] // self.n_resp_levels * self.n_resp_levels], 0, ( codes.shape[0] // self.n_resp_levels, self.n_resp_levels ) )
"""
if codes.dim() == 1:
codes = torch.unflatten( codes[:codes.shape[0] // self.n_resp_levels * self.n_resp_levels], 0, ( codes.shape[0] // self.n_resp_levels, self.n_resp_levels ) )
pattern_provider = _get_pattern_provider( self.interleave_pattern )( self.n_resp_levels )
pattern = pattern_provider.get_pattern( codes.shape[0] )
res, _, _ = pattern.revert_pattern_sequence( codes, special_token=self.interleaved_token)
return res[0].t()
"""
def __init__(
self,
n_tokens: int = 1024,
d_model: int = 512,
n_heads: int = 8,
n_layers: int = 12,
p_dropout: float = 0.1,
config: dict | None = None
):
super().__init__()
self._cfg = config
self.n_tokens = n_tokens
self.d_model = d_model
self.n_heads = n_heads
self.n_layers = n_layers
# + tasks for each token they represent in the prom
n_prom_tokens = n_tokens + (self.n_tasks - 1) + (1 if self.interleave_pattern else 0) # - 1 because tts is an inherent task
# +1 to include the stop token + 1 to include interleave token
n_resp_tokens = n_tokens + (1 if self.use_stop_token else 0) + (1 if self.interleave_pattern else 0) # AR requires a stop token to... know when to stop
self.text_emb = Embedding(n_tokens, d_model)
self.proms_emb = MultiEmbedding(self.n_prom_levels, n_prom_tokens, d_model)
self.resps_emb = MultiEmbedding(1, n_resp_tokens, d_model)
self.sep = nn.Parameter(torch.randn(d_model))
if self.arch_type == "transformer":
self.sin_emb = SinusoidalEmbedding(d_model)
self.blocks = nn.ModuleList([TransformerBlock(
d_model=d_model,
n_heads=n_heads,
p_dropout=p_dropout,
causal=self.causal,
norm_type=self.norm_type,
n_levels=1,
) for _ in range(n_layers) ])
elif self.arch_type == "retnet":
self.retnet = RetNetDecoder(RetNetConfig(
vocab_size=n_tokens,
decoder_embed_dim=d_model,
decoder_retention_heads=n_heads,
decoder_ffn_embed_dim=d_model * 4,
decoder_layers=n_layers,
dropout=p_dropout,
checkpoint_activations=True,
chunkwise_recurrent=self.causal and self.recurrent_chunk_size > 0,
recurrent_chunkwise_size=self.recurrent_chunk_size if self.causal else 0,
no_output_layer=True,
decoder_normalize_before=True,
))
# I imagine because each step returns `resp_level`s tokens at once, so we need to have a classifier for each level
#self.classifier = nn.ModuleList([ nn.Linear(d_model, n_resp_tokens) for _ in range(self.n_resp_levels) ]) if self.interleave_pattern else nn.Linear(d_model, n_resp_tokens)
self.classifier = nn.Linear(d_model, n_resp_tokens)
self.accuracy_metric = MulticlassAccuracy(
n_resp_tokens,
top_k=10,
average="micro",
multidim_average="global",
ignore_index=self.ignore_index,
)
self.precision_metric = MulticlassPrecision(
n_resp_tokens,
top_k=10,
average="micro",
multidim_average="global",
ignore_index=self.ignore_index,
)
@overload
def forward(
self,
text_list: list[Tensor],
proms_list: list[Tensor],
resps_list: list[Tensor],
targ_list: list[Tensor] | None = None,
quant_levels: Tensor | None = None,
shift_targ_list: bool = False,
return_all: Literal[False] = False,
return_all_resp: Literal[False] = False,
sampling_temperature: float = 1.0,
) -> Tensor:
...
@overload
def forward(
self,
text_list: list[Tensor],
proms_list: list[Tensor],
resps_list: list[Tensor],
targ_list: list[Tensor] | None = None,
quant_levels: Tensor | None = None,
shift_targ_list: bool = False,
return_all: Literal[True] = True,
return_all_resp: Literal[True] = True,
sampling_temperature: float = 1.0,
) -> list[Tensor]:
...
def _forward(
self,
text_list: list[Tensor],
proms_list: list[Tensor],
resps_list: list[Tensor],
targ_list: list[Tensor] | None = None,
quant_levels: Tensor | None = None,
shift_targ_list: bool = False,
return_all: bool = False,
return_all_resp: bool = False,
sampling_temperature: float = 1.0,
state: dict | None = None,
):
"""
Args:
text_list: [t] * b
proms_list: [t' l] * b, l quantization levels.
resps_list: [t'' l] * b, l quantization levels.
targ_list: [t''] * b, one quantization level only; when given, loss will be computed
quant_levels: specify which quant_levels to feed forward, used in NAR mode.
shift_targ_list: whether to shift target list when computing loss. True if AR.
return_all_resp: True if NAR.
sampling_temperature: a lower temperature makes the result more robust but less diverse.
Returns:
y: sampled tokens
"""
batch_size = len(text_list)
x_list = self._samplewise_merge_tensors(
self.text_emb(text_list),
self.proms_emb(proms_list),
self.resps_emb(resps_list),
sep=self.sep,
)
x, m = list_to_tensor(x_list)
device = x.device
if state is not None:
# prefill
prefill_size = x.shape[1]
# run the initial prompt to fill the KV cache
if len(state) == 0:
for n in range(prefill_size):
xi = x[:, n, :].unsqueeze(1)
self.retnet(xi, incremental_state=state, token_embeddings=xi, features_only=True)
# grab last token(s)
x = x[:, -1, :].unsqueeze(1)
if self.arch_type == "transformer":
x = self.sin_emb.add_pe(x)
for block in self.blocks:
x = block(x, m, quant_levels)
elif self.arch_type == "retnet":
# to-do: actually make this work and verify it works with recurrent_forward / chunkwise_forward
x, _ = self.retnet(x, incremental_state=state, token_embeddings=x, features_only=True)
x = self.classifier(x) * m
# Remove padding
h_list = [hi[:li] for hi, li in zip(x, map(len, x_list))]
if True:
logits = [hi[:] for hi, li in zip(h_list, map(len, resps_list))]
ret = [ Categorical(logits=hi / sampling_temperature).sample() for hi in logits ]
print( [ r for r in ret ] )
# compute loss if the target is given
if targ_list is not None:
if any([l == 0 for l in map(len, targ_list)]):
raise ValueError("Cannot compute loss given empty targ_list.")
ignore_sep = torch.tensor(self.ignore_index, device=device)
# ignore the prompt when computing loss
prom_list = [
torch.full_like(t[..., 0], self.ignore_index) for t in proms_list
]
# remake input with ignored input prompt
text_prom_list = self._samplewise_merge_tensors(
text_list, prom_list, sep=ignore_sep
)
for i in range(len(text_prom_list)):
# ignore computing loss against text/prompt portion of input
# the NAR doesn't need to compute the loss for it
if self.resp_loss_only:
text_prom_list[i][:] = self.ignore_index
# roll the text/prompt for loss computing
# the AR benefits from this, for some reason I'll figure out later
else:
text_prom_list[i] = text_prom_list[i].roll(-1, dims=0)
text_prom_list[i][-1] = self.ignore_index
# for the AR, roll by one and mark the ending with a stop token
# this coerces the model into properly inferencing causally
# why we don't just append a stop token in the dataloader, who knows
if shift_targ_list:
targ_list = [*targ_list]
for i in range(len(targ_list)):
targ_list[i] = targ_list[i].roll(-self.n_resp_levels, dims=0)
for j in range(self.n_resp_levels):
targ_list[i][-j-1] = self.stop_token
# create the new target sequence to compute the loss against
y_list = self._samplewise_merge_tensors( text_prom_list, targ_list, sep=ignore_sep )
self.loss = dict(
nll=F.cross_entropy(
torch.cat(h_list), # input / predicted logits
torch.cat(y_list), # target / ground truth
ignore_index=self.ignore_index,
)
)
self.stats = dict(
acc = self.accuracy_metric( torch.cat(h_list), torch.cat(y_list) ),
precision = self.precision_metric( torch.cat(h_list), torch.cat(y_list) ),
)
# return the entire generated token string
if return_all:
logits = [hi[:] for hi, li in zip(h_list, map(len, resps_list))]
# return the entire generated response
elif return_all_resp:
logits = [hi[-li:] for hi, li in zip(h_list, map(len, resps_list))]
# return the last chunkwise piece
elif self.causal and self.recurrent_chunk_size > 0:
logits = [hi[-self.recurrent_chunk_size:] for hi, li in zip(h_list, map(len, resps_list))]
# return just the last code
else:
logits = [ hi[-1:] for hi in h_list ]
return [ Categorical(logits=hi / sampling_temperature).sample() for hi in logits ]
def forward(
self,
text_list: list[Tensor],
proms_list: list[Tensor],
resps_list: list[Tensor] | None = None,
max_steps: int = 1000,
sampling_temperature: float = 1.0,
):
if resps_list is not None:
resps_list = [self._interleave(r) for r in resps_list] # guarantees we only have the first levels
return self._forward(
text_list=text_list,
proms_list=proms_list,
resps_list=self._unsqueeze_list(resps_list),
targ_list=resps_list,
quant_levels=None,
shift_targ_list=True,
return_all_resp=False,
)
device = text_list[0].device
batch_size = len(text_list)
resps_list: list[Tensor] = [ torch.zeros(0, device=device).to(torch.int16) for _ in text_list ]
stopped = torch.zeros(batch_size, device=device).bool()
state = {} if cfg.inference.recurrent_forward else None
for n in range(max_steps // max(1, self.recurrent_chunk_size)):
# get next in sequence
r = self._forward(
text_list,
proms_list,
self._unsqueeze_list(resps_list),
sampling_temperature=sampling_temperature,
state=state
)
# append tokens
for i, ri in enumerate(r):
if self.stop_token in ri:
stopped[i] = True
resps_list[i] = torch.cat([resps_list[i], ri])
# stop token found
stopped |= r == self.stop_token
if stopped.all().item():
break
pruned = [self._prune(r) for r in resps_list]
print( [ r for r in pruned ] )
deinterleaved = [ self._deinterleave(r) for r in pruned ]
print( [ r for r in deinterleaved ] )
return deinterleaved
def example_usage():
from ..config import cfg
cfg.trainer.backend = "local"
cfg.trainer.check_for_oom = False
from functools import partial
from einops import repeat
from ..emb.qnt import decode_to_file
from ..engines import Engine, Engines
from tqdm import tqdm, trange
device = "cuda"
x8 = partial(repeat, pattern="t -> t l", l=cfg.models.prom_levels)
symmap = {'<s>': 1, '</s>': 2, ' ': 3, '.': 4, ',': 5, '!': 6, '?': 7, 'p': 7, 'iː': 8, 'ɚ': 9, 'ˌ': 10, 'dˌ': 11, 'mˌ': 12, 'd': 13, 'ɹ': 14, 'tˈ': 15, 'pˌ': 16, 'uː': 17, 'l': 18, 'æ': 19, 'ɛ': 20, 'ɪ': 21, 'j': 22, 'ʊ': 23, 't': 24, 'n': 25, 'v': 26, 'a': 27, 'o': 28, 'ŋ': 29, 'w': 30, 'ʌ': 31, 'hˈ': 32, 'ɡˈ': 33, 'ə': 34, 'θˈ': 35, 'dˈ': 36, 'wˌ': 37, 'h': 38, 'z': 39, 'k': 40, 'ð': 41, 'ɡˌ': 42, 'ˈ': 43, 'fˈ': 44, 'i': 45, 's': 46, 'ʃ': 47, 'wˈ': 48, 'ðˈ': 49, 'ɹˈ': 50, 'lˈ': 51, 'ɡ': 52, 'oː': 53, 'mˈ': 54, 'e': 55, 'ɑː': 56, 'nˈ': 57, 'm': 58, 'θˌ': 59, 'sˈ': 60, 'f': 61, 'ɔː': 62, 'hˌ': 63, 'b': 64, 'jˈ': 65, 'ɐ': 66, 'ʒˈ': 67, 'θ': 68, 'bˈ': 69, 'ɾ': 70, 'ɜː': 71, 'ʌˈ': 72, 'ʃˌ': 73, 'bˌ': 74, 'kˈ': 75, 'ɔ': 76, 'zˈ': 77, 'ᵻ': 78, 'kˌ': 79, 'vˈ': 80, 'fˌ': 81, 'ʒ': 82, 'ʃˈ': 83, 'ɹˌ': 84, 'tˌ': 85, 'pˈ': 86, 'ðˌ': 87, 'sˌ': 88, 'nˌ': 89, 'lˌ': 90, '̩': 91, 'ʔ': 92, 'vˌ': 93, 'ɪˈ': 94, '"': 95, 'ɪˌ': 96, 'ʒˌ': 97, 'uːˌ': 98, 'ʊˈ': 99, 'jˌ': 100, 'uːˈ': 101, 'iːˈ': 102, 'zˌ': 103, '.ˈ': 104, '…': 105, 'ŋˌ': 106, 'ɐˌ': 107, '—ˈ': 108, 'iˌ': 109, 'iːˌ': 110, 'ɛː': 111, ')': 112, ')ˈ': 113, '(': 114, 'u': 115, '-': 116, 'ɖˈ': 117, 'iˈ': 118, 'ʰˈ': 119, 'ɟˈ': 120, '̃': 121, 'eː': 122, 'ɾˈ': 123, 'r': 124, 'ʰ': 125, '-ˌ': 126, 'ɫ': 127, 'q': 128, '—': 129, 'ʊˌ': 130, 'aː': 131, 'cˈ': 132, '…ˈ': 133, 'c': 134, 'ɳ': 135, 'ɐˈ': 136, 'x': 137, 'ʔˌ': 138, '.ˌ': 139, 'ɑ': 140, '?ˈ': 141, '̩ˈ': 142, '"ˈ': 143, ',ˈ': 144, 'ŋˈ': 145, 'əˌ': 146, '!ˈ': 147, '"ˌ': 148, '?ˌ': 149, ',ˌ': 150, '—ˌ': 151, '̩ˌ': 152, 'əˈ': 153, '!ˌ': 154, 'ɬ': 155, 'ʲ': 156, '¡': 157, 'ɯ': 158, 'qˌ': 159, 'ʑ': 160, 'ʑˈ': 161, '¿': 162, 'ɑːˈ': 163, 'iːː': 164, 'ɛˈ': 165, '¡ˈ': 166, 'æˈ': 167, 'ç': 168, 'ɾˌ': 169, 'ᵻˈ': 170, 'xˈ': 171, 'ɔːˈ': 172, ';': 173, 'ɬˌ': 174, ':': 175, 'ʔˈ': 176, 'ɑːˌ': 177, 'ɬˈ': 178}
def tokenize(content, lang_marker="en"):
split = content.split(" ")
phones = [f"<s>"] + [ " " if not p else p for p in split ] + [f"</s>"]
return torch.tensor([*map(symmap.get, phones)]).to()
kwargs = {
'n_tokens': 1024,
'd_model': 1024,
'n_heads': 16,
'n_layers': 12,
}
models = { "ar": Base(**kwargs).to(device) }
for name, model in models.items():
print(f"{name} parameter count: {sum(p.numel() for p in model.parameters() if p.requires_grad)}")
engines = Engines({ name: Engine(model=model, optimizer=torch.optim.AdamW(model.parameters(), lr=1e-4)) for name, model in models.items() })
train = True
qnt = torch.load("data/qnt.pt")[0].t()[:, :cfg.models.prom_levels].to(device)
text_list = [
tokenize("ˈ a ɪ w ɪ l nˌ ɑː t ˈ æ s k ɐ sˈ ɛ k ə n d tˈ a ɪ m").to(device),
#tokenize("ˌ ɔ n ɡˌ o ʊ ɪ ŋ hˈ o ʊ m ð ə tˈ uː f ɹˈ ɛ n d z fˈ a ʊ n d ɐ lˈ ɛ ɾ ɚ f ɹ ʌ m ˈ æ θ o ʊ z , hˌ uː d ɪ zˈ a ɪ ɚ d ðˌ ɛ m t ə mˈ iː t hˌ ɪ m æ t ð ə ɡ ɹˈ æ n d t ʃˈ ɑː ɹ l ɪ mˌ æ ɡ n i ɔ n ð ə fˈ ɑː l o ʊ ɪ ŋ dˈ e ɪ .").to(device),
]
proms_list = [
qnt.to(device),
]
resps_list = [
qnt.to(device),
]
def sample( filename, steps=400 ):
AR = None
engines.eval()
for name, engine in engines.items():
if name[:2] == "ar":
AR = engine
resps_list = AR(text_list, proms_list, max_steps=steps, sampling_temperature=1.0)
decode_to_file(resps_list[0].cpu(), f"./data/{filename}.wav", device="cpu")
if train:
sample("init", 15)
engines.train()
t = trange(100)
for i in t:
stats = {"step": i}
"""
for name, engine in engines.items():
stats |= engine.traverse(text_list=text_list, proms_list=proms_list, resps_list=resps_list)
"""
stats = engines.step({"text_list": text_list, "proms_list": proms_list, "resps_list": resps_list})
tqdm.write(f"{stats}")
else:
for name, engine in engines.items():
engine.module.load_state_dict(torch.load(f"./data/{name}.pth"))
sample("final")
if __name__ == "__main__":
example_usage()