File size: 17,424 Bytes
11e6f7b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
import logging
from abc import abstractmethod
from typing import Dict, Iterator, Literal, Optional, Tuple, Union

import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F
from einops import rearrange
from torch import einsum

from .base import AbstractRegularizer, measure_perplexity

logpy = logging.getLogger(__name__)


class AbstractQuantizer(AbstractRegularizer):
    def __init__(self):
        super().__init__()
        # Define these in your init
        # shape (N,)
        self.used: Optional[torch.Tensor]
        self.re_embed: int
        self.unknown_index: Union[Literal["random"], int]

    def remap_to_used(self, inds: torch.Tensor) -> torch.Tensor:
        assert self.used is not None, "You need to define used indices for remap"
        ishape = inds.shape
        assert len(ishape) > 1
        inds = inds.reshape(ishape[0], -1)
        used = self.used.to(inds)
        match = (inds[:, :, None] == used[None, None, ...]).long()
        new = match.argmax(-1)
        unknown = match.sum(2) < 1
        if self.unknown_index == "random":
            new[unknown] = torch.randint(0, self.re_embed, size=new[unknown].shape).to(
                device=new.device
            )
        else:
            new[unknown] = self.unknown_index
        return new.reshape(ishape)

    def unmap_to_all(self, inds: torch.Tensor) -> torch.Tensor:
        assert self.used is not None, "You need to define used indices for remap"
        ishape = inds.shape
        assert len(ishape) > 1
        inds = inds.reshape(ishape[0], -1)
        used = self.used.to(inds)
        if self.re_embed > self.used.shape[0]:  # extra token
            inds[inds >= self.used.shape[0]] = 0  # simply set to zero
        back = torch.gather(used[None, :][inds.shape[0] * [0], :], 1, inds)
        return back.reshape(ishape)

    @abstractmethod
    def get_codebook_entry(
        self, indices: torch.Tensor, shape: Optional[Tuple[int, ...]] = None
    ) -> torch.Tensor:
        raise NotImplementedError()

    def get_trainable_parameters(self) -> Iterator[torch.nn.Parameter]:
        yield from self.parameters()


class GumbelQuantizer(AbstractQuantizer):
    """
    credit to @karpathy:
    https://github.com/karpathy/deep-vector-quantization/blob/main/model.py (thanks!)
    Gumbel Softmax trick quantizer
    Categorical Reparameterization with Gumbel-Softmax, Jang et al. 2016
    https://arxiv.org/abs/1611.01144
    """

    def __init__(
        self,
        num_hiddens: int,
        embedding_dim: int,
        n_embed: int,
        straight_through: bool = True,
        kl_weight: float = 5e-4,
        temp_init: float = 1.0,
        remap: Optional[str] = None,
        unknown_index: str = "random",
        loss_key: str = "loss/vq",
    ) -> None:
        super().__init__()

        self.loss_key = loss_key
        self.embedding_dim = embedding_dim
        self.n_embed = n_embed

        self.straight_through = straight_through
        self.temperature = temp_init
        self.kl_weight = kl_weight

        self.proj = nn.Conv2d(num_hiddens, n_embed, 1)
        self.embed = nn.Embedding(n_embed, embedding_dim)

        self.remap = remap
        if self.remap is not None:
            self.register_buffer("used", torch.tensor(np.load(self.remap)))
            self.re_embed = self.used.shape[0]
        else:
            self.used = None
            self.re_embed = n_embed
        if unknown_index == "extra":
            self.unknown_index = self.re_embed
            self.re_embed = self.re_embed + 1
        else:
            assert unknown_index == "random" or isinstance(
                unknown_index, int
            ), "unknown index needs to be 'random', 'extra' or any integer"
            self.unknown_index = unknown_index  # "random" or "extra" or integer
        if self.remap is not None:
            logpy.info(
                f"Remapping {self.n_embed} indices to {self.re_embed} indices. "
                f"Using {self.unknown_index} for unknown indices."
            )

    def forward(
        self, z: torch.Tensor, temp: Optional[float] = None, return_logits: bool = False
    ) -> Tuple[torch.Tensor, Dict]:
        # force hard = True when we are in eval mode, as we must quantize.
        # actually, always true seems to work
        hard = self.straight_through if self.training else True
        temp = self.temperature if temp is None else temp
        out_dict = {}
        logits = self.proj(z)
        if self.remap is not None:
            # continue only with used logits
            full_zeros = torch.zeros_like(logits)
            logits = logits[:, self.used, ...]

        soft_one_hot = F.gumbel_softmax(logits, tau=temp, dim=1, hard=hard)
        if self.remap is not None:
            # go back to all entries but unused set to zero
            full_zeros[:, self.used, ...] = soft_one_hot
            soft_one_hot = full_zeros
        z_q = einsum("b n h w, n d -> b d h w", soft_one_hot, self.embed.weight)

        # + kl divergence to the prior loss
        qy = F.softmax(logits, dim=1)
        diff = (
            self.kl_weight
            * torch.sum(qy * torch.log(qy * self.n_embed + 1e-10), dim=1).mean()
        )
        out_dict[self.loss_key] = diff

        ind = soft_one_hot.argmax(dim=1)
        out_dict["indices"] = ind
        if self.remap is not None:
            ind = self.remap_to_used(ind)

        if return_logits:
            out_dict["logits"] = logits

        return z_q, out_dict

    def get_codebook_entry(self, indices, shape):
        # TODO: shape not yet optional
        b, h, w, c = shape
        assert b * h * w == indices.shape[0]
        indices = rearrange(indices, "(b h w) -> b h w", b=b, h=h, w=w)
        if self.remap is not None:
            indices = self.unmap_to_all(indices)
        one_hot = (
            F.one_hot(indices, num_classes=self.n_embed).permute(0, 3, 1, 2).float()
        )
        z_q = einsum("b n h w, n d -> b d h w", one_hot, self.embed.weight)
        return z_q


class VectorQuantizer(AbstractQuantizer):
    """
    ____________________________________________
    Discretization bottleneck part of the VQ-VAE.
    Inputs:
    - n_e : number of embeddings
    - e_dim : dimension of embedding
    - beta : commitment cost used in loss term,
        beta * ||z_e(x)-sg[e]||^2
    _____________________________________________
    """

    def __init__(
        self,
        n_e: int,
        e_dim: int,
        beta: float = 0.25,
        remap: Optional[str] = None,
        unknown_index: str = "random",
        sane_index_shape: bool = False,
        log_perplexity: bool = False,
        embedding_weight_norm: bool = False,
        loss_key: str = "loss/vq",
    ):
        super().__init__()
        self.n_e = n_e
        self.e_dim = e_dim
        self.beta = beta
        self.loss_key = loss_key

        if not embedding_weight_norm:
            self.embedding = nn.Embedding(self.n_e, self.e_dim)
            self.embedding.weight.data.uniform_(-1.0 / self.n_e, 1.0 / self.n_e)
        else:
            self.embedding = torch.nn.utils.weight_norm(
                nn.Embedding(self.n_e, self.e_dim), dim=1
            )

        self.remap = remap
        if self.remap is not None:
            self.register_buffer("used", torch.tensor(np.load(self.remap)))
            self.re_embed = self.used.shape[0]
        else:
            self.used = None
            self.re_embed = n_e
        if unknown_index == "extra":
            self.unknown_index = self.re_embed
            self.re_embed = self.re_embed + 1
        else:
            assert unknown_index == "random" or isinstance(
                unknown_index, int
            ), "unknown index needs to be 'random', 'extra' or any integer"
            self.unknown_index = unknown_index  # "random" or "extra" or integer
        if self.remap is not None:
            logpy.info(
                f"Remapping {self.n_e} indices to {self.re_embed} indices. "
                f"Using {self.unknown_index} for unknown indices."
            )

        self.sane_index_shape = sane_index_shape
        self.log_perplexity = log_perplexity

    def forward(
        self,
        z: torch.Tensor,
    ) -> Tuple[torch.Tensor, Dict]:
        do_reshape = z.ndim == 4
        if do_reshape:
            #     # reshape z -> (batch, height, width, channel) and flatten
            z = rearrange(z, "b c h w -> b h w c").contiguous()

        else:
            assert z.ndim < 4, "No reshaping strategy for inputs > 4 dimensions defined"
            z = z.contiguous()

        z_flattened = z.view(-1, self.e_dim)
        # distances from z to embeddings e_j (z - e)^2 = z^2 + e^2 - 2 e * z

        d = (
            torch.sum(z_flattened**2, dim=1, keepdim=True)
            + torch.sum(self.embedding.weight**2, dim=1)
            - 2
            * torch.einsum(
                "bd,dn->bn", z_flattened, rearrange(self.embedding.weight, "n d -> d n")
            )
        )

        min_encoding_indices = torch.argmin(d, dim=1)
        z_q = self.embedding(min_encoding_indices).view(z.shape)
        loss_dict = {}
        if self.log_perplexity:
            perplexity, cluster_usage = measure_perplexity(
                min_encoding_indices.detach(), self.n_e
            )
            loss_dict.update({"perplexity": perplexity, "cluster_usage": cluster_usage})

        # compute loss for embedding
        loss = self.beta * torch.mean((z_q.detach() - z) ** 2) + torch.mean(
            (z_q - z.detach()) ** 2
        )
        loss_dict[self.loss_key] = loss

        # preserve gradients
        z_q = z + (z_q - z).detach()

        # reshape back to match original input shape
        if do_reshape:
            z_q = rearrange(z_q, "b h w c -> b c h w").contiguous()

        if self.remap is not None:
            min_encoding_indices = min_encoding_indices.reshape(
                z.shape[0], -1
            )  # add batch axis
            min_encoding_indices = self.remap_to_used(min_encoding_indices)
            min_encoding_indices = min_encoding_indices.reshape(-1, 1)  # flatten

        if self.sane_index_shape:
            if do_reshape:
                min_encoding_indices = min_encoding_indices.reshape(
                    z_q.shape[0], z_q.shape[2], z_q.shape[3]
                )
            else:
                min_encoding_indices = rearrange(
                    min_encoding_indices, "(b s) 1 -> b s", b=z_q.shape[0]
                )

        loss_dict["min_encoding_indices"] = min_encoding_indices

        return z_q, loss_dict

    def get_codebook_entry(
        self, indices: torch.Tensor, shape: Optional[Tuple[int, ...]] = None
    ) -> torch.Tensor:
        # shape specifying (batch, height, width, channel)
        if self.remap is not None:
            assert shape is not None, "Need to give shape for remap"
            indices = indices.reshape(shape[0], -1)  # add batch axis
            indices = self.unmap_to_all(indices)
            indices = indices.reshape(-1)  # flatten again

        # get quantized latent vectors
        z_q = self.embedding(indices)

        if shape is not None:
            z_q = z_q.view(shape)
            # reshape back to match original input shape
            z_q = z_q.permute(0, 3, 1, 2).contiguous()

        return z_q


class EmbeddingEMA(nn.Module):
    def __init__(self, num_tokens, codebook_dim, decay=0.99, eps=1e-5):
        super().__init__()
        self.decay = decay
        self.eps = eps
        weight = torch.randn(num_tokens, codebook_dim)
        self.weight = nn.Parameter(weight, requires_grad=False)
        self.cluster_size = nn.Parameter(torch.zeros(num_tokens), requires_grad=False)
        self.embed_avg = nn.Parameter(weight.clone(), requires_grad=False)
        self.update = True

    def forward(self, embed_id):
        return F.embedding(embed_id, self.weight)

    def cluster_size_ema_update(self, new_cluster_size):
        self.cluster_size.data.mul_(self.decay).add_(
            new_cluster_size, alpha=1 - self.decay
        )

    def embed_avg_ema_update(self, new_embed_avg):
        self.embed_avg.data.mul_(self.decay).add_(new_embed_avg, alpha=1 - self.decay)

    def weight_update(self, num_tokens):
        n = self.cluster_size.sum()
        smoothed_cluster_size = (
            (self.cluster_size + self.eps) / (n + num_tokens * self.eps) * n
        )
        # normalize embedding average with smoothed cluster size
        embed_normalized = self.embed_avg / smoothed_cluster_size.unsqueeze(1)
        self.weight.data.copy_(embed_normalized)


class EMAVectorQuantizer(AbstractQuantizer):
    def __init__(
        self,
        n_embed: int,
        embedding_dim: int,
        beta: float,
        decay: float = 0.99,
        eps: float = 1e-5,
        remap: Optional[str] = None,
        unknown_index: str = "random",
        loss_key: str = "loss/vq",
    ):
        super().__init__()
        self.codebook_dim = embedding_dim
        self.num_tokens = n_embed
        self.beta = beta
        self.loss_key = loss_key

        self.embedding = EmbeddingEMA(self.num_tokens, self.codebook_dim, decay, eps)

        self.remap = remap
        if self.remap is not None:
            self.register_buffer("used", torch.tensor(np.load(self.remap)))
            self.re_embed = self.used.shape[0]
        else:
            self.used = None
            self.re_embed = n_embed
        if unknown_index == "extra":
            self.unknown_index = self.re_embed
            self.re_embed = self.re_embed + 1
        else:
            assert unknown_index == "random" or isinstance(
                unknown_index, int
            ), "unknown index needs to be 'random', 'extra' or any integer"
            self.unknown_index = unknown_index  # "random" or "extra" or integer
        if self.remap is not None:
            logpy.info(
                f"Remapping {self.n_embed} indices to {self.re_embed} indices. "
                f"Using {self.unknown_index} for unknown indices."
            )

    def forward(self, z: torch.Tensor) -> Tuple[torch.Tensor, Dict]:
        # reshape z -> (batch, height, width, channel) and flatten
        # z, 'b c h w -> b h w c'
        z = rearrange(z, "b c h w -> b h w c")
        z_flattened = z.reshape(-1, self.codebook_dim)

        # distances from z to embeddings e_j (z - e)^2 = z^2 + e^2 - 2 e * z
        d = (
            z_flattened.pow(2).sum(dim=1, keepdim=True)
            + self.embedding.weight.pow(2).sum(dim=1)
            - 2 * torch.einsum("bd,nd->bn", z_flattened, self.embedding.weight)
        )  # 'n d -> d n'

        encoding_indices = torch.argmin(d, dim=1)

        z_q = self.embedding(encoding_indices).view(z.shape)
        encodings = F.one_hot(encoding_indices, self.num_tokens).type(z.dtype)
        avg_probs = torch.mean(encodings, dim=0)
        perplexity = torch.exp(-torch.sum(avg_probs * torch.log(avg_probs + 1e-10)))

        if self.training and self.embedding.update:
            # EMA cluster size
            encodings_sum = encodings.sum(0)
            self.embedding.cluster_size_ema_update(encodings_sum)
            # EMA embedding average
            embed_sum = encodings.transpose(0, 1) @ z_flattened
            self.embedding.embed_avg_ema_update(embed_sum)
            # normalize embed_avg and update weight
            self.embedding.weight_update(self.num_tokens)

        # compute loss for embedding
        loss = self.beta * F.mse_loss(z_q.detach(), z)

        # preserve gradients
        z_q = z + (z_q - z).detach()

        # reshape back to match original input shape
        # z_q, 'b h w c -> b c h w'
        z_q = rearrange(z_q, "b h w c -> b c h w")

        out_dict = {
            self.loss_key: loss,
            "encodings": encodings,
            "encoding_indices": encoding_indices,
            "perplexity": perplexity,
        }

        return z_q, out_dict


class VectorQuantizerWithInputProjection(VectorQuantizer):
    def __init__(
        self,
        input_dim: int,
        n_codes: int,
        codebook_dim: int,
        beta: float = 1.0,
        output_dim: Optional[int] = None,
        **kwargs,
    ):
        super().__init__(n_codes, codebook_dim, beta, **kwargs)
        self.proj_in = nn.Linear(input_dim, codebook_dim)
        self.output_dim = output_dim
        if output_dim is not None:
            self.proj_out = nn.Linear(codebook_dim, output_dim)
        else:
            self.proj_out = nn.Identity()

    def forward(self, z: torch.Tensor) -> Tuple[torch.Tensor, Dict]:
        rearr = False
        in_shape = z.shape

        if z.ndim > 3:
            rearr = self.output_dim is not None
            z = rearrange(z, "b c ... -> b (...) c")
        z = self.proj_in(z)
        z_q, loss_dict = super().forward(z)

        z_q = self.proj_out(z_q)
        if rearr:
            if len(in_shape) == 4:
                z_q = rearrange(z_q, "b (h w) c -> b c h w ", w=in_shape[-1])
            elif len(in_shape) == 5:
                z_q = rearrange(
                    z_q, "b (t h w) c -> b c t h w ", w=in_shape[-1], h=in_shape[-2]
                )
            else:
                raise NotImplementedError(
                    f"rearranging not available for {len(in_shape)}-dimensional input."
                )

        return z_q, loss_dict