distillation_loss.py 15.8 KB
Newer Older
littletomatodonkey's avatar
littletomatodonkey 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
#copyright (c) 2021 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
#    http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.

import paddle
import paddle.nn as nn
L
fix bug  
LDOUBLEV 已提交
17 18
import numpy as np
import cv2
littletomatodonkey's avatar
littletomatodonkey 已提交
19 20

from .rec_ctc_loss import CTCLoss
A
andyjpaddle 已提交
21
from .rec_sar_loss import SARLoss
littletomatodonkey's avatar
littletomatodonkey 已提交
22
from .basic_loss import DMLLoss
23
from .basic_loss import DistanceLoss
littletomatodonkey's avatar
littletomatodonkey 已提交
24
from .basic_loss import LossFromOutput
L
LDOUBLEV 已提交
25 26
from .det_db_loss import DBLoss
from .det_basic_loss import BalanceLoss, MaskL1Loss, DiceLoss
littletomatodonkey's avatar
littletomatodonkey 已提交
27
from .vqa_token_layoutlm_loss import VQASerTokenLayoutLMLoss
littletomatodonkey's avatar
littletomatodonkey 已提交
28 29


L
LDOUBLEV 已提交
30 31 32 33 34 35 36 37 38 39 40 41
def _sum_loss(loss_dict):
    if "loss" in loss_dict.keys():
        return loss_dict
    else:
        loss_dict["loss"] = 0.
        for k, value in loss_dict.items():
            if k == "loss":
                continue
            else:
                loss_dict["loss"] += value
        return loss_dict

L
LDOUBLEV 已提交
42 43

class DistillationDMLLoss(DMLLoss):
littletomatodonkey's avatar
littletomatodonkey 已提交
44 45 46
    """
    """

L
LDOUBLEV 已提交
47 48 49
    def __init__(self,
                 model_name_pairs=[],
                 act=None,
50
                 use_log=False,
L
LDOUBLEV 已提交
51
                 key=None,
A
andyjpaddle 已提交
52 53
                 multi_head=False,
                 dis_head='ctc',
L
LDOUBLEV 已提交
54
                 maps_name=None,
L
LDOUBLEV 已提交
55
                 name="dml"):
56
        super().__init__(act=act, use_log=use_log)
57
        assert isinstance(model_name_pairs, list)
littletomatodonkey's avatar
littletomatodonkey 已提交
58
        self.key = key
A
andyjpaddle 已提交
59 60
        self.multi_head = multi_head
        self.dis_head = dis_head
L
fix bug  
LDOUBLEV 已提交
61
        self.model_name_pairs = self._check_model_name_pairs(model_name_pairs)
littletomatodonkey's avatar
littletomatodonkey 已提交
62
        self.name = name
L
LDOUBLEV 已提交
63
        self.maps_name = self._check_maps_name(maps_name)
64

L
fix bug  
LDOUBLEV 已提交
65 66 67
    def _check_model_name_pairs(self, model_name_pairs):
        if not isinstance(model_name_pairs, list):
            return []
68 69
        elif isinstance(model_name_pairs[0], list) and isinstance(
                model_name_pairs[0][0], str):
L
fix bug  
LDOUBLEV 已提交
70 71 72
            return model_name_pairs
        else:
            return [model_name_pairs]
L
LDOUBLEV 已提交
73 74 75 76 77 78 79 80 81 82 83 84 85 86 87

    def _check_maps_name(self, maps_name):
        if maps_name is None:
            return None
        elif type(maps_name) == str:
            return [maps_name]
        elif type(maps_name) == list:
            return [maps_name]
        else:
            return None

    def _slice_out(self, outs):
        new_outs = {}
        for k in self.maps_name:
            if k == "thrink_maps":
L
LDOUBLEV 已提交
88
                new_outs[k] = outs[:, 0, :, :]
L
LDOUBLEV 已提交
89
            elif k == "threshold_maps":
L
LDOUBLEV 已提交
90
                new_outs[k] = outs[:, 1, :, :]
L
LDOUBLEV 已提交
91
            elif k == "binary_maps":
L
LDOUBLEV 已提交
92
                new_outs[k] = outs[:, 2, :, :]
L
LDOUBLEV 已提交
93 94
            else:
                continue
L
fix bug  
LDOUBLEV 已提交
95
        return new_outs
littletomatodonkey's avatar
littletomatodonkey 已提交
96 97 98

    def forward(self, predicts, batch):
        loss_dict = dict()
99 100 101
        for idx, pair in enumerate(self.model_name_pairs):
            out1 = predicts[pair[0]]
            out2 = predicts[pair[1]]
littletomatodonkey's avatar
littletomatodonkey 已提交
102 103 104
            if self.key is not None:
                out1 = out1[self.key]
                out2 = out2[self.key]
L
LDOUBLEV 已提交
105 106

            if self.maps_name is None:
A
andyjpaddle 已提交
107 108 109 110 111
                if self.multi_head:
                    loss = super().forward(out1[self.dis_head],
                                           out2[self.dis_head])
                else:
                    loss = super().forward(out1, out2)
L
LDOUBLEV 已提交
112 113 114 115 116 117
                if isinstance(loss, dict):
                    for key in loss:
                        loss_dict["{}_{}_{}_{}".format(key, pair[0], pair[1],
                                                       idx)] = loss[key]
                else:
                    loss_dict["{}_{}".format(self.name, idx)] = loss
118
            else:
L
LDOUBLEV 已提交
119 120
                outs1 = self._slice_out(out1)
                outs2 = self._slice_out(out2)
L
LDOUBLEV 已提交
121
                for _c, k in enumerate(outs1.keys()):
L
LDOUBLEV 已提交
122 123 124 125
                    loss = super().forward(outs1[k], outs2[k])
                    if isinstance(loss, dict):
                        for key in loss:
                            loss_dict["{}_{}_{}_{}_{}".format(key, pair[
littletomatodonkey's avatar
littletomatodonkey 已提交
126
                                0], pair[1], self.maps_name, idx)] = loss[key]
L
LDOUBLEV 已提交
127
                    else:
128 129 130
                        loss_dict["{}_{}_{}".format(self.name, self.maps_name[
                            _c], idx)] = loss

L
LDOUBLEV 已提交
131 132
        loss_dict = _sum_loss(loss_dict)

littletomatodonkey's avatar
littletomatodonkey 已提交
133 134 135 136
        return loss_dict


class DistillationCTCLoss(CTCLoss):
A
andyjpaddle 已提交
137 138 139 140 141
    def __init__(self,
                 model_name_list=[],
                 key=None,
                 multi_head=False,
                 name="loss_ctc"):
littletomatodonkey's avatar
littletomatodonkey 已提交
142 143 144 145
        super().__init__()
        self.model_name_list = model_name_list
        self.key = key
        self.name = name
A
andyjpaddle 已提交
146
        self.multi_head = multi_head
littletomatodonkey's avatar
littletomatodonkey 已提交
147 148 149

    def forward(self, predicts, batch):
        loss_dict = dict()
150
        for idx, model_name in enumerate(self.model_name_list):
littletomatodonkey's avatar
littletomatodonkey 已提交
151 152 153
            out = predicts[model_name]
            if self.key is not None:
                out = out[self.key]
A
andyjpaddle 已提交
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
            if self.multi_head:
                assert 'ctc' in out, 'multi head has multi out'
                loss = super().forward(out['ctc'], batch[:2] + batch[3:])
            else:
                loss = super().forward(out, batch)
            if isinstance(loss, dict):
                for key in loss:
                    loss_dict["{}_{}_{}".format(self.name, model_name,
                                                idx)] = loss[key]
            else:
                loss_dict["{}_{}".format(self.name, model_name)] = loss
        return loss_dict


class DistillationSARLoss(SARLoss):
    def __init__(self,
                 model_name_list=[],
                 key=None,
                 multi_head=False,
                 name="loss_sar",
                 **kwargs):
        ignore_index = kwargs.get('ignore_index', 92)
        super().__init__(ignore_index=ignore_index)
        self.model_name_list = model_name_list
        self.key = key
        self.name = name
        self.multi_head = multi_head

    def forward(self, predicts, batch):
        loss_dict = dict()
        for idx, model_name in enumerate(self.model_name_list):
            out = predicts[model_name]
            if self.key is not None:
                out = out[self.key]
            if self.multi_head:
                assert 'sar' in out, 'multi head has multi out'
                loss = super().forward(out['sar'], batch[:1] + batch[2:])
            else:
                loss = super().forward(out, batch)
littletomatodonkey's avatar
littletomatodonkey 已提交
193
            if isinstance(loss, dict):
194 195 196 197 198
                for key in loss:
                    loss_dict["{}_{}_{}".format(self.name, model_name,
                                                idx)] = loss[key]
            else:
                loss_dict["{}_{}".format(self.name, model_name)] = loss
littletomatodonkey's avatar
littletomatodonkey 已提交
199
        return loss_dict
200 201


L
LDOUBLEV 已提交
202 203 204 205 206 207 208 209 210
class DistillationDBLoss(DBLoss):
    def __init__(self,
                 model_name_list=[],
                 balance_loss=True,
                 main_loss_type='DiceLoss',
                 alpha=5,
                 beta=10,
                 ohem_ratio=3,
                 eps=1e-6,
L
LDOUBLEV 已提交
211
                 name="db",
L
LDOUBLEV 已提交
212 213 214 215 216 217
                 **kwargs):
        super().__init__()
        self.model_name_list = model_name_list
        self.name = name
        self.key = None

L
fix bug  
LDOUBLEV 已提交
218
    def forward(self, predicts, batch):
L
LDOUBLEV 已提交
219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241
        loss_dict = {}
        for idx, model_name in enumerate(self.model_name_list):
            out = predicts[model_name]
            if self.key is not None:
                out = out[self.key]
            loss = super().forward(out, batch)

            if isinstance(loss, dict):
                for key in loss.keys():
                    if key == "loss":
                        continue
                    name = "{}_{}_{}".format(self.name, model_name, key)
                    loss_dict[name] = loss[key]
            else:
                loss_dict["{}_{}".format(self.name, model_name)] = loss

        loss_dict = _sum_loss(loss_dict)
        return loss_dict


class DistillationDilaDBLoss(DBLoss):
    def __init__(self,
                 model_name_pairs=[],
L
fix bug  
LDOUBLEV 已提交
242
                 key=None,
L
LDOUBLEV 已提交
243 244 245 246 247 248 249 250 251 252
                 balance_loss=True,
                 main_loss_type='DiceLoss',
                 alpha=5,
                 beta=10,
                 ohem_ratio=3,
                 eps=1e-6,
                 name="dila_dbloss"):
        super().__init__()
        self.model_name_pairs = model_name_pairs
        self.name = name
L
fix bug  
LDOUBLEV 已提交
253
        self.key = key
L
LDOUBLEV 已提交
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

    def forward(self, predicts, batch):
        loss_dict = dict()
        for idx, pair in enumerate(self.model_name_pairs):
            stu_outs = predicts[pair[0]]
            tch_outs = predicts[pair[1]]
            if self.key is not None:
                stu_preds = stu_outs[self.key]
                tch_preds = tch_outs[self.key]

            stu_shrink_maps = stu_preds[:, 0, :, :]
            stu_binary_maps = stu_preds[:, 2, :, :]

            # dilation to teacher prediction
            dilation_w = np.array([[1, 1], [1, 1]])
            th_shrink_maps = tch_preds[:, 0, :, :]
            th_shrink_maps = th_shrink_maps.numpy() > 0.3  # thresh = 0.3 
            dilate_maps = np.zeros_like(th_shrink_maps).astype(np.float32)
            for i in range(th_shrink_maps.shape[0]):
                dilate_maps[i] = cv2.dilate(
                    th_shrink_maps[i, :, :].astype(np.uint8), dilation_w)
            th_shrink_maps = paddle.to_tensor(dilate_maps)

            label_threshold_map, label_threshold_mask, label_shrink_map, label_shrink_mask = batch[
                1:]

            # calculate the shrink map loss
            bce_loss = self.alpha * self.bce_loss(
                stu_shrink_maps, th_shrink_maps, label_shrink_mask)
            loss_binary_maps = self.dice_loss(stu_binary_maps, th_shrink_maps,
                                              label_shrink_mask)

            # k = f"{self.name}_{pair[0]}_{pair[1]}"
            k = "{}_{}_{}".format(self.name, pair[0], pair[1])
            loss_dict[k] = bce_loss + loss_binary_maps

        loss_dict = _sum_loss(loss_dict)
L
fix bug  
LDOUBLEV 已提交
291
        return loss_dict
L
LDOUBLEV 已提交
292 293


294 295 296 297 298 299 300 301 302 303
class DistillationDistanceLoss(DistanceLoss):
    """
    """

    def __init__(self,
                 mode="l2",
                 model_name_pairs=[],
                 key=None,
                 name="loss_distance",
                 **kargs):
littletomatodonkey's avatar
littletomatodonkey 已提交
304
        super().__init__(mode=mode, **kargs)
305 306 307
        assert isinstance(model_name_pairs, list)
        self.key = key
        self.model_name_pairs = model_name_pairs
littletomatodonkey's avatar
littletomatodonkey 已提交
308
        self.name = name + "_l2"
309 310 311 312 313 314 315 316 317 318 319 320 321 322 323

    def forward(self, predicts, batch):
        loss_dict = dict()
        for idx, pair in enumerate(self.model_name_pairs):
            out1 = predicts[pair[0]]
            out2 = predicts[pair[1]]
            if self.key is not None:
                out1 = out1[self.key]
                out2 = out2[self.key]
            loss = super().forward(out1, out2)
            if isinstance(loss, dict):
                for key in loss:
                    loss_dict["{}_{}_{}".format(self.name, key, idx)] = loss[
                        key]
            else:
littletomatodonkey's avatar
littletomatodonkey 已提交
324 325
                loss_dict["{}_{}_{}_{}".format(self.name, pair[0], pair[1],
                                               idx)] = loss
326
        return loss_dict
littletomatodonkey's avatar
littletomatodonkey 已提交
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


class DistillationVQASerTokenLayoutLMLoss(VQASerTokenLayoutLMLoss):
    def __init__(self,
                 num_classes,
                 model_name_list=[],
                 key=None,
                 name="loss_ser"):
        super().__init__(num_classes=num_classes)
        self.model_name_list = model_name_list
        self.key = key
        self.name = name

    def forward(self, predicts, batch):
        loss_dict = dict()
        for idx, model_name in enumerate(self.model_name_list):
            out = predicts[model_name]
            if self.key is not None:
                out = out[self.key]
            loss = super().forward(out, batch)
            loss_dict["{}_{}".format(self.name, model_name)] = loss["loss"]
        return loss_dict


class DistillationLossFromOutput(LossFromOutput):
    def __init__(self,
                 reduction="none",
                 model_name_list=[],
                 dist_key=None,
                 key="loss",
                 name="loss_re"):
        super().__init__(key=key, reduction=reduction)
        self.model_name_list = model_name_list
        self.name = name
        self.dist_key = dist_key

    def forward(self, predicts, batch):
        loss_dict = dict()
        for idx, model_name in enumerate(self.model_name_list):
            out = predicts[model_name]
            if self.dist_key is not None:
                out = out[self.dist_key]
            loss = super().forward(out, batch)
            loss_dict["{}_{}".format(self.name, model_name)] = loss["loss"]
        return loss_dict


class DistillationSERDMLLoss(DMLLoss):
    """
    """

    def __init__(self,
                 act="softmax",
                 use_log=True,
                 num_classes=7,
                 model_name_pairs=[],
                 key=None,
                 name="loss_dml_ser"):
        super().__init__(act=act, use_log=use_log)
        assert isinstance(model_name_pairs, list)
        self.key = key
        self.name = name
        self.num_classes = num_classes
        self.model_name_pairs = model_name_pairs

    def forward(self, predicts, batch):
        loss_dict = dict()
        for idx, pair in enumerate(self.model_name_pairs):
            out1 = predicts[pair[0]]
            out2 = predicts[pair[1]]
            if self.key is not None:
                out1 = out1[self.key]
                out2 = out2[self.key]
            out1 = out1.reshape([-1, out1.shape[-1]])
            out2 = out2.reshape([-1, out2.shape[-1]])

            attention_mask = batch[2]
            if attention_mask is not None:
                active_output = attention_mask.reshape([-1, ]) == 1
                out1 = out1[active_output]
                out2 = out2[active_output]

            loss_dict["{}_{}".format(self.name, idx)] = super().forward(out1,
                                                                        out2)

        return loss_dict


class DistillationVQADistanceLoss(DistanceLoss):
    def __init__(self,
                 mode="l2",
                 model_name_pairs=[],
                 key=None,
                 name="loss_distance",
                 **kargs):
        super().__init__(mode=mode, **kargs)
        assert isinstance(model_name_pairs, list)
        self.key = key
        self.model_name_pairs = model_name_pairs
        self.name = name + "_l2"

    def forward(self, predicts, batch):
        loss_dict = dict()
        for idx, pair in enumerate(self.model_name_pairs):
            out1 = predicts[pair[0]]
            out2 = predicts[pair[1]]
            attention_mask = batch[2]
            if self.key is not None:
                out1 = out1[self.key]
                out2 = out2[self.key]
                if attention_mask is not None:
                    max_len = attention_mask.shape[-1]
                    out1 = out1[:, :max_len]
                    out2 = out2[:, :max_len]
                out1 = out1.reshape([-1, out1.shape[-1]])
                out2 = out2.reshape([-1, out2.shape[-1]])
            if attention_mask is not None:
                active_output = attention_mask.reshape([-1, ]) == 1
                out1 = out1[active_output]
                out2 = out2[active_output]

            loss = super().forward(out1, out2)
            if isinstance(loss, dict):
                for key in loss:
                    loss_dict["{}_{}nohu_{}".format(self.name, key,
                                                    idx)] = loss[key]
            else:
                loss_dict["{}_{}_{}_{}".format(self.name, pair[0], pair[1],
                                               idx)] = loss
        return loss_dict