File size: 33,748 Bytes
663494c | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 | # Copyright (c) OpenMMLab. All rights reserved.
import copy
import warnings
import torch
import torch.nn as nn
import torch.utils.checkpoint as cp
import torch.nn.functional as F
from mmcv import ConfigDict, deprecated_api_warning
from mmcv.cnn.bricks.wrappers import Linear
from mmcv.cnn.bricks.activation import build_activation_layer
from mmcv.cnn import build_norm_layer
from mmcv.cnn.bricks.drop import build_dropout
# from mmcv.models.bricks import Linear, build_activation_layer, build_norm_layer
from mmcv.runner.base_module import BaseModule, ModuleList, Sequential
from mmcv.utils import build_from_cfg
from mmcv.cnn.bricks.registry import (ATTENTION, FEEDFORWARD_NETWORK, POSITIONAL_ENCODING,
TRANSFORMER_LAYER, TRANSFORMER_LAYER_SEQUENCE)
from mmdet3d_plugin.uniad.custom_modules.peft import (LoRALinear, ZeroAdapter, LoRACLAdapter, LoRAMoECLAdapter, MOELoRALinear,
finetuning_detach, frozen_grad, peft_wrapper_forward, lora_wrapper)
from mmdet3d_plugin.uniad.custom_modules.custom_mha_function import custom_multi_head_attention_forward
def build_positional_encoding(cfg, default_args=None):
"""Builder for Position Encoding."""
return build_from_cfg(cfg, POSITIONAL_ENCODING, default_args)
def build_attention(cfg, default_args=None):
"""Builder for attention."""
return build_from_cfg(cfg, ATTENTION, default_args)
def build_feedforward_network(cfg, default_args=None):
"""Builder for feed-forward network (FFN)."""
return build_from_cfg(cfg, FEEDFORWARD_NETWORK, default_args)
def build_transformer_layer(cfg, default_args=None):
"""Builder for transformer layer."""
return build_from_cfg(cfg, TRANSFORMER_LAYER, default_args)
def build_transformer_layer_sequence(cfg, default_args=None):
"""Builder for transformer encoder and transformer decoder."""
return build_from_cfg(cfg, TRANSFORMER_LAYER_SEQUENCE, default_args)
@ATTENTION.register_module(force=True)
class MultiheadAttention(BaseModule):
"""A wrapper for ``torch.nn.MultiheadAttention``.
This module implements MultiheadAttention with identity connection,
and positional encoding is also passed as input.
Args:
embed_dims (int): The embedding dimension.
num_heads (int): Parallel attention heads.
attn_drop (float): A Dropout layer on attn_output_weights.
Default: 0.0.
proj_drop (float): A Dropout layer after `nn.MultiheadAttention`.
Default: 0.0.
dropout_layer (obj:`ConfigDict`): The dropout_layer used
when adding the shortcut.
init_cfg (obj:`mmcv.ConfigDict`): The Config for initialization.
Default: None.
batch_first (bool): When it is True, Key, Query and Value are shape of
(batch, n, embed_dim), otherwise (n, batch, embed_dim).
Default to False.
"""
def __init__(self,
embed_dims,
num_heads,
attn_drop=0.,
proj_drop=0.,
dropout_layer=dict(type='Dropout', drop_prob=0.),
init_cfg=None,
batch_first=False,
with_cp=False,
use_lora=False,
multi_lora_task=False,
lora_rank=16,
num_task=6,
moe_lora=False,
**kwargs):
super(MultiheadAttention, self).__init__(init_cfg)
if 'dropout' in kwargs:
warnings.warn('The arguments `dropout` in MultiheadAttention '
'has been deprecated, now you can separately '
'set `attn_drop`(float), proj_drop(float), '
'and `dropout_layer`(dict) ')
attn_drop = kwargs['dropout']
dropout_layer['drop_prob'] = kwargs.pop('dropout')
self.embed_dims = embed_dims
self.num_heads = num_heads
self.batch_first = batch_first
self.moe_lora = moe_lora
self.attn = nn.MultiheadAttention(embed_dims, num_heads, attn_drop,
**kwargs)
self.use_lora = use_lora
self.multi_lora_task = multi_lora_task
self.proj_drop = nn.Dropout(proj_drop)
self.dropout_layer = build_dropout(
dropout_layer) if dropout_layer else nn.Identity()
self.with_cp = with_cp
if self.use_lora:
# freeze param in attn
for param in self.attn.parameters():
param.requires_grad = False
if self.multi_lora_task:
self.q_lora = LoRALinear(embed_dims, embed_dims, r=lora_rank)
self.k_lora = LoRACLAdapter(embed_dims, embed_dims, num_task=num_task, r=lora_rank)
self.v_lora = LoRACLAdapter(embed_dims, embed_dims, num_task=num_task, r=lora_rank)
elif moe_lora:
self.q_lora = MOELoRALinear(embed_dims, embed_dims, num_task=num_task, r=lora_rank)
self.k_lora = MOELoRALinear(embed_dims, embed_dims, num_task=num_task, r=lora_rank)
self.v_lora = MOELoRALinear(embed_dims, embed_dims, num_task=num_task, r=lora_rank)
else:
self.q_lora = LoRALinear(embed_dims, embed_dims, r=lora_rank)
self.k_lora = LoRALinear(embed_dims, embed_dims, r=lora_rank)
self.v_lora = LoRALinear(embed_dims, embed_dims, r=lora_rank)
if self.moe_lora:
self.out_lora = MOELoRALinear(embed_dims, embed_dims, num_task=num_task, r=lora_rank)
else:
self.out_lora = LoRALinear(embed_dims, embed_dims, r=lora_rank)
self.q_proj_weight = self.attn.in_proj_weight[:embed_dims, :]
self.k_proj_weight = self.attn.in_proj_weight[embed_dims:2*embed_dims, :]
self.v_proj_weight = self.attn.in_proj_weight[2*embed_dims:, :]
self.q_proj_bias = self.attn.in_proj_bias[:embed_dims]
self.k_proj_bias = self.attn.in_proj_bias[embed_dims:2*embed_dims]
self.v_proj_bias = self.attn.in_proj_bias[2*embed_dims:]
finetuning_detach(self)
def lora_attn_forward(
self,
query,
key,
value,
attn_mask,
key_padding_mask,
task_mask=None,
task_idx=None,
):
if self.attn.batch_first:
query, key, value = [x.transpose(1, 0) for x in (query, key, value)]
lora_query = self.q_lora(query, task_mask, task_idx=task_idx)
lora_key = self.k_lora(key, task_mask, task_idx=task_idx)
lora_value = self.v_lora(value, task_mask, task_idx=task_idx)
# print(query.device, self.q_proj_weight.device, self.q_proj_bias.device)
self.q_proj_weight = self.q_proj_weight.cuda()
self.q_proj_bias = self.q_proj_bias.cuda()
self.k_proj_weight = self.k_proj_weight.cuda()
self.k_proj_bias = self.k_proj_bias.cuda()
self.v_proj_weight = self.v_proj_weight.cuda()
self.v_proj_bias = self.v_proj_bias.cuda()
new_query = F.linear(query, self.q_proj_weight, self.q_proj_bias)
new_key = F.linear(key, self.k_proj_weight, self.k_proj_bias)
new_value = F.linear(value, self.v_proj_weight, self.v_proj_bias)
query = new_query.detach() + lora_query
key = new_key.detach() + lora_key
value = new_value.detach() + lora_value
attn_output, attn_output_weights = custom_multi_head_attention_forward(
query, key, value, self.attn.embed_dim, self.attn.num_heads,
in_proj_weight=None, # Projections are already done manually
in_proj_bias=None, # Bias is not used
bias_k=None, bias_v=None, add_zero_attn=self.attn.add_zero_attn,
dropout_p=self.attn.dropout,
out_proj_weight=self.attn.out_proj.weight, out_proj_bias=self.attn.out_proj.bias,
training=self.attn.training,
key_padding_mask=key_padding_mask, need_weights=True,
use_separate_proj_weight=True,
use_direct_input=True,
attn_mask=attn_mask,
)
if self.attn.batch_first:
return attn_output.transpose(1, 0), attn_output_weights
else:
return attn_output, attn_output_weights
@deprecated_api_warning({'residual': 'identity'},
cls_name='MultiheadAttention')
def forward(self,
query,
key=None,
value=None,
identity=None,
query_pos=None,
key_pos=None,
attn_mask=None,
key_padding_mask=None,
task_mask=None,
forward_origin=False,
task_idx=None,
**kwargs):
"""Forward function for `MultiheadAttention`.
**kwargs allow passing a more general data flow when combining
with other operations in `transformerlayer`.
Args:
query (Tensor): The input query with shape [num_queries, bs,
embed_dims] if self.batch_first is False, else
[bs, num_queries embed_dims].
key (Tensor): The key tensor with shape [num_keys, bs,
embed_dims] if self.batch_first is False, else
[bs, num_keys, embed_dims] .
If None, the ``query`` will be used. Defaults to None.
value (Tensor): The value tensor with same shape as `key`.
Same in `nn.MultiheadAttention.forward`. Defaults to None.
If None, the `key` will be used.
identity (Tensor): This tensor, with the same shape as x,
will be used for the identity link.
If None, `x` will be used. Defaults to None.
query_pos (Tensor): The positional encoding for query, with
the same shape as `x`. If not None, it will
be added to `x` before forward function. Defaults to None.
key_pos (Tensor): The positional encoding for `key`, with the
same shape as `key`. Defaults to None. If not None, it will
be added to `key` before forward function. If None, and
`query_pos` has the same shape as `key`, then `query_pos`
will be used for `key_pos`. Defaults to None.
attn_mask (Tensor): ByteTensor mask with shape [num_queries,
num_keys]. Same in `nn.MultiheadAttention.forward`.
Defaults to None.
key_padding_mask (Tensor): ByteTensor with shape [bs, num_keys].
Defaults to None.
Returns:
Tensor: forwarded results with shape
[num_queries, bs, embed_dims]
if self.batch_first is False, else
[bs, num_queries embed_dims].
"""
if key is None:
key = query
if value is None:
value = key
if identity is None:
identity = query
if key_pos is None:
if query_pos is not None:
# use query_pos if key_pos is not available
if query_pos.shape == key.shape:
key_pos = query_pos
else:
warnings.warn(f'position encoding of key is'
f'missing in {self.__class__.__name__}.')
if query_pos is not None:
query = query + query_pos
if key_pos is not None:
key = key + key_pos
# Because the dataflow('key', 'query', 'value') of
# ``torch.nn.MultiheadAttention`` is (num_query, batch,
# embed_dims), We should adjust the shape of dataflow from
# batch_first (batch, num_query, embed_dims) to num_query_first
# (num_query ,batch, embed_dims), and recover ``attn_output``
# from num_query_first to batch_first.
if self.batch_first:
query = query.transpose(0, 1)
key = key.transpose(0, 1)
value = value.transpose(0, 1)
if self.with_cp:
out = cp.checkpoint(self.attn, use_reentrant=False, query=query,
key=key,
value=value,
attn_mask=attn_mask,
key_padding_mask=key_padding_mask)[0]
else:
if self.use_lora and forward_origin==False:
out = self.lora_attn_forward(
query=query,
key=key,
value=value,
attn_mask=attn_mask,
key_padding_mask=key_padding_mask,
task_mask=task_mask,
task_idx=task_idx)[0]
else:
out = self.attn(
query=query,
key=key,
value=value,
attn_mask=attn_mask,
key_padding_mask=key_padding_mask)[0]
if self.batch_first:
out = out.transpose(0, 1)
if self.use_lora:
out = out + self.out_lora(out, i=task_idx)
return identity + self.dropout_layer(self.proj_drop(out))
#forceful register
@FEEDFORWARD_NETWORK.register_module(force=True)
class FFN(BaseModule):
"""Implements feed-forward networks (FFNs) with identity connection.
Args:
embed_dims (int): The feature dimension. Same as
`MultiheadAttention`. Defaults: 256.
feedforward_channels (int): The hidden dimension of FFNs.
Defaults: 1024.
num_fcs (int, optional): The number of fully-connected layers in
FFNs. Default: 2.
act_cfg (dict, optional): The activation config for FFNs.
Default: dict(type='ReLU')
ffn_drop (float, optional): Probability of an element to be
zeroed in FFN. Default 0.0.
add_identity (bool, optional): Whether to add the
identity connection. Default: `True`.
dropout_layer (obj:`ConfigDict`): The dropout_layer used
when adding the shortcut.
init_cfg (obj:`mmcv.ConfigDict`): The Config for initialization.
Default: None.
"""
@deprecated_api_warning(
{
'dropout': 'ffn_drop',
'add_residual': 'add_identity'
},
cls_name='FFN')
def __init__(self,
embed_dims=256,
feedforward_channels=1024,
num_fcs=2,
act_cfg=dict(type='ReLU', inplace=True),
ffn_drop=0.,
dropout_layer=None,
add_identity=True,
init_cfg=None,
use_lora=False,
lora_rank=16,
use_adapter=False,
adatper_num=6,
lora_moe=False,
num_task=6,
**kwargs):
super(FFN, self).__init__(init_cfg)
assert num_fcs >= 2, 'num_fcs should be no less ' \
f'than 2. got {num_fcs}.'
self.embed_dims = embed_dims
self.feedforward_channels = feedforward_channels
self.num_fcs = num_fcs
self.act_cfg = act_cfg
self.activate = build_activation_layer(act_cfg)
self.lora_moe = lora_moe
layers = []
in_channels = embed_dims
for _ in range(num_fcs - 1):
layers.append(
Sequential(
Linear(in_channels, feedforward_channels), self.activate,
nn.Dropout(ffn_drop)))
in_channels = feedforward_channels
layers.append(Linear(feedforward_channels, embed_dims))
layers.append(nn.Dropout(ffn_drop))
self.use_lora = use_lora
self.use_adapter = use_adapter
if use_adapter:
assert use_lora==True
self.layers = Sequential(*layers)
if self.use_lora:
if self.use_adapter:
self.lora_layers = LoRAMoECLAdapter(embed_dims, feedforward_channels, embed_dims,
num_task=adatper_num, r=lora_rank, dropout=ffn_drop)
else:
lora_layer = MOELoRALinear if self.lora_moe else LoRALinear
self.lora_layers = lora_wrapper(self.layers, LoraLayer=lora_layer, rank=lora_rank, dropout=0.0,num_task=num_task)
self.dropout_layer = build_dropout(
dropout_layer) if dropout_layer else torch.nn.Identity()
self.add_identity = add_identity
if self.use_lora:
self.layers = frozen_grad(self.layers)
finetuning_detach(self)
@deprecated_api_warning({'residual': 'identity'}, cls_name='FFN')
def forward(self, x, identity=None, forward_origin=False, task_idx=None,**kwargs):
"""Forward function for `FFN`.
The function would add x to the output tensor if residue is None.
"""
org_x = x.clone()
if (not self.use_lora) or forward_origin:
out = self.layers(x)
else:
if self.use_adapter:
out = self.layers(x).detach() + self.lora_layers(x)
else:
out = peft_wrapper_forward(x, self.layers, self.lora_layers, task_idx=task_idx)
if not self.add_identity:
return self.dropout_layer(out)
if identity is None:
identity = org_x
return identity + self.dropout_layer(out)
@TRANSFORMER_LAYER.register_module(force=True)
class BaseTransformerLayer(BaseModule):
"""Base `TransformerLayer` for vision transformer.
It can be built from `mmcv.ConfigDict` and support more flexible
customization, for example, using any number of `FFN or LN ` and
use different kinds of `attention` by specifying a list of `ConfigDict`
named `attn_cfgs`. It is worth mentioning that it supports `prenorm`
when you specifying `norm` as the first element of `operation_order`.
More details about the `prenorm`: `On Layer Normalization in the
Transformer Architecture <https://arxiv.org/abs/2002.04745>`_ .
Args:
attn_cfgs (list[`mmcv.ConfigDict`] | obj:`mmcv.ConfigDict` | None )):
Configs for `self_attention` or `cross_attention` modules,
The order of the configs in the list should be consistent with
corresponding attentions in operation_order.
If it is a dict, all of the attention modules in operation_order
will be built with this config. Default: None.
ffn_cfgs (list[`mmcv.ConfigDict`] | obj:`mmcv.ConfigDict` | None )):
Configs for FFN, The order of the configs in the list should be
consistent with corresponding ffn in operation_order.
If it is a dict, all of the attention modules in operation_order
will be built with this config.
operation_order (tuple[str]): The execution order of operation
in transformer. Such as ('self_attn', 'norm', 'ffn', 'norm').
Support `prenorm` when you specifying first element as `norm`.
Default：None.
norm_cfg (dict): Config dict for normalization layer.
Default: dict(type='LN').
init_cfg (obj:`mmcv.ConfigDict`): The Config for initialization.
Default: None.
batch_first (bool): Key, Query and Value are shape
of (batch, n, embed_dim)
or (n, batch, embed_dim). Default to False.
"""
def __init__(self,
attn_cfgs=None,
with_cp=False,
ffn_cfgs=dict(
type='FFN',
embed_dims=256,
feedforward_channels=1024,
num_fcs=2,
ffn_drop=0.,
act_cfg=dict(type='ReLU', inplace=True),
),
operation_order=None,
norm_cfg=dict(type='LN'),
init_cfg=None,
batch_first=False,
use_lora=False,
lora_rank=16,
ffn_use_lora=False,
ffn_lora_rank=16,
ffn_use_adapter=False,
ffn_adapter_num=6,
moe_lora=False,
num_task=6,
**kwargs):
deprecated_args = dict(
feedforward_channels='feedforward_channels',
ffn_dropout='ffn_drop',
ffn_num_fcs='num_fcs')
for ori_name, new_name in deprecated_args.items():
if ori_name in kwargs:
warnings.warn(
f'The arguments `{ori_name}` in BaseTransformerLayer '
f'has been deprecated, now you should set `{new_name}` '
f'and other FFN related arguments '
f'to a dict named `ffn_cfgs`. ')
ffn_cfgs[new_name] = kwargs[ori_name]
super(BaseTransformerLayer, self).__init__(init_cfg)
self.batch_first = batch_first
assert set(operation_order) & set(
['self_attn', 'norm', 'ffn', 'cross_attn']) == \
set(operation_order), f'The operation_order of' \
f' {self.__class__.__name__} should ' \
f'contains all four operation type ' \
f"{['self_attn', 'norm', 'ffn', 'cross_attn']}"
num_attn = operation_order.count('self_attn') + operation_order.count(
'cross_attn')
attn_use_lora = False
if isinstance(attn_cfgs, dict):
if 'use_lora' in attn_cfgs:
attn_use_lora = attn_cfgs['use_lora']
else:
attn_use_lora = False
if isinstance(attn_cfgs, dict):
attn_cfgs = [copy.deepcopy(attn_cfgs) for _ in range(num_attn)]
else:
assert num_attn == len(attn_cfgs), f'The length ' \
f'of attn_cfg {num_attn} is ' \
f'not consistent with the number of attention' \
f'in operation_order {operation_order}.'
self.num_attn = num_attn
self.operation_order = operation_order
self.norm_cfg = norm_cfg
self.pre_norm = operation_order[0] == 'norm'
self.attentions = ModuleList()
self.use_lora = use_lora
index = 0
have_use_lora = False
for attn_cfg in attn_cfgs:
if 'use_lora' in attn_cfg:
if attn_cfg['use_lora'] == True:
have_use_lora = True
for operation_name in operation_order:
if operation_name in ['self_attn', 'cross_attn']:
if 'batch_first' in attn_cfgs[index]:
assert self.batch_first == attn_cfgs[index]['batch_first']
else:
attn_cfgs[index]['batch_first'] = self.batch_first
attention = build_attention(attn_cfgs[index])
if have_use_lora:
if 'use_lora' not in attn_cfgs[index].keys():
for param in attention.parameters():
param.requires_grad = False
else:
if attn_cfgs[index]['use_lora']==False:
for param in attention.parameters():
param.requires_grad = False
# Some custom attentions used as `self_attn`
# or `cross_attn` can have different behavior.
attention.operation_name = operation_name
self.attentions.append(attention)
index += 1
self.embed_dims = self.attentions[0].embed_dims
self.ffns = ModuleList()
ffn_cfgs['use_lora'] = ffn_use_lora
ffn_cfgs['lora_rank'] = ffn_lora_rank
ffn_cfgs['use_adapter'] = ffn_use_adapter
ffn_cfgs['adapter_num'] = ffn_adapter_num
ffn_cfgs['moe_lora'] = moe_lora
ffn_cfgs['num_task'] = num_task
self.freeze_ffn = False
if ffn_use_lora==False:
self.freeze_ffn = True
num_ffns = operation_order.count('ffn')
if isinstance(ffn_cfgs, dict):
ffn_cfgs = ConfigDict(ffn_cfgs)
if isinstance(ffn_cfgs, dict):
ffn_cfgs = [copy.deepcopy(ffn_cfgs) for _ in range(num_ffns)]
assert len(ffn_cfgs) == num_ffns
for ffn_index in range(num_ffns):
if 'embed_dims' not in ffn_cfgs[ffn_index]:
ffn_cfgs['embed_dims'] = self.embed_dims
else:
assert ffn_cfgs[ffn_index]['embed_dims'] == self.embed_dims
self.ffns.append(
build_feedforward_network(ffn_cfgs[ffn_index],
dict(type='FFN')))
self.norms = ModuleList()
num_norms = operation_order.count('norm')
for _ in range(num_norms):
self.norms.append(build_norm_layer(norm_cfg, self.embed_dims)[1])
self.with_cp = with_cp
if self.freeze_ffn:
for param in self.ffns.parameters():
param.requires_grad = False
def forward(self,
query,
key=None,
value=None,
query_pos=None,
key_pos=None,
attn_masks=None,
query_key_padding_mask=None,
key_padding_mask=None,
task_mask=None,
forward_origin=False,
task_idx=None,
**kwargs):
"""Forward function for `TransformerDecoderLayer`.
**kwargs contains some specific arguments of attentions.
Args:
query (Tensor): The input query with shape
[num_queries, bs, embed_dims] if
self.batch_first is False, else
[bs, num_queries embed_dims].
key (Tensor): The key tensor with shape [num_keys, bs,
embed_dims] if self.batch_first is False, else
[bs, num_keys, embed_dims] .
value (Tensor): The value tensor with same shape as `key`.
query_pos (Tensor): The positional encoding for `query`.
Default: None.
key_pos (Tensor): The positional encoding for `key`.
Default: None.
attn_masks (List[Tensor] | None): 2D Tensor used in
calculation of corresponding attention. The length of
it should equal to the number of `attention` in
`operation_order`. Default: None.
query_key_padding_mask (Tensor): ByteTensor for `query`, with
shape [bs, num_queries]. Only used in `self_attn` layer.
Defaults to None.
key_padding_mask (Tensor): ByteTensor for `query`, with
shape [bs, num_keys]. Default: None.
Returns:
Tensor: forwarded results with shape [num_queries, bs, embed_dims].
"""
norm_index = 0
attn_index = 0
ffn_index = 0
identity = query
if attn_masks is None:
attn_masks = [None for _ in range(self.num_attn)]
elif isinstance(attn_masks, torch.Tensor):
attn_masks = [
copy.deepcopy(attn_masks) for _ in range(self.num_attn)
]
warnings.warn(f'Use same attn_mask in all attentions in '
f'{self.__class__.__name__} ')
else:
assert len(attn_masks) == self.num_attn, f'The length of ' \
f'attn_masks {len(attn_masks)} must be equal ' \
f'to the number of attention in ' \
f'operation_order {self.num_attn}'
for layer in self.operation_order:
if layer == 'self_attn':
temp_key = temp_value = query
query = self.attentions[attn_index](
query,
temp_key,
temp_value,
identity if self.pre_norm else None,
query_pos=query_pos,
key_pos=query_pos,
attn_mask=attn_masks[attn_index],
key_padding_mask=query_key_padding_mask,
forward_origin=forward_origin,
task_idx=task_idx,
**kwargs)
attn_index += 1
identity = query
elif layer == 'norm':
query = self.norms[norm_index](query)
norm_index += 1
elif layer == 'cross_attn':
query = self.attentions[attn_index](
query,
key,
value,
identity if self.pre_norm else None,
query_pos=query_pos,
key_pos=key_pos,
attn_mask=attn_masks[attn_index],
key_padding_mask=key_padding_mask,
task_mask=task_mask,
forward_origin=forward_origin,
task_idx=task_idx,
**kwargs)
attn_index += 1
identity = query
elif layer == 'ffn':
if self.with_cp:
query = cp.checkpoint(self.ffns[ffn_index], query)
else:
query = self.ffns[ffn_index](
query, identity if self.pre_norm else None,
forward_origin=forward_origin,
task_idx=task_idx,)
ffn_index += 1
return query
@TRANSFORMER_LAYER_SEQUENCE.register_module(force=True)
class TransformerLayerSequence(BaseModule):
"""Base class for TransformerEncoder and TransformerDecoder in vision
transformer.
As base-class of Encoder and Decoder in vision transformer.
Support customization such as specifying different kind
of `transformer_layer` in `transformer_coder`.
Args:
transformerlayer (list[obj:`mmcv.ConfigDict`] |
obj:`mmcv.ConfigDict`): Config of transformerlayer
in TransformerCoder. If it is obj:`mmcv.ConfigDict`,
it would be repeated `num_layer` times to a
list[`mmcv.ConfigDict`]. Default: None.
num_layers (int): The number of `TransformerLayer`. Default: None.
init_cfg (obj:`mmcv.ConfigDict`): The Config for initialization.
Default: None.
"""
def __init__(self, transformerlayers=None, num_layers=None, init_cfg=None):
super(TransformerLayerSequence, self).__init__(init_cfg)
if isinstance(transformerlayers, dict):
transformerlayers = [
copy.deepcopy(transformerlayers) for _ in range(num_layers)
]
else:
assert isinstance(transformerlayers, list) and \
len(transformerlayers) == num_layers
self.num_layers = num_layers
self.layers = ModuleList()
for i in range(num_layers):
self.layers.append(build_transformer_layer(transformerlayers[i]))
self.embed_dims = self.layers[0].embed_dims
self.pre_norm = self.layers[0].pre_norm
def forward(self,
query,
key,
value,
query_pos=None,
key_pos=None,
attn_masks=None,
query_key_padding_mask=None,
key_padding_mask=None,
task_mask=None,
forward_origin=False,
task_idx=None,
**kwargs):
"""Forward function for `TransformerCoder`.
Args:
query (Tensor): Input query with shape
`(num_queries, bs, embed_dims)`.
key (Tensor): The key tensor with shape
`(num_keys, bs, embed_dims)`.
value (Tensor): The value tensor with shape
`(num_keys, bs, embed_dims)`.
query_pos (Tensor): The positional encoding for `query`.
Default: None.
key_pos (Tensor): The positional encoding for `key`.
Default: None.
attn_masks (List[Tensor], optional): Each element is 2D Tensor
which is used in calculation of corresponding attention in
operation_order. Default: None.
query_key_padding_mask (Tensor): ByteTensor for `query`, with
shape [bs, num_queries]. Only used in self-attention
Default: None.
key_padding_mask (Tensor): ByteTensor for `query`, with
shape [bs, num_keys]. Default: None.
Returns:
Tensor: results with shape [num_queries, bs, embed_dims].
"""
for layer in self.layers:
query = layer(
query,
key,
value,
query_pos=query_pos,
key_pos=key_pos,
attn_masks=attn_masks,
query_key_padding_mask=query_key_padding_mask,
key_padding_mask=key_padding_mask,
task_mask=task_mask,
forward_origin=forward_origin,
task_idx=task_idx,
**kwargs)
return query
def transformer_test():
model = MultiheadAttention(
64, 4, use_lora=True
)
model = FFN(use_lora=True)
# finetuning_detach(model)
print("Model structure after attaching LoRA layers:\n", model)
for name, param in model.named_parameters():
print(name, param.shape, param.requires_grad)
if __name__=='__main__':
transformer_test()
|