rec_aster_loss.py 3.6 KB
Newer Older
T
tink2123 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
# 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.

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import paddle
from paddle import nn
T
tink2123 已提交
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40


class CosineEmbeddingLoss(nn.Layer):
    def __init__(self, margin=0.):
        super(CosineEmbeddingLoss, self).__init__()
        self.margin = margin

    def forward(self, x1, x2, target):
        similarity = paddle.fluid.layers.reduce_sum(
            x1 * x2, dim=-1) / (paddle.norm(
                x1, axis=-1) * paddle.norm(
                    x2, axis=-1))
        one_list = paddle.full_like(target, fill_value=1)
        out = paddle.fluid.layers.reduce_mean(
            paddle.where(
                paddle.equal(target, one_list), 1. - similarity,
                paddle.maximum(
                    paddle.zeros_like(similarity), similarity - self.margin)))

        return out
T
tink2123 已提交
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56


class AsterLoss(nn.Layer):
    def __init__(self,
                 weight=None,
                 size_average=True,
                 ignore_index=-100,
                 sequence_normalize=False,
                 sample_normalize=True,
                 **kwargs):
        super(AsterLoss, self).__init__()
        self.weight = weight
        self.size_average = size_average
        self.ignore_index = ignore_index
        self.sequence_normalize = sequence_normalize
        self.sample_normalize = sample_normalize
T
tink2123 已提交
57 58 59
        self.loss_sem = CosineEmbeddingLoss()
        self.is_cosin_loss = True
        self.loss_func_rec = nn.CrossEntropyLoss(weight=None, reduction='none')
T
tink2123 已提交
60 61 62 63

    def forward(self, predicts, batch):
        targets = batch[1].astype("int64")
        label_lengths = batch[2].astype('int64')
T
tink2123 已提交
64
        sem_target = batch[3].astype('float32')
T
tink2123 已提交
65 66 67
        embedding_vectors = predicts['embedding_vectors']
        rec_pred = predicts['rec_pred']

T
tink2123 已提交
68 69 70 71 72 73
        if not self.is_cosin_loss:
            sem_loss = paddle.sum(self.loss_sem(embedding_vectors, sem_target))
        else:
            label_target = paddle.ones([embedding_vectors.shape[0]])
            sem_loss = paddle.sum(
                self.loss_sem(embedding_vectors, sem_target, label_target))
T
tink2123 已提交
74 75

        # rec loss
T
tink2123 已提交
76
        batch_size, def_max_length = targets.shape[0], targets.shape[1]
T
tink2123 已提交
77

T
tink2123 已提交
78
        mask = paddle.zeros([batch_size, def_max_length])
T
tink2123 已提交
79 80 81 82 83 84 85
        for i in range(batch_size):
            mask[i, :label_lengths[i]] = 1
        mask = paddle.cast(mask, "float32")
        max_length = max(label_lengths)
        assert max_length == rec_pred.shape[1]
        targets = targets[:, :max_length]
        mask = mask[:, :max_length]
T
tink2123 已提交
86
        rec_pred = paddle.reshape(rec_pred, [-1, rec_pred.shape[2]])
T
tink2123 已提交
87 88 89
        input = nn.functional.log_softmax(rec_pred, axis=1)
        targets = paddle.reshape(targets, [-1, 1])
        mask = paddle.reshape(mask, [-1, 1])
T
tink2123 已提交
90
        output = -paddle.index_sample(input, index=targets) * mask
T
tink2123 已提交
91 92 93 94 95
        output = paddle.sum(output)
        if self.sequence_normalize:
            output = output / paddle.sum(mask)
        if self.sample_normalize:
            output = output / batch_size
T
tink2123 已提交
96 97 98

        loss = output + sem_loss * 0.1
        return {'loss': loss}