cleanup Vq
Browse files- audiocraft/builders.py +4 -10
- audiocraft/encodec.py +0 -3
- audiocraft/lm.py +4 -2
- audiocraft/quantization/__init__.py +0 -9
- audiocraft/quantization/base.py +0 -99
- audiocraft/quantization/core_vq.py +0 -405
- audiocraft/{quantization/vq.py → vq.py} +149 -14
- demo.py +4 -3
audiocraft/builders.py
CHANGED
@@ -4,15 +4,9 @@
|
|
4 |
# This source code is licensed under the license found in the
|
5 |
# LICENSE file in the root directory of this source tree.
|
6 |
|
7 |
-
"""
|
8 |
-
All the functions to build the relevant models and modules
|
9 |
-
from the Hydra config.
|
10 |
-
"""
|
11 |
-
|
12 |
import typing as tp
|
13 |
import omegaconf
|
14 |
import torch
|
15 |
-
|
16 |
from .encodec import CompressionModel, EncodecModel
|
17 |
from .lm import LMModel
|
18 |
from .seanet import SEANetDecoder
|
@@ -24,15 +18,15 @@ from .conditioners import (
|
|
24 |
T5Conditioner,
|
25 |
)
|
26 |
from .unet import DiffusionUnet
|
27 |
-
|
28 |
from .utils.utils import dict_from_config
|
29 |
from .diffusion_schedule import MultiBandProcessor, SampleProcessor
|
30 |
|
31 |
|
32 |
-
def get_quantizer(quantizer
|
33 |
klass = {
|
34 |
-
'no_quant':
|
35 |
-
'rvq':
|
36 |
}[quantizer]
|
37 |
kwargs = dict_from_config(getattr(cfg, quantizer))
|
38 |
if quantizer != 'no_quant':
|
|
|
4 |
# This source code is licensed under the license found in the
|
5 |
# LICENSE file in the root directory of this source tree.
|
6 |
|
|
|
|
|
|
|
|
|
|
|
7 |
import typing as tp
|
8 |
import omegaconf
|
9 |
import torch
|
|
|
10 |
from .encodec import CompressionModel, EncodecModel
|
11 |
from .lm import LMModel
|
12 |
from .seanet import SEANetDecoder
|
|
|
18 |
T5Conditioner,
|
19 |
)
|
20 |
from .unet import DiffusionUnet
|
21 |
+
from .vq import ResidualVectorQuantizer
|
22 |
from .utils.utils import dict_from_config
|
23 |
from .diffusion_schedule import MultiBandProcessor, SampleProcessor
|
24 |
|
25 |
|
26 |
+
def get_quantizer(quantizer, cfg, dimension):
|
27 |
klass = {
|
28 |
+
'no_quant': None,
|
29 |
+
'rvq': ResidualVectorQuantizer
|
30 |
}[quantizer]
|
31 |
kwargs = dict_from_config(getattr(cfg, quantizer))
|
32 |
if quantizer != 'no_quant':
|
audiocraft/encodec.py
CHANGED
@@ -9,7 +9,6 @@ Also defines the main interface that a model must follow to be usable as an audi
|
|
9 |
|
10 |
from abc import ABC, abstractmethod
|
11 |
import logging
|
12 |
-
import math
|
13 |
from pathlib import Path
|
14 |
import typing as tp
|
15 |
|
@@ -19,8 +18,6 @@ import torch
|
|
19 |
from torch import nn
|
20 |
from transformers import EncodecModel as HFEncodecModel
|
21 |
|
22 |
-
import audiocraft.quantization as qt
|
23 |
-
|
24 |
|
25 |
logger = logging.getLogger()
|
26 |
|
|
|
9 |
|
10 |
from abc import ABC, abstractmethod
|
11 |
import logging
|
|
|
12 |
from pathlib import Path
|
13 |
import typing as tp
|
14 |
|
|
|
18 |
from torch import nn
|
19 |
from transformers import EncodecModel as HFEncodecModel
|
20 |
|
|
|
|
|
21 |
|
22 |
logger = logging.getLogger()
|
23 |
|
audiocraft/lm.py
CHANGED
@@ -433,7 +433,9 @@ class LMModel(StreamingModule):
|
|
433 |
|
434 |
# print(f'{unconditional_state=} \n
|
435 |
# print('Set All to Special')
|
436 |
-
|
|
|
|
|
437 |
|
438 |
|
439 |
|
@@ -449,7 +451,7 @@ class LMModel(StreamingModule):
|
|
449 |
unconditional_state.clear()
|
450 |
|
451 |
out_codes, _, _ = pattern.revert_pattern_sequence(gen_sequence, special_token=unknown_token)
|
452 |
-
|
453 |
out_start_offset = start_offset if remove_prompts else 0
|
454 |
out_codes = out_codes[..., out_start_offset:max_gen_len]
|
455 |
|
|
|
433 |
|
434 |
# print(f'{unconditional_state=} \n
|
435 |
# print('Set All to Special')
|
436 |
+
|
437 |
+
# RUNS with = 2047 just different of self.special_token_id -> 2047 is drill noise
|
438 |
+
# next_token[:] = self.special_token_id
|
439 |
|
440 |
|
441 |
|
|
|
451 |
unconditional_state.clear()
|
452 |
|
453 |
out_codes, _, _ = pattern.revert_pattern_sequence(gen_sequence, special_token=unknown_token)
|
454 |
+
print(f'{out_codes.shape=} {out_codes.min()} {out_codes.max()}\n')
|
455 |
out_start_offset = start_offset if remove_prompts else 0
|
456 |
out_codes = out_codes[..., out_start_offset:max_gen_len]
|
457 |
|
audiocraft/quantization/__init__.py
DELETED
@@ -1,9 +0,0 @@
|
|
1 |
-
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
2 |
-
# All rights reserved.
|
3 |
-
#
|
4 |
-
# This source code is licensed under the license found in the
|
5 |
-
# LICENSE file in the root directory of this source tree.
|
6 |
-
"""RVQ."""
|
7 |
-
# flake8: noqa
|
8 |
-
from .vq import ResidualVectorQuantizer
|
9 |
-
from .base import BaseQuantizer, DummyQuantizer, QuantizedResult
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
audiocraft/quantization/base.py
DELETED
@@ -1,99 +0,0 @@
|
|
1 |
-
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
2 |
-
# All rights reserved.
|
3 |
-
#
|
4 |
-
# This source code is licensed under the license found in the
|
5 |
-
# LICENSE file in the root directory of this source tree.
|
6 |
-
|
7 |
-
"""
|
8 |
-
Base class for all quantizers.
|
9 |
-
"""
|
10 |
-
|
11 |
-
from dataclasses import dataclass, field
|
12 |
-
import typing as tp
|
13 |
-
|
14 |
-
import torch
|
15 |
-
from torch import nn
|
16 |
-
|
17 |
-
|
18 |
-
@dataclass
|
19 |
-
class QuantizedResult:
|
20 |
-
x: torch.Tensor
|
21 |
-
codes: torch.Tensor
|
22 |
-
bandwidth: torch.Tensor # bandwidth in kb/s used, per batch item.
|
23 |
-
penalty: tp.Optional[torch.Tensor] = None
|
24 |
-
metrics: dict = field(default_factory=dict)
|
25 |
-
|
26 |
-
|
27 |
-
class BaseQuantizer(nn.Module):
|
28 |
-
"""Base class for quantizers.
|
29 |
-
"""
|
30 |
-
|
31 |
-
def forward(self, x: torch.Tensor, frame_rate: int) -> QuantizedResult:
|
32 |
-
"""
|
33 |
-
Given input tensor x, returns first the quantized (or approximately quantized)
|
34 |
-
representation along with quantized codes, bandwidth, and any penalty term for the loss.
|
35 |
-
Finally, this returns a dict of metrics to update logging etc.
|
36 |
-
Frame rate must be passed so that the bandwidth is properly computed.
|
37 |
-
"""
|
38 |
-
raise NotImplementedError()
|
39 |
-
|
40 |
-
def encode(self, x: torch.Tensor) -> torch.Tensor:
|
41 |
-
"""Encode a given input tensor with the specified sample rate at the given bandwidth."""
|
42 |
-
raise NotImplementedError()
|
43 |
-
|
44 |
-
def decode(self, codes: torch.Tensor) -> torch.Tensor:
|
45 |
-
"""Decode the given codes to the quantized representation."""
|
46 |
-
raise NotImplementedError()
|
47 |
-
|
48 |
-
@property
|
49 |
-
def total_codebooks(self):
|
50 |
-
"""Total number of codebooks."""
|
51 |
-
raise NotImplementedError()
|
52 |
-
|
53 |
-
@property
|
54 |
-
def num_codebooks(self):
|
55 |
-
"""Number of active codebooks."""
|
56 |
-
raise NotImplementedError()
|
57 |
-
|
58 |
-
def set_num_codebooks(self, n: int):
|
59 |
-
"""Set the number of active codebooks."""
|
60 |
-
raise NotImplementedError()
|
61 |
-
|
62 |
-
|
63 |
-
class DummyQuantizer(BaseQuantizer):
|
64 |
-
"""Fake quantizer that actually does not perform any quantization.
|
65 |
-
"""
|
66 |
-
def __init__(self):
|
67 |
-
super().__init__()
|
68 |
-
|
69 |
-
def forward(self, x: torch.Tensor, frame_rate: int):
|
70 |
-
q = x.unsqueeze(1)
|
71 |
-
return QuantizedResult(x, q, torch.tensor(q.numel() * 32 * frame_rate / 1000 / len(x)).to(x))
|
72 |
-
|
73 |
-
def encode(self, x: torch.Tensor) -> torch.Tensor:
|
74 |
-
"""Encode a given input tensor with the specified sample rate at the given bandwidth.
|
75 |
-
In the case of the DummyQuantizer, the codes are actually identical
|
76 |
-
to the input and resulting quantized representation as no quantization is done.
|
77 |
-
"""
|
78 |
-
return x.unsqueeze(1)
|
79 |
-
|
80 |
-
def decode(self, codes: torch.Tensor) -> torch.Tensor:
|
81 |
-
"""Decode the given codes to the quantized representation.
|
82 |
-
In the case of the DummyQuantizer, the codes are actually identical
|
83 |
-
to the input and resulting quantized representation as no quantization is done.
|
84 |
-
"""
|
85 |
-
return codes.squeeze(1)
|
86 |
-
|
87 |
-
@property
|
88 |
-
def total_codebooks(self):
|
89 |
-
"""Total number of codebooks."""
|
90 |
-
return 1
|
91 |
-
|
92 |
-
@property
|
93 |
-
def num_codebooks(self):
|
94 |
-
"""Total number of codebooks."""
|
95 |
-
return self.total_codebooks
|
96 |
-
|
97 |
-
def set_num_codebooks(self, n: int):
|
98 |
-
"""Set the number of active codebooks."""
|
99 |
-
raise AttributeError("Cannot override the number of codebooks for the dummy quantizer")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
audiocraft/quantization/core_vq.py
DELETED
@@ -1,405 +0,0 @@
|
|
1 |
-
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
2 |
-
# All rights reserved.
|
3 |
-
#
|
4 |
-
# This source code is licensed under the license found in the
|
5 |
-
# LICENSE file in the root directory of this source tree.
|
6 |
-
|
7 |
-
import typing as tp
|
8 |
-
|
9 |
-
from einops import rearrange, repeat
|
10 |
-
import flashy
|
11 |
-
import torch
|
12 |
-
from torch import nn, einsum
|
13 |
-
import torch.nn.functional as F
|
14 |
-
|
15 |
-
|
16 |
-
def exists(val: tp.Optional[tp.Any]) -> bool:
|
17 |
-
return val is not None
|
18 |
-
|
19 |
-
|
20 |
-
def default(val: tp.Any, d: tp.Any) -> tp.Any:
|
21 |
-
return val if exists(val) else d
|
22 |
-
|
23 |
-
|
24 |
-
def l2norm(t):
|
25 |
-
return F.normalize(t, p=2, dim=-1)
|
26 |
-
|
27 |
-
|
28 |
-
def ema_inplace(moving_avg, new, decay: float):
|
29 |
-
moving_avg.data.mul_(decay).add_(new, alpha=(1 - decay))
|
30 |
-
|
31 |
-
|
32 |
-
def laplace_smoothing(x, n_categories: int, epsilon: float = 1e-5):
|
33 |
-
return (x + epsilon) / (x.sum() + n_categories * epsilon)
|
34 |
-
|
35 |
-
|
36 |
-
def uniform_init(*shape: int):
|
37 |
-
t = torch.empty(shape)
|
38 |
-
nn.init.kaiming_uniform_(t)
|
39 |
-
return t
|
40 |
-
|
41 |
-
|
42 |
-
def sample_vectors(samples, num: int):
|
43 |
-
num_samples, device = samples.shape[0], samples.device
|
44 |
-
|
45 |
-
if num_samples >= num:
|
46 |
-
indices = torch.randperm(num_samples, device=device)[:num]
|
47 |
-
else:
|
48 |
-
indices = torch.randint(0, num_samples, (num,), device=device)
|
49 |
-
|
50 |
-
return samples[indices]
|
51 |
-
|
52 |
-
|
53 |
-
def kmeans(samples, num_clusters: int, num_iters: int = 10):
|
54 |
-
dim, dtype = samples.shape[-1], samples.dtype
|
55 |
-
|
56 |
-
means = sample_vectors(samples, num_clusters)
|
57 |
-
|
58 |
-
for _ in range(num_iters):
|
59 |
-
diffs = rearrange(samples, "n d -> n () d") - rearrange(
|
60 |
-
means, "c d -> () c d"
|
61 |
-
)
|
62 |
-
dists = -(diffs ** 2).sum(dim=-1)
|
63 |
-
|
64 |
-
buckets = dists.max(dim=-1).indices
|
65 |
-
bins = torch.bincount(buckets, minlength=num_clusters)
|
66 |
-
zero_mask = bins == 0
|
67 |
-
bins_min_clamped = bins.masked_fill(zero_mask, 1)
|
68 |
-
|
69 |
-
new_means = buckets.new_zeros(num_clusters, dim, dtype=dtype)
|
70 |
-
new_means.scatter_add_(0, repeat(buckets, "n -> n d", d=dim), samples)
|
71 |
-
new_means = new_means / bins_min_clamped[..., None]
|
72 |
-
|
73 |
-
means = torch.where(zero_mask[..., None], means, new_means)
|
74 |
-
|
75 |
-
return means, bins
|
76 |
-
|
77 |
-
|
78 |
-
def orthogonal_loss_fn(t):
|
79 |
-
# eq (2) from https://arxiv.org/abs/2112.00384
|
80 |
-
n = t.shape[0]
|
81 |
-
normed_codes = l2norm(t)
|
82 |
-
identity = torch.eye(n, device=t.device)
|
83 |
-
cosine_sim = einsum("i d, j d -> i j", normed_codes, normed_codes)
|
84 |
-
return ((cosine_sim - identity) ** 2).sum() / (n ** 2)
|
85 |
-
|
86 |
-
|
87 |
-
class EuclideanCodebook(nn.Module):
|
88 |
-
"""Codebook with Euclidean distance.
|
89 |
-
|
90 |
-
Args:
|
91 |
-
dim (int): Dimension.
|
92 |
-
codebook_size (int): Codebook size.
|
93 |
-
kmeans_init (bool): Whether to use k-means to initialize the codebooks.
|
94 |
-
If set to true, run the k-means algorithm on the first training batch and use
|
95 |
-
the learned centroids as initialization.
|
96 |
-
kmeans_iters (int): Number of iterations used for k-means algorithm at initialization.
|
97 |
-
decay (float): Decay for exponential moving average over the codebooks.
|
98 |
-
epsilon (float): Epsilon value for numerical stability.
|
99 |
-
threshold_ema_dead_code (int): Threshold for dead code expiration. Replace any codes
|
100 |
-
that have an exponential moving average cluster size less than the specified threshold with
|
101 |
-
randomly selected vector from the current batch.
|
102 |
-
"""
|
103 |
-
def __init__(
|
104 |
-
self,
|
105 |
-
dim: int,
|
106 |
-
codebook_size: int,
|
107 |
-
kmeans_init: int = False,
|
108 |
-
kmeans_iters: int = 10,
|
109 |
-
decay: float = 0.8,
|
110 |
-
epsilon: float = 1e-5,
|
111 |
-
threshold_ema_dead_code: int = 2,
|
112 |
-
):
|
113 |
-
super().__init__()
|
114 |
-
self.decay = decay
|
115 |
-
init_fn: tp.Union[tp.Callable[..., torch.Tensor], tp.Any] = uniform_init if not kmeans_init else torch.zeros
|
116 |
-
embed = init_fn(codebook_size, dim)
|
117 |
-
|
118 |
-
self.codebook_size = codebook_size
|
119 |
-
|
120 |
-
self.kmeans_iters = kmeans_iters
|
121 |
-
self.epsilon = epsilon
|
122 |
-
self.threshold_ema_dead_code = threshold_ema_dead_code
|
123 |
-
|
124 |
-
self.register_buffer("inited", torch.Tensor([not kmeans_init]))
|
125 |
-
self.register_buffer("cluster_size", torch.zeros(codebook_size))
|
126 |
-
self.register_buffer("embed", embed)
|
127 |
-
self.register_buffer("embed_avg", embed.clone())
|
128 |
-
|
129 |
-
@torch.jit.ignore
|
130 |
-
def init_embed_(self, data):
|
131 |
-
if self.inited:
|
132 |
-
return
|
133 |
-
|
134 |
-
embed, cluster_size = kmeans(data, self.codebook_size, self.kmeans_iters)
|
135 |
-
self.embed.data.copy_(embed)
|
136 |
-
self.embed_avg.data.copy_(embed.clone())
|
137 |
-
self.cluster_size.data.copy_(cluster_size)
|
138 |
-
self.inited.data.copy_(torch.Tensor([True]))
|
139 |
-
# Make sure all buffers across workers are in sync after initialization
|
140 |
-
flashy.distrib.broadcast_tensors(self.buffers())
|
141 |
-
|
142 |
-
def replace_(self, samples, mask):
|
143 |
-
modified_codebook = torch.where(
|
144 |
-
mask[..., None], sample_vectors(samples, self.codebook_size), self.embed
|
145 |
-
)
|
146 |
-
self.embed.data.copy_(modified_codebook)
|
147 |
-
|
148 |
-
def expire_codes_(self, batch_samples):
|
149 |
-
if self.threshold_ema_dead_code == 0:
|
150 |
-
return
|
151 |
-
|
152 |
-
expired_codes = self.cluster_size < self.threshold_ema_dead_code
|
153 |
-
if not torch.any(expired_codes):
|
154 |
-
return
|
155 |
-
|
156 |
-
batch_samples = rearrange(batch_samples, "... d -> (...) d")
|
157 |
-
self.replace_(batch_samples, mask=expired_codes)
|
158 |
-
flashy.distrib.broadcast_tensors(self.buffers())
|
159 |
-
|
160 |
-
def preprocess(self, x):
|
161 |
-
x = rearrange(x, "... d -> (...) d")
|
162 |
-
return x
|
163 |
-
|
164 |
-
def quantize(self, x):
|
165 |
-
embed = self.embed.t()
|
166 |
-
dist = -(
|
167 |
-
x.pow(2).sum(1, keepdim=True)
|
168 |
-
- 2 * x @ embed
|
169 |
-
+ embed.pow(2).sum(0, keepdim=True)
|
170 |
-
)
|
171 |
-
embed_ind = dist.max(dim=-1).indices
|
172 |
-
return embed_ind
|
173 |
-
|
174 |
-
def postprocess_emb(self, embed_ind, shape):
|
175 |
-
return embed_ind.view(*shape[:-1])
|
176 |
-
|
177 |
-
def dequantize(self, embed_ind):
|
178 |
-
quantize = F.embedding(embed_ind, self.embed)
|
179 |
-
return quantize
|
180 |
-
|
181 |
-
def encode(self, x):
|
182 |
-
shape = x.shape
|
183 |
-
# pre-process
|
184 |
-
x = self.preprocess(x)
|
185 |
-
# quantize
|
186 |
-
embed_ind = self.quantize(x)
|
187 |
-
# post-process
|
188 |
-
embed_ind = self.postprocess_emb(embed_ind, shape)
|
189 |
-
return embed_ind
|
190 |
-
|
191 |
-
def decode(self, embed_ind):
|
192 |
-
quantize = self.dequantize(embed_ind)
|
193 |
-
return quantize
|
194 |
-
|
195 |
-
def forward(self, x):
|
196 |
-
shape, dtype = x.shape, x.dtype
|
197 |
-
x = self.preprocess(x)
|
198 |
-
self.init_embed_(x)
|
199 |
-
|
200 |
-
embed_ind = self.quantize(x)
|
201 |
-
embed_onehot = F.one_hot(embed_ind, self.codebook_size).type(dtype)
|
202 |
-
embed_ind = self.postprocess_emb(embed_ind, shape)
|
203 |
-
quantize = self.dequantize(embed_ind)
|
204 |
-
|
205 |
-
if self.training:
|
206 |
-
# We do the expiry of code at that point as buffers are in sync
|
207 |
-
# and all the workers will take the same decision.
|
208 |
-
self.expire_codes_(x)
|
209 |
-
ema_inplace(self.cluster_size, embed_onehot.sum(0), self.decay)
|
210 |
-
embed_sum = x.t() @ embed_onehot
|
211 |
-
ema_inplace(self.embed_avg, embed_sum.t(), self.decay)
|
212 |
-
cluster_size = (
|
213 |
-
laplace_smoothing(self.cluster_size, self.codebook_size, self.epsilon)
|
214 |
-
* self.cluster_size.sum()
|
215 |
-
)
|
216 |
-
embed_normalized = self.embed_avg / cluster_size.unsqueeze(1)
|
217 |
-
self.embed.data.copy_(embed_normalized)
|
218 |
-
|
219 |
-
return quantize, embed_ind
|
220 |
-
|
221 |
-
|
222 |
-
class VectorQuantization(nn.Module):
|
223 |
-
"""Vector quantization implementation.
|
224 |
-
Currently supports only euclidean distance.
|
225 |
-
|
226 |
-
Args:
|
227 |
-
dim (int): Dimension
|
228 |
-
codebook_size (int): Codebook size
|
229 |
-
codebook_dim (int): Codebook dimension. If not defined, uses the specified dimension in dim.
|
230 |
-
decay (float): Decay for exponential moving average over the codebooks.
|
231 |
-
epsilon (float): Epsilon value for numerical stability.
|
232 |
-
kmeans_init (bool): Whether to use kmeans to initialize the codebooks.
|
233 |
-
kmeans_iters (int): Number of iterations used for kmeans initialization.
|
234 |
-
threshold_ema_dead_code (int):
|
235 |
-
channels_last (bool): Channels are the last dimension in the input tensors.
|
236 |
-
commitment_weight (float): Weight for commitment loss.
|
237 |
-
orthogonal_reg_weight (float): Orthogonal regularization weights.
|
238 |
-
orthogonal_reg_active_codes_only (bool): Apply orthogonal regularization only on active codes.
|
239 |
-
orthogonal_reg_max_codes (optional int): Maximum number of codes to consider
|
240 |
-
for orthogonal regularization.
|
241 |
-
threshold_ema_dead_code (int): Threshold for dead code expiration. Replace any codes
|
242 |
-
that have an exponential moving average cluster size less than the specified threshold with
|
243 |
-
randomly selected vector from the current batch.
|
244 |
-
"""
|
245 |
-
def __init__(
|
246 |
-
self,
|
247 |
-
dim: int,
|
248 |
-
codebook_size: int,
|
249 |
-
codebook_dim: tp.Optional[int] = None,
|
250 |
-
decay: float = 0.8,
|
251 |
-
epsilon: float = 1e-5,
|
252 |
-
kmeans_init: bool = False,
|
253 |
-
kmeans_iters: int = 10,
|
254 |
-
threshold_ema_dead_code: int = 2,
|
255 |
-
channels_last: bool = False,
|
256 |
-
commitment_weight: float = 1.,
|
257 |
-
orthogonal_reg_weight: float = 0.0,
|
258 |
-
orthogonal_reg_active_codes_only: bool = False,
|
259 |
-
orthogonal_reg_max_codes: tp.Optional[int] = None,
|
260 |
-
):
|
261 |
-
super().__init__()
|
262 |
-
_codebook_dim: int = default(codebook_dim, dim)
|
263 |
-
|
264 |
-
requires_projection = _codebook_dim != dim
|
265 |
-
self.project_in = (nn.Linear(dim, _codebook_dim) if requires_projection else nn.Identity())
|
266 |
-
self.project_out = (nn.Linear(_codebook_dim, dim) if requires_projection else nn.Identity())
|
267 |
-
|
268 |
-
self.epsilon = epsilon
|
269 |
-
self.commitment_weight = commitment_weight
|
270 |
-
|
271 |
-
self.orthogonal_reg_weight = orthogonal_reg_weight
|
272 |
-
self.orthogonal_reg_active_codes_only = orthogonal_reg_active_codes_only
|
273 |
-
self.orthogonal_reg_max_codes = orthogonal_reg_max_codes
|
274 |
-
|
275 |
-
self._codebook = EuclideanCodebook(dim=_codebook_dim, codebook_size=codebook_size,
|
276 |
-
kmeans_init=kmeans_init, kmeans_iters=kmeans_iters,
|
277 |
-
decay=decay, epsilon=epsilon,
|
278 |
-
threshold_ema_dead_code=threshold_ema_dead_code)
|
279 |
-
self.codebook_size = codebook_size
|
280 |
-
|
281 |
-
self.channels_last = channels_last
|
282 |
-
|
283 |
-
@property
|
284 |
-
def codebook(self):
|
285 |
-
return self._codebook.embed
|
286 |
-
|
287 |
-
@property
|
288 |
-
def inited(self):
|
289 |
-
return self._codebook.inited
|
290 |
-
|
291 |
-
def _preprocess(self, x):
|
292 |
-
if not self.channels_last:
|
293 |
-
x = rearrange(x, "b d n -> b n d")
|
294 |
-
return x
|
295 |
-
|
296 |
-
def _postprocess(self, quantize):
|
297 |
-
if not self.channels_last:
|
298 |
-
quantize = rearrange(quantize, "b n d -> b d n")
|
299 |
-
return quantize
|
300 |
-
|
301 |
-
def encode(self, x):
|
302 |
-
x = self._preprocess(x)
|
303 |
-
x = self.project_in(x)
|
304 |
-
embed_in = self._codebook.encode(x)
|
305 |
-
return embed_in
|
306 |
-
|
307 |
-
def decode(self, embed_ind):
|
308 |
-
quantize = self._codebook.decode(embed_ind)
|
309 |
-
quantize = self.project_out(quantize)
|
310 |
-
quantize = self._postprocess(quantize)
|
311 |
-
return quantize
|
312 |
-
|
313 |
-
def forward(self, x):
|
314 |
-
device = x.device
|
315 |
-
x = self._preprocess(x)
|
316 |
-
|
317 |
-
x = self.project_in(x)
|
318 |
-
quantize, embed_ind = self._codebook(x)
|
319 |
-
|
320 |
-
if self.training:
|
321 |
-
quantize = x + (quantize - x).detach()
|
322 |
-
|
323 |
-
loss = torch.tensor([0.0], device=device, requires_grad=self.training)
|
324 |
-
|
325 |
-
if self.training:
|
326 |
-
if self.commitment_weight > 0:
|
327 |
-
commit_loss = F.mse_loss(quantize.detach(), x)
|
328 |
-
loss = loss + commit_loss * self.commitment_weight
|
329 |
-
|
330 |
-
if self.orthogonal_reg_weight > 0:
|
331 |
-
codebook = self.codebook
|
332 |
-
|
333 |
-
if self.orthogonal_reg_active_codes_only:
|
334 |
-
# only calculate orthogonal loss for the activated codes for this batch
|
335 |
-
unique_code_ids = torch.unique(embed_ind)
|
336 |
-
codebook = codebook[unique_code_ids]
|
337 |
-
|
338 |
-
num_codes = codebook.shape[0]
|
339 |
-
if exists(self.orthogonal_reg_max_codes) and num_codes > self.orthogonal_reg_max_codes:
|
340 |
-
rand_ids = torch.randperm(num_codes, device=device)[:self.orthogonal_reg_max_codes]
|
341 |
-
codebook = codebook[rand_ids]
|
342 |
-
|
343 |
-
orthogonal_reg_loss = orthogonal_loss_fn(codebook)
|
344 |
-
loss = loss + orthogonal_reg_loss * self.orthogonal_reg_weight
|
345 |
-
|
346 |
-
quantize = self.project_out(quantize)
|
347 |
-
quantize = self._postprocess(quantize)
|
348 |
-
|
349 |
-
return quantize, embed_ind, loss
|
350 |
-
|
351 |
-
|
352 |
-
class ResidualVectorQuantization(nn.Module):
|
353 |
-
"""Residual vector quantization implementation.
|
354 |
-
|
355 |
-
Follows Algorithm 1. in https://arxiv.org/pdf/2107.03312.pdf
|
356 |
-
"""
|
357 |
-
def __init__(self, *, num_quantizers, **kwargs):
|
358 |
-
super().__init__()
|
359 |
-
self.layers = nn.ModuleList(
|
360 |
-
[VectorQuantization(**kwargs) for _ in range(num_quantizers)]
|
361 |
-
)
|
362 |
-
|
363 |
-
def forward(self, x, n_q: tp.Optional[int] = None):
|
364 |
-
quantized_out = 0.0
|
365 |
-
residual = x
|
366 |
-
|
367 |
-
all_losses = []
|
368 |
-
all_indices = []
|
369 |
-
|
370 |
-
n_q = n_q or len(self.layers)
|
371 |
-
|
372 |
-
for i, layer in enumerate(self.layers[:n_q]):
|
373 |
-
quantized, indices, loss = layer(residual)
|
374 |
-
quantized = quantized.detach()
|
375 |
-
residual = residual - quantized
|
376 |
-
quantized_out = quantized_out + quantized
|
377 |
-
all_indices.append(indices)
|
378 |
-
all_losses.append(loss)
|
379 |
-
|
380 |
-
if self.training:
|
381 |
-
# Solving subtle bug with STE and RVQ: https://github.com/facebookresearch/encodec/issues/25
|
382 |
-
quantized_out = x + (quantized_out - x).detach()
|
383 |
-
|
384 |
-
out_losses, out_indices = map(torch.stack, (all_losses, all_indices))
|
385 |
-
return quantized_out, out_indices, out_losses
|
386 |
-
|
387 |
-
def encode(self, x: torch.Tensor, n_q: tp.Optional[int] = None) -> torch.Tensor:
|
388 |
-
residual = x
|
389 |
-
all_indices = []
|
390 |
-
n_q = n_q or len(self.layers)
|
391 |
-
for layer in self.layers[:n_q]:
|
392 |
-
indices = layer.encode(residual)
|
393 |
-
quantized = layer.decode(indices)
|
394 |
-
residual = residual - quantized
|
395 |
-
all_indices.append(indices)
|
396 |
-
out_indices = torch.stack(all_indices)
|
397 |
-
return out_indices
|
398 |
-
|
399 |
-
def decode(self, q_indices: torch.Tensor) -> torch.Tensor:
|
400 |
-
quantized_out = torch.tensor(0.0, device=q_indices.device)
|
401 |
-
for i, indices in enumerate(q_indices):
|
402 |
-
layer = self.layers[i]
|
403 |
-
quantized = layer.decode(indices)
|
404 |
-
quantized_out = quantized_out + quantized
|
405 |
-
return quantized_out
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
audiocraft/{quantization/vq.py → vq.py}
RENAMED
@@ -1,19 +1,157 @@
|
|
1 |
-
# Copyright (c) Meta Platforms, Inc. and affiliates.
|
2 |
-
# All rights reserved.
|
3 |
-
#
|
4 |
-
# This source code is licensed under the license found in the
|
5 |
-
# LICENSE file in the root directory of this source tree.
|
6 |
-
|
7 |
import math
|
8 |
import typing as tp
|
9 |
-
|
|
|
10 |
import torch
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
11 |
|
12 |
-
|
13 |
-
|
|
|
|
|
14 |
|
15 |
|
16 |
-
class ResidualVectorQuantizer(
|
17 |
"""Residual Vector Quantizer.
|
18 |
|
19 |
Args:
|
@@ -59,6 +197,7 @@ class ResidualVectorQuantizer(BaseQuantizer):
|
|
59 |
self.orthogonal_reg_weight = orthogonal_reg_weight
|
60 |
self.orthogonal_reg_active_codes_only = orthogonal_reg_active_codes_only
|
61 |
self.orthogonal_reg_max_codes = orthogonal_reg_max_codes
|
|
|
62 |
self.vq = ResidualVectorQuantization(
|
63 |
dim=self.dimension,
|
64 |
codebook_size=self.bins,
|
@@ -66,10 +205,6 @@ class ResidualVectorQuantizer(BaseQuantizer):
|
|
66 |
decay=self.decay,
|
67 |
kmeans_init=self.kmeans_init,
|
68 |
kmeans_iters=self.kmeans_iters,
|
69 |
-
threshold_ema_dead_code=self.threshold_ema_dead_code,
|
70 |
-
orthogonal_reg_weight=self.orthogonal_reg_weight,
|
71 |
-
orthogonal_reg_active_codes_only=self.orthogonal_reg_active_codes_only,
|
72 |
-
orthogonal_reg_max_codes=self.orthogonal_reg_max_codes,
|
73 |
channels_last=False
|
74 |
)
|
75 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
import math
|
2 |
import typing as tp
|
3 |
+
from dataclasses import dataclass, field
|
4 |
+
import typing as tp
|
5 |
import torch
|
6 |
+
from torch import nn
|
7 |
+
from einops import rearrange
|
8 |
+
import torch.nn.functional as F
|
9 |
+
|
10 |
+
@dataclass
|
11 |
+
class QuantizedResult:
|
12 |
+
x: torch.Tensor
|
13 |
+
codes: torch.Tensor
|
14 |
+
bandwidth: torch.Tensor # bandwidth in kb/s used, per batch item.
|
15 |
+
penalty: tp.Optional[torch.Tensor] = None
|
16 |
+
metrics: dict = field(default_factory=dict)
|
17 |
+
|
18 |
+
|
19 |
+
|
20 |
+
|
21 |
+
|
22 |
+
|
23 |
+
class EuclideanCodebook(nn.Module):
|
24 |
+
def __init__(
|
25 |
+
self,
|
26 |
+
dim,
|
27 |
+
codebook_size,
|
28 |
+
kmeans_init=False,
|
29 |
+
kmeans_iters=10,
|
30 |
+
decay=0.8,
|
31 |
+
epsilon=1e-5,
|
32 |
+
):
|
33 |
+
super().__init__()
|
34 |
+
self.decay=decay
|
35 |
+
init_fn=uniform_init if not kmeans_init else torch.zeros
|
36 |
+
embed = init_fn(codebook_size, dim)
|
37 |
+
|
38 |
+
self.codebook_size = codebook_size
|
39 |
+
|
40 |
+
self.kmeans_iters = kmeans_iters
|
41 |
+
self.epsilon = epsilon
|
42 |
+
|
43 |
+
self.register_buffer("inited", torch.Tensor([not kmeans_init]))
|
44 |
+
self.register_buffer("cluster_size", torch.zeros(codebook_size))
|
45 |
+
self.register_buffer("embed", embed)
|
46 |
+
self.register_buffer("embed_avg", embed.clone())
|
47 |
+
|
48 |
+
@torch.jit.ignore
|
49 |
+
def init_embed_(self, data):
|
50 |
+
if self.inited:
|
51 |
+
return
|
52 |
+
|
53 |
+
embed, cluster_size = kmeans(data, self.codebook_size, self.kmeans_iters)
|
54 |
+
self.embed.data.copy_(embed)
|
55 |
+
self.embed_avg.data.copy_(embed.clone())
|
56 |
+
self.cluster_size.data.copy_(cluster_size)
|
57 |
+
self.inited.data.copy_(torch.Tensor([True]))
|
58 |
+
# Make sure all buffers across workers are in sync after initialization
|
59 |
+
# flashy.distrib.broadcast_tensors(self.buffers()) # brodcast param values to all GPUS
|
60 |
+
|
61 |
+
|
62 |
+
|
63 |
+
def postprocess_emb(self, embed_ind, shape):
|
64 |
+
return embed_ind.view(*shape[:-1])
|
65 |
+
|
66 |
+
def dequantize(self, embed_ind):
|
67 |
+
quantize = F.embedding(embed_ind, self.embed)
|
68 |
+
# print('\n\nDE QUANT\n\n', quantize.shape) # (1, 35, 128) -> also arrives here for special_token
|
69 |
+
return quantize
|
70 |
+
|
71 |
+
def decode(self, embed_ind):
|
72 |
+
quantize = self.dequantize(embed_ind)
|
73 |
+
return quantize
|
74 |
+
|
75 |
+
|
76 |
+
|
77 |
+
class VectorQuantization(nn.Module):
|
78 |
+
|
79 |
+
def __init__(
|
80 |
+
self,
|
81 |
+
dim,
|
82 |
+
codebook_size,
|
83 |
+
codebook_dim=None,
|
84 |
+
decay=0.8,
|
85 |
+
epsilon=1e-5,
|
86 |
+
kmeans_init=False,
|
87 |
+
kmeans_iters=10,
|
88 |
+
channels_last=False,
|
89 |
+
):
|
90 |
+
super().__init__()
|
91 |
+
# _codebook_dim: int = default(codebook_dim, dim)
|
92 |
+
_codebook_dim = codebook_dim if codebook_dim is not None else dim
|
93 |
+
|
94 |
+
requires_projection = _codebook_dim != dim
|
95 |
+
self.project_in = (nn.Linear(dim, _codebook_dim) if requires_projection else nn.Identity())
|
96 |
+
self.project_out = (nn.Linear(_codebook_dim, dim) if requires_projection else nn.Identity())
|
97 |
+
self._codebook = EuclideanCodebook(dim=_codebook_dim,
|
98 |
+
codebook_size=codebook_size,
|
99 |
+
kmeans_init=kmeans_init,
|
100 |
+
kmeans_iters=kmeans_iters,
|
101 |
+
decay=decay,
|
102 |
+
epsilon=epsilon)
|
103 |
+
self.codebook_size = codebook_size
|
104 |
+
|
105 |
+
self.channels_last = channels_last
|
106 |
+
|
107 |
+
@property
|
108 |
+
def codebook(self):
|
109 |
+
return self._codebook.embed
|
110 |
+
|
111 |
+
@property
|
112 |
+
def inited(self):
|
113 |
+
return self._codebook.inited
|
114 |
+
|
115 |
+
def _postprocess(self, quantize):
|
116 |
+
if not self.channels_last:
|
117 |
+
quantize = rearrange(quantize, "b n d -> b d n")
|
118 |
+
return quantize
|
119 |
+
|
120 |
+
def decode(self, embed_ind):
|
121 |
+
quantize = self._codebook.decode(embed_ind)
|
122 |
+
quantize = self.project_out(quantize)
|
123 |
+
quantize = self._postprocess(quantize)
|
124 |
+
return quantize
|
125 |
+
|
126 |
+
|
127 |
+
|
128 |
+
|
129 |
+
class ResidualVectorQuantization(nn.Module):
|
130 |
+
"""Residual vector quantization implementation.
|
131 |
+
|
132 |
+
Follows Algorithm 1. in https://arxiv.org/pdf/2107.03312.pdf
|
133 |
+
"""
|
134 |
+
def __init__(self, *, num_quantizers, **kwargs):
|
135 |
+
super().__init__()
|
136 |
+
self.layers = nn.ModuleList(
|
137 |
+
[VectorQuantization(**kwargs) for _ in range(num_quantizers)]
|
138 |
+
)
|
139 |
+
|
140 |
+
def decode(self, q_indices: torch.Tensor) -> torch.Tensor:
|
141 |
+
quantized_out = torch.tensor(0.0, device=q_indices.device)
|
142 |
+
for i, indices in enumerate(q_indices):
|
143 |
+
layer = self.layers[i]
|
144 |
+
quantized = layer.decode(indices)
|
145 |
+
quantized_out = quantized_out + quantized
|
146 |
+
return quantized_out
|
147 |
|
148 |
+
|
149 |
+
|
150 |
+
|
151 |
+
# ------------------------------------- END core_vq.py
|
152 |
|
153 |
|
154 |
+
class ResidualVectorQuantizer(nn.Module):
|
155 |
"""Residual Vector Quantizer.
|
156 |
|
157 |
Args:
|
|
|
197 |
self.orthogonal_reg_weight = orthogonal_reg_weight
|
198 |
self.orthogonal_reg_active_codes_only = orthogonal_reg_active_codes_only
|
199 |
self.orthogonal_reg_max_codes = orthogonal_reg_max_codes
|
200 |
+
print(f' {kmeans_init=}\n\n\n\n')
|
201 |
self.vq = ResidualVectorQuantization(
|
202 |
dim=self.dimension,
|
203 |
codebook_size=self.bins,
|
|
|
205 |
decay=self.decay,
|
206 |
kmeans_init=self.kmeans_init,
|
207 |
kmeans_iters=self.kmeans_iters,
|
|
|
|
|
|
|
|
|
208 |
channels_last=False
|
209 |
)
|
210 |
|
demo.py
CHANGED
@@ -1,12 +1,13 @@
|
|
1 |
from audiocraft.audiogen import AudioGen #, audio_write
|
2 |
-
|
|
|
3 |
|
4 |
print('\n\n\n\n___________________')
|
5 |
|
6 |
-
txt = '
|
7 |
|
8 |
sound_generator = AudioGen.get_pretrained('facebook/audiogen-medium')
|
9 |
-
sound_generator.set_generation_params(duration
|
10 |
|
11 |
x = sound_generator.generate([txt])[0].detach().cpu().numpy()[0, :]
|
12 |
x /= np.abs(x).max() + 1e-7
|
|
|
1 |
from audiocraft.audiogen import AudioGen #, audio_write
|
2 |
+
import audiofile
|
3 |
+
import numpy as np
|
4 |
|
5 |
print('\n\n\n\n___________________')
|
6 |
|
7 |
+
txt = 'sea waves rock crash pirates'
|
8 |
|
9 |
sound_generator = AudioGen.get_pretrained('facebook/audiogen-medium')
|
10 |
+
sound_generator.set_generation_params(duration=.7) # why is generating so long at 14 seconds
|
11 |
|
12 |
x = sound_generator.generate([txt])[0].detach().cpu().numpy()[0, :]
|
13 |
x /= np.abs(x).max() + 1e-7
|