# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
# Copyright 2021 Microsoft Research and The HuggingFace Inc. team. All rights reserved.
#
# 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.
""" Modeling classes for LayoutXLM model."""
import copy
import math
import paddle
import paddle.nn as nn
import paddle.nn.functional as F
from paddle.nn import CrossEntropyLoss, Layer
from paddlenlp.utils.log import logger
from .. import PretrainedModel, register_base_model
from .configuration import (
LAYOUTXLM_PRETRAINED_INIT_CONFIGURATION,
LAYOUTXLM_PRETRAINED_RESOURCE_FILES_MAP,
LayoutXLMConfig,
)
from .visual_backbone import build_resnet_fpn_backbone, read_config
__all__ = [
"LayoutXLMModel",
"LayoutXLMPretrainedModel",
"LayoutXLMForTokenClassification",
"LayoutXLMForSequenceClassification",
"LayoutXLMForPretraining",
"LayoutXLMForRelationExtraction",
"LayoutXLMForQuestionAnswering",
]
def relative_position_bucket(relative_position, bidirectional=True, num_buckets=32, max_distance=128):
ret = 0
if bidirectional:
num_buckets //= 2
ret += (relative_position > 0).astype(paddle.int64) * num_buckets
n = paddle.abs(relative_position)
else:
n = paddle.max(-relative_position, paddle.zeros_like(relative_position))
# Now n is in the range [0, inf)
# half of the buckets are for exact increments in positions
max_exact = num_buckets // 2
is_small = n < max_exact
# The other half of the buckets are for logarithmically bigger bins in positions up to max_distance
val_if_large = max_exact + (
paddle.log(n.astype(paddle.float32) / max_exact)
/ math.log(max_distance / max_exact)
* (num_buckets - max_exact)
).astype(paddle.int64)
val_if_large = paddle.minimum(val_if_large, paddle.full_like(val_if_large, num_buckets - 1))
ret += paddle.where(is_small, n, val_if_large)
return ret
def token_featue_to_sequence_feature(input_ids, seq_length, sequence_output):
"""
used to transform token feature into sequence feature by
averaging all the token features of certain sequence
"""
batches = input_ids.shape[0]
for batch_id in range(batches):
start_idx = -1
for i in range(0, seq_length):
if input_ids[batch_id, i] == 6:
if start_idx > -1:
feature_block = sequence_output[batch_id, start_idx + 1 : i]
sequence_output[batch_id, start_idx] = paddle.mean(feature_block, axis=0)
start_idx = i
if input_ids[batch_id, i] == 1:
feature_block = sequence_output[batch_id, start_idx + 1 : i]
sequence_output[batch_id, start_idx] = paddle.mean(feature_block, axis=0)
break
if i == seq_length - 1:
sequence_output[batch_id, start_idx] = paddle.mean(feature_block, axis=0)
return
class LayoutXLMPooler(Layer):
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMPooler, self).__init__()
self.dense = nn.Linear(config.hidden_size, config.hidden_size)
self.activation = nn.Tanh()
self.with_pool = config.with_pool
def forward(self, hidden_states):
# We "pool" the model by simply taking the hidden state corresponding
# to the first token.
first_token_tensor = hidden_states[:, 0]
pooled_output = self.dense(first_token_tensor)
if self.with_pool == "tanh":
pooled_output = self.activation(pooled_output)
return pooled_output
class LayoutXLMEmbeddings(Layer):
"""
Include embeddings from word, position and token_type embeddings
"""
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMEmbeddings, self).__init__()
self.word_embeddings = nn.Embedding(config.vocab_size, config.hidden_size, padding_idx=0)
self.position_embeddings = nn.Embedding(config.max_position_embeddings, config.hidden_size)
self.x_position_embeddings = nn.Embedding(config.max_2d_position_embeddings, config.coordinate_size)
self.y_position_embeddings = nn.Embedding(config.max_2d_position_embeddings, config.coordinate_size)
self.h_position_embeddings = nn.Embedding(config.max_2d_position_embeddings, config.coordinate_size)
self.w_position_embeddings = nn.Embedding(config.max_2d_position_embeddings, config.coordinate_size)
self.token_type_embeddings = nn.Embedding(config.type_vocab_size, config.hidden_size)
self.LayerNorm = nn.LayerNorm(config.hidden_size, epsilon=config.layer_norm_eps)
self.dropout = nn.Dropout(config.hidden_dropout_prob)
self.register_buffer(
"position_ids", paddle.arange(config.max_position_embeddings, dtype="int64").expand((1, -1))
)
def _cal_spatial_position_embeddings(self, bbox):
try:
left_position_embeddings = self.x_position_embeddings(bbox[:, :, 0])
upper_position_embeddings = self.y_position_embeddings(bbox[:, :, 1])
right_position_embeddings = self.x_position_embeddings(bbox[:, :, 2])
lower_position_embeddings = self.y_position_embeddings(bbox[:, :, 3])
except IndexError as e:
raise IndexError("The :obj:`bbox`coordinate values should be within 0-1000 range.") from e
h_position_embeddings = self.h_position_embeddings(bbox[:, :, 3] - bbox[:, :, 1])
w_position_embeddings = self.w_position_embeddings(bbox[:, :, 2] - bbox[:, :, 0])
spatial_position_embeddings = paddle.concat(
[
left_position_embeddings,
upper_position_embeddings,
right_position_embeddings,
lower_position_embeddings,
h_position_embeddings,
w_position_embeddings,
],
axis=-1,
)
return spatial_position_embeddings
def forward(self, input_ids, bbox=None, token_type_ids=None, position_ids=None):
if position_ids is None:
ones = paddle.ones_like(input_ids, dtype="int64")
seq_length = paddle.cumsum(ones, axis=-1)
position_ids = seq_length - ones
position_ids.stop_gradient = True
if token_type_ids is None:
token_type_ids = paddle.zeros_like(input_ids, dtype="int64")
input_embedings = self.word_embeddings(input_ids)
position_embeddings = self.position_embeddings(position_ids)
try:
left_position_embeddings = self.x_position_embeddings(bbox[:, :, 0])
upper_position_embeddings = self.y_position_embeddings(bbox[:, :, 1])
right_position_embeddings = self.x_position_embeddings(bbox[:, :, 2])
lower_position_embeddings = self.y_position_embeddings(bbox[:, :, 3])
except IndexError as e:
raise IndexError("The :obj:`bbox`coordinate values should be within 0-1000 range.") from e
h_position_embeddings = self.h_position_embeddings(bbox[:, :, 3] - bbox[:, :, 1])
w_position_embeddings = self.w_position_embeddings(bbox[:, :, 2] - bbox[:, :, 0])
token_type_embeddings = self.token_type_embeddings(token_type_ids)
embeddings = (
input_embedings
+ position_embeddings
+ left_position_embeddings
+ upper_position_embeddings
+ right_position_embeddings
+ lower_position_embeddings
+ h_position_embeddings
+ w_position_embeddings
+ token_type_embeddings
)
embeddings = self.layer_norm(embeddings)
embeddings = self.dropout(embeddings)
return embeddings
[文档]class LayoutXLMPretrainedModel(PretrainedModel):
config_class = LayoutXLMConfig
pretrained_init_configuration = LAYOUTXLM_PRETRAINED_INIT_CONFIGURATION
pretrained_resource_files_map = LAYOUTXLM_PRETRAINED_RESOURCE_FILES_MAP
base_model_prefix = "layoutxlm"
def _init_weights(self, layer):
"""Initialization hook"""
if isinstance(layer, (nn.Linear, nn.Embedding)):
if isinstance(layer.weight, paddle.Tensor):
layer.weight.set_value(
paddle.tensor.normal(
mean=0.0,
std=self.pretrained_init_configuration["initializer_range"]
if "initializer_range" in self.pretrained_init_configuration
else 0.02,
shape=layer.weight.shape,
)
)
class LayoutXLMSelfOutput(nn.Layer):
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMSelfOutput, self).__init__()
self.dense = nn.Linear(config.hidden_size, config.hidden_size)
self.LayerNorm = nn.LayerNorm(config.hidden_size, epsilon=config.layer_norm_eps)
self.dropout = nn.Dropout(config.hidden_dropout_prob)
def forward(self, hidden_states, input_tensor):
hidden_states = self.dense(hidden_states)
hidden_states = self.dropout(hidden_states)
hidden_states = self.LayerNorm(hidden_states + input_tensor.astype(hidden_states.dtype))
return hidden_states
class LayoutXLMSelfAttention(nn.Layer):
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMSelfAttention, self).__init__()
if config.hidden_size % config.num_attention_heads != 0 and not hasattr(config, "embedding_size"):
raise ValueError(
"The hidden size {} is not a multiple of the number of attention "
"heads {}".format(config.hidden_size, config.num_attention_heads)
)
self.fast_qkv = config.fast_qkv
self.num_attention_heads = config.num_attention_heads
self.attention_head_size = int(config.hidden_size / config.num_attention_heads)
self.all_head_size = self.num_attention_heads * self.attention_head_size
self.has_relative_attention_bias = config.has_relative_attention_bias
self.has_spatial_attention_bias = config.has_spatial_attention_bias
if config.fast_qkv:
self.qkv_linear = nn.Linear(config.hidden_size, 3 * self.all_head_size, bias_attr=False)
self.q_bias = self.create_parameter(
shape=[1, 1, self.all_head_size], default_initializer=nn.initializer.Constant(0.0)
)
self.v_bias = self.create_parameter(
shape=[1, 1, self.all_head_size], default_initializer=nn.initializer.Constant(0.0)
)
else:
self.query = nn.Linear(config.hidden_size, self.all_head_size)
self.key = nn.Linear(config.hidden_size, self.all_head_size)
self.value = nn.Linear(config.hidden_size, self.all_head_size)
self.dropout = nn.Dropout(config.attention_probs_dropout_prob)
def transpose_for_scores(self, x):
new_x_shape = list(x.shape[:-1]) + [self.num_attention_heads, self.attention_head_size]
x = x.reshape(new_x_shape)
return x.transpose([0, 2, 1, 3])
def compute_qkv(self, hidden_states):
if self.fast_qkv:
qkv = self.qkv_linear(hidden_states)
q, k, v = paddle.chunk(qkv, 3, axis=-1)
if q.ndimension() == self.q_bias.ndimension():
q = q + self.q_bias
v = v + self.v_bias
else:
_sz = (1,) * (q.ndimension() - 1) + (-1,)
q = q + self.q_bias.reshape(_sz)
v = v + self.v_bias.vreshape(_sz)
else:
q = self.query(hidden_states)
k = self.key(hidden_states)
v = self.value(hidden_states)
return q, k, v
def forward(
self,
hidden_states,
attention_mask=None,
head_mask=None,
encoder_hidden_states=None,
encoder_attention_mask=None,
past_key_value=None,
output_attentions=False,
rel_pos=None,
rel_2d_pos=None,
):
q, k, v = self.compute_qkv(hidden_states)
# (B, L, H*D) -> (B, H, L, D)
query_layer = self.transpose_for_scores(q)
key_layer = self.transpose_for_scores(k)
value_layer = self.transpose_for_scores(v)
query_layer = query_layer / math.sqrt(self.attention_head_size)
# [BSZ, NAT, L, L]
attention_scores = paddle.matmul(query_layer, key_layer.transpose([0, 1, 3, 2]))
if self.has_relative_attention_bias:
attention_scores += rel_pos
if self.has_spatial_attention_bias:
attention_scores += rel_2d_pos
bool_attention_mask = attention_mask.astype(paddle.bool)
bool_attention_mask.stop_gradient = True
attention_scores_shape = paddle.shape(attention_scores)
attention_scores = paddle.where(
bool_attention_mask.expand(attention_scores_shape),
paddle.ones(attention_scores_shape) * float("-1e10"),
attention_scores,
)
attention_probs = F.softmax(attention_scores, axis=-1)
# This is actually dropping out entire tokens to attend to, which might
# seem a bit unusual, but is taken from the original Transformer paper.
attention_probs = self.dropout(attention_probs)
context_layer = paddle.matmul(attention_probs, value_layer)
context_layer = context_layer.transpose([0, 2, 1, 3])
new_context_layer_shape = list(context_layer.shape[:-2]) + [self.all_head_size]
context_layer = context_layer.reshape(new_context_layer_shape)
if output_attentions:
outputs = [context_layer, attention_probs]
else:
outputs = [context_layer]
return outputs
class LayoutXLMAttention(nn.Layer):
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMAttention, self).__init__()
self.self = LayoutXLMSelfAttention(config)
self.output = LayoutXLMSelfOutput(config)
def forward(
self,
hidden_states,
attention_mask=None,
head_mask=None,
encoder_hidden_states=None,
encoder_attention_mask=None,
past_key_value=None,
output_attentions=False,
rel_pos=None,
rel_2d_pos=None,
):
self_outputs = self.self(
hidden_states,
attention_mask,
head_mask,
encoder_hidden_states,
encoder_attention_mask,
past_key_value,
output_attentions,
rel_pos=rel_pos,
rel_2d_pos=rel_2d_pos,
)
attention_output = self.output(self_outputs[0], hidden_states)
# add attentions if we output them
if output_attentions:
outputs = [
attention_output,
] + self_outputs[1:]
else:
outputs = [attention_output]
return outputs
class LayoutXLMEncoder(nn.Layer):
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMEncoder, self).__init__()
self.config = config
self.layer = nn.LayerList([LayoutXLMLayer(config) for _ in range(config.num_hidden_layers)])
self.has_relative_attention_bias = config.has_relative_attention_bias
self.has_spatial_attention_bias = config.has_spatial_attention_bias
if self.has_relative_attention_bias:
self.rel_pos_bins = config.rel_pos_bins
self.max_rel_pos = config.max_rel_pos
self.rel_pos_onehot_size = config.rel_pos_bins
self.rel_pos_bias = nn.Linear(self.rel_pos_onehot_size, config.num_attention_heads, bias_attr=False)
if self.has_spatial_attention_bias:
self.max_rel_2d_pos = config.max_rel_2d_pos
self.rel_2d_pos_bins = config.rel_2d_pos_bins
self.rel_2d_pos_onehot_size = config.rel_2d_pos_bins
self.rel_pos_x_bias = nn.Linear(self.rel_2d_pos_onehot_size, config.num_attention_heads, bias_attr=False)
self.rel_pos_y_bias = nn.Linear(self.rel_2d_pos_onehot_size, config.num_attention_heads, bias_attr=False)
def _cal_1d_pos_emb(self, hidden_states, position_ids):
rel_pos_mat = position_ids.unsqueeze(-2) - position_ids.unsqueeze(-1)
rel_pos = relative_position_bucket(
rel_pos_mat,
num_buckets=self.rel_pos_bins,
max_distance=self.max_rel_pos,
)
rel_pos = paddle.nn.functional.one_hot(rel_pos, num_classes=self.rel_pos_onehot_size).astype(
hidden_states.dtype
)
rel_pos = self.rel_pos_bias(rel_pos).transpose([0, 3, 1, 2])
return rel_pos
def _cal_2d_pos_emb(self, hidden_states, bbox):
position_coord_x = bbox[:, :, 0]
position_coord_y = bbox[:, :, 3]
rel_pos_x_2d_mat = position_coord_x.unsqueeze(-2) - position_coord_x.unsqueeze(-1)
rel_pos_y_2d_mat = position_coord_y.unsqueeze(-2) - position_coord_y.unsqueeze(-1)
rel_pos_x = relative_position_bucket(
rel_pos_x_2d_mat,
num_buckets=self.rel_2d_pos_bins,
max_distance=self.max_rel_2d_pos,
)
rel_pos_y = relative_position_bucket(
rel_pos_y_2d_mat,
num_buckets=self.rel_2d_pos_bins,
max_distance=self.max_rel_2d_pos,
)
rel_pos_x = F.one_hot(rel_pos_x, num_classes=self.rel_2d_pos_onehot_size).astype(hidden_states.dtype)
rel_pos_y = F.one_hot(rel_pos_y, num_classes=self.rel_2d_pos_onehot_size).astype(hidden_states.dtype)
rel_pos_x = self.rel_pos_x_bias(rel_pos_x).transpose([0, 3, 1, 2])
rel_pos_y = self.rel_pos_y_bias(rel_pos_y).transpose([0, 3, 1, 2])
rel_2d_pos = rel_pos_x + rel_pos_y
return rel_2d_pos
def forward(
self,
hidden_states,
attention_mask=None,
head_mask=None,
encoder_hidden_states=None,
encoder_attention_mask=None,
past_key_values=None,
output_attentions=False,
output_hidden_states=False,
bbox=None,
position_ids=None,
):
all_hidden_states = () if output_hidden_states else None
rel_pos = self._cal_1d_pos_emb(hidden_states, position_ids) if self.has_relative_attention_bias else None
rel_2d_pos = self._cal_2d_pos_emb(hidden_states, bbox) if self.has_spatial_attention_bias else None
hidden_save = dict()
hidden_save["input_hidden_states"] = hidden_states
for i, layer_module in enumerate(self.layer):
if output_hidden_states:
all_hidden_states = all_hidden_states + (hidden_states,)
layer_head_mask = head_mask[i] if head_mask is not None else None
past_key_value = past_key_values[i] if past_key_values is not None else None
# gradient_checkpointing is set as False here so we remove some codes here
hidden_save["input_attention_mask"] = attention_mask
hidden_save["input_layer_head_mask"] = layer_head_mask
layer_outputs = layer_module(
hidden_states,
attention_mask,
layer_head_mask,
encoder_hidden_states,
encoder_attention_mask,
past_key_value,
output_attentions,
rel_pos=rel_pos,
rel_2d_pos=rel_2d_pos,
)
hidden_states = layer_outputs[0]
hidden_save["{}_data".format(i)] = hidden_states
return hidden_states, hidden_save
class LayoutXLMIntermediate(nn.Layer):
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMIntermediate, self).__init__()
self.dense = nn.Linear(config.hidden_size, config.intermediate_size)
if config.hidden_act == "gelu":
self.intermediate_act_fn = nn.GELU()
else:
assert False, "hidden_act is set as: {}, please check it..".format(config.hidden_act)
def forward(self, hidden_states):
hidden_states = self.dense(hidden_states)
hidden_states = self.intermediate_act_fn(hidden_states)
return hidden_states
class LayoutXLMOutput(nn.Layer):
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMOutput, self).__init__()
self.dense = nn.Linear(config.intermediate_size, config.hidden_size)
self.LayerNorm = nn.LayerNorm(config.hidden_size, epsilon=config.layer_norm_eps)
self.dropout = nn.Dropout(config.hidden_dropout_prob)
def forward(self, hidden_states, input_tensor):
hidden_states = self.dense(hidden_states)
hidden_states = self.dropout(hidden_states)
hidden_states = self.LayerNorm(hidden_states + input_tensor)
return hidden_states
class LayoutXLMLayer(nn.Layer):
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMLayer, self).__init__()
# since chunk_size_feed_forward is 0 as default, no chunk is needed here.
self.seq_len_dim = 1
self.attention = LayoutXLMAttention(config)
self.add_cross_attention = False # default as false
self.intermediate = LayoutXLMIntermediate(config)
self.output = LayoutXLMOutput(config)
def feed_forward_chunk(self, attention_output):
intermediate_output = self.intermediate(attention_output)
layer_output = self.output(intermediate_output, attention_output)
return layer_output
def forward(
self,
hidden_states,
attention_mask=None,
head_mask=None,
encoder_hidden_states=None,
encoder_attention_mask=None,
past_key_value=None,
output_attentions=False,
rel_pos=None,
rel_2d_pos=None,
):
# decoder uni-directional self-attention cached key/values tuple is at positions 1,2
self_attn_past_key_value = past_key_value[:2] if past_key_value is not None else None
self_attention_outputs = self.attention(
hidden_states,
attention_mask,
head_mask,
output_attentions=output_attentions,
past_key_value=self_attn_past_key_value,
rel_pos=rel_pos,
rel_2d_pos=rel_2d_pos,
)
attention_output = self_attention_outputs[0]
layer_output = self.feed_forward_chunk(attention_output)
if output_attentions:
outputs = self_attention_outputs[1:] # add self attentions if we output attention weights
outputs = [
layer_output,
] + list(outputs)
else:
outputs = [layer_output]
return outputs
class VisualBackbone(nn.Layer):
def __init__(self, config: LayoutXLMConfig):
super(VisualBackbone, self).__init__()
self.cfg = read_config()
self.backbone = build_resnet_fpn_backbone(self.cfg)
assert len(self.cfg.MODEL.PIXEL_MEAN) == len(self.cfg.MODEL.PIXEL_STD)
num_channels = len(self.cfg.MODEL.PIXEL_MEAN)
self.register_buffer("pixel_mean", paddle.to_tensor(self.cfg.MODEL.PIXEL_MEAN).reshape([num_channels, 1, 1]))
self.register_buffer("pixel_std", paddle.to_tensor(self.cfg.MODEL.PIXEL_STD).reshape([num_channels, 1, 1]))
self.out_feature_key = "p2"
# is_deterministic is disabled here.
self.pool = nn.AdaptiveAvgPool2D(config.image_feature_pool_shape[:2])
if len(config.image_feature_pool_shape) == 2:
config.image_feature_pool_shape.append(self.backbone.output_shape()[self.out_feature_key].channels)
assert self.backbone.output_shape()[self.out_feature_key].channels == config.image_feature_pool_shape[2]
def forward(self, images):
images_input = (paddle.to_tensor(images) - self.pixel_mean) / self.pixel_std
features = self.backbone(images_input)
features = features[self.out_feature_key]
features = self.pool(features).flatten(start_axis=2).transpose([0, 2, 1])
return features
[文档]@register_base_model
class LayoutXLMModel(LayoutXLMPretrainedModel):
"""
The bare LayoutXLM Model outputting raw hidden-states.
This model inherits from :class:`~paddlenlp.transformers.model_utils.PretrainedModel`.
Refer to the superclass documentation for the generic methods.
This model is also a Paddle `paddle.nn.Layer <https://www.paddlepaddle.org.cn/documentation
/docs/zh/api/paddle/nn/Layer_cn.html>`__ subclass. Use it as a regular Paddle Layer
and refer to the Paddle documentation for all matter related to general usage and behavior.
Args:
vocab_size (`int`):
Vocabulary size of the XLNet model. Defines the number of different tokens that can
be represented by the `inputs_ids` passed when calling XLNetModel.
hidden_size (`int`, optional):
Dimensionality of the encoder layers and the pooler layer. Defaults to ``768``.
num_hidden_layers (`int`, optional):
Number of hidden layers in the Transformer encoder. Defaults to ``12``.
num_attention_heads (`int`, optional):
Number of attention heads for each attention layer in the Transformer encoder.
Defaults to ``12``.
intermediate_size (`int`, optional):
Dimensionality of the "intermediate" (often named feed-forward) layer in the Transformer encoder.
Defaults to ``3072``.
hidden_act (`str`, optional):
The non-linear activation function in the feed-forward layer.
``"gelu"``, ``"relu"`` and any other paddle supported activation functions
are supported. Defaults to ``"gelu"``.
hidden_dropout_prob (`float`, optional):
The dropout probability for all fully connected layers in the embeddings and encoder.
Defaults to ``0.1``.
attention_probs_dropout_prob (`float`, optional):
The dropout probability for all fully connected layers in the pooler.
Defaults to ``0.1``.
initializer_range (`float`, optional):
The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
Defaults to ``0.02``.
"""
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMModel, self).__init__(config)
self.config = config
self.use_visual_backbone = config.use_visual_backbone
self.has_visual_segment_embedding = config.has_visual_segment_embedding
self.embeddings = LayoutXLMEmbeddings(config)
if self.use_visual_backbone is True:
self.visual = VisualBackbone(config)
self.visual.stop_gradient = True
self.visual_proj = nn.Linear(config.image_feature_pool_shape[-1], config.hidden_size)
if self.has_visual_segment_embedding:
self.visual_segment_embedding = self.create_parameter(
shape=[
config.hidden_size,
],
dtype=paddle.float32,
)
self.visual_LayerNorm = nn.LayerNorm(config.hidden_size, epsilon=config.layer_norm_eps)
self.visual_dropout = nn.Dropout(config.hidden_dropout_prob)
self.encoder = LayoutXLMEncoder(config)
self.pooler = LayoutXLMPooler(config)
def _calc_text_embeddings(self, input_ids, bbox, position_ids, token_type_ids):
words_embeddings = self.embeddings.word_embeddings(input_ids)
position_embeddings = self.embeddings.position_embeddings(position_ids)
spatial_position_embeddings = self.embeddings._cal_spatial_position_embeddings(bbox)
token_type_embeddings = self.embeddings.token_type_embeddings(token_type_ids)
embeddings = words_embeddings + position_embeddings + spatial_position_embeddings + token_type_embeddings
embeddings = self.embeddings.LayerNorm(embeddings)
embeddings = self.embeddings.dropout(embeddings)
return embeddings
def _calc_visual_bbox(self, image_feature_pool_shape, bbox, visual_shape):
visual_bbox_x = (
paddle.arange(
0,
1000 * (image_feature_pool_shape[1] + 1),
1000,
dtype=bbox.dtype,
)
// image_feature_pool_shape[1]
)
visual_bbox_y = (
paddle.arange(
0,
1000 * (image_feature_pool_shape[0] + 1),
1000,
dtype=bbox.dtype,
)
// image_feature_pool_shape[0]
)
expand_shape = image_feature_pool_shape[0:2]
visual_bbox = paddle.stack(
[
visual_bbox_x[:-1].expand(expand_shape),
visual_bbox_y[:-1].expand(expand_shape[::-1]).transpose([1, 0]),
visual_bbox_x[1:].expand(expand_shape),
visual_bbox_y[1:].expand(expand_shape[::-1]).transpose([1, 0]),
],
axis=-1,
).reshape([expand_shape[0] * expand_shape[1], paddle.shape(bbox)[-1]])
visual_bbox = visual_bbox.expand([visual_shape[0], visual_bbox.shape[0], visual_bbox.shape[1]])
return visual_bbox
def _calc_img_embeddings(self, image, bbox, position_ids):
use_image_info = self.use_visual_backbone and image is not None
position_embeddings = self.embeddings.position_embeddings(position_ids)
spatial_position_embeddings = self.embeddings._cal_spatial_position_embeddings(bbox)
if use_image_info is True:
visual_embeddings = self.visual_proj(self.visual(image.astype(paddle.float32)))
embeddings = visual_embeddings + position_embeddings + spatial_position_embeddings
else:
embeddings = position_embeddings + spatial_position_embeddings
if self.has_visual_segment_embedding:
embeddings += self.visual_segment_embedding
embeddings = self.visual_LayerNorm(embeddings)
embeddings = self.visual_dropout(embeddings)
return embeddings
[文档] def resize_position_embeddings(self, new_num_position_embeddings):
"""
Resizes position embeddings of the model if `new_num_position_embeddings != config["max_position_embeddings"]`.
Arguments:
new_num_position_embeddings (`int`):
The number of new position embedding matrix. If position embeddings are learned, increasing the size
will add newly initialized vectors at the end, whereas reducing the size will remove vectors from the
end.
"""
num_position_embeds_diff = new_num_position_embeddings - self.config.max_position_embeddings
# no resizing needs to be done if the length stays the same
if num_position_embeds_diff == 0:
return
logger.info(f"Setting `config.max_position_embeddings={new_num_position_embeddings}`...")
self.config.max_position_embeddings = new_num_position_embeddings
old_position_embeddings_weight = self.embeddings.position_embeddings.weight
self.embeddings.position_embeddings = nn.Embedding(
self.config.max_position_embeddings, self.config.hidden_size
)
with paddle.no_grad():
if num_position_embeds_diff > 0:
self.embeddings.position_embeddings.weight[:-num_position_embeds_diff] = old_position_embeddings_weight
else:
self.embeddings.position_embeddings.weight = old_position_embeddings_weight[:num_position_embeds_diff]
[文档] def forward(
self,
input_ids=None,
bbox=None,
image=None,
token_type_ids=None,
position_ids=None,
attention_mask=None,
head_mask=None,
output_hidden_states=False,
output_attentions=False,
):
input_shape = paddle.shape(input_ids)
visual_shape = list(input_shape)
visual_shape[1] = self.config.image_feature_pool_shape[0] * self.config.image_feature_pool_shape[1]
visual_bbox = self._calc_visual_bbox(self.config.image_feature_pool_shape, bbox, visual_shape)
final_bbox = paddle.concat([bbox, visual_bbox], axis=1)
if attention_mask is None:
attention_mask = paddle.ones(input_shape)
if self.use_visual_backbone is True:
visual_attention_mask = paddle.ones(visual_shape)
else:
visual_attention_mask = paddle.zeros(visual_shape)
attention_mask = attention_mask.astype(visual_attention_mask.dtype)
final_attention_mask = paddle.concat([attention_mask, visual_attention_mask], axis=1)
if token_type_ids is None:
token_type_ids = paddle.zeros(input_shape, dtype=paddle.int64)
if position_ids is None:
seq_length = input_shape[1]
position_ids = self.embeddings.position_ids[:, :seq_length]
position_ids = position_ids.expand(input_shape)
visual_position_ids = paddle.arange(0, visual_shape[1]).expand([input_shape[0], visual_shape[1]])
final_position_ids = paddle.concat([position_ids, visual_position_ids], axis=1)
if bbox is None:
bbox = paddle.zeros(input_shape + [4])
text_layout_emb = self._calc_text_embeddings(
input_ids=input_ids,
bbox=bbox,
token_type_ids=token_type_ids,
position_ids=position_ids,
)
visual_emb = self._calc_img_embeddings(
image=image,
bbox=visual_bbox,
position_ids=visual_position_ids,
)
final_emb = paddle.concat([text_layout_emb, visual_emb], axis=1)
extended_attention_mask = final_attention_mask.unsqueeze(1).unsqueeze(2)
extended_attention_mask = (1.0 - extended_attention_mask) * -10000.0
if head_mask is not None:
if head_mask.dim() == 1:
head_mask = head_mask.unsqueeze(0).unsqueeze(0).unsqueeze(-1).unsqueeze(-1)
head_mask = head_mask.expand(self.config.num_hidden_layers, -1, -1, -1, -1)
elif head_mask.dim() == 2:
head_mask = head_mask.unsqueeze(1).unsqueeze(-1).unsqueeze(-1)
else:
head_mask = [None] * self.config.num_hidden_layers
encoder_outputs = self.encoder(
final_emb,
extended_attention_mask,
bbox=final_bbox,
position_ids=final_position_ids,
head_mask=head_mask,
output_attentions=output_attentions,
output_hidden_states=output_hidden_states,
)
sequence_output = encoder_outputs[0]
pooled_output = self.pooler(sequence_output)
return sequence_output, pooled_output, encoder_outputs[1]
[文档]class LayoutXLMForTokenClassification(LayoutXLMPretrainedModel):
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMForTokenClassification, self).__init__(config)
self.num_classes = config.num_labels
self.layoutxlm = LayoutXLMModel(config)
self.dropout = nn.Dropout(config.hidden_dropout_prob)
self.classifier = nn.Linear(config.hidden_size, self.num_classes)
[文档] def resize_position_embeddings(self, new_num_position_embeddings):
"""
Resizes position embeddings of the model if `new_num_position_embeddings != config["max_position_embeddings"]`.
Arguments:
new_num_position_embeddings (`int`):
The number of new position embedding matrix. If position embeddings are learned, increasing the size
will add newly initialized vectors at the end, whereas reducing the size will remove vectors from the
end.
"""
self.layoutxlm.resize_position_embeddings(new_num_position_embeddings)
[文档] def forward(
self,
input_ids=None,
bbox=None,
image=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
head_mask=None,
labels=None,
):
outputs = self.layoutxlm(
input_ids=input_ids,
bbox=bbox,
image=image,
attention_mask=attention_mask,
token_type_ids=token_type_ids,
position_ids=position_ids,
head_mask=head_mask,
)
seq_length = input_ids.shape[1]
# sequence out and image out
sequence_output = outputs[0][:, :seq_length]
sequence_output = self.dropout(sequence_output)
logits = self.classifier(sequence_output)
hidden_states = {
f"hidden_states_{idx}": outputs[2][f"{idx}_data"] for idx in range(self.layoutxlm.config.num_hidden_layers)
}
if self.training:
outputs = (logits, hidden_states)
else:
outputs = (logits,)
if labels is not None:
loss_fct = nn.CrossEntropyLoss()
if attention_mask is not None:
active_loss = (
attention_mask.reshape(
[
-1,
]
)
== 1
)
active_logits = logits.reshape([-1, self.num_classes])[active_loss]
active_labels = labels.reshape(
[
-1,
]
)[active_loss]
loss = loss_fct(active_logits, active_labels)
else:
loss = loss_fct(
logits.reshape([-1, self.num_classes]),
labels.reshape(
[
-1,
]
),
)
outputs = (loss,) + outputs
return outputs
[文档]class LayoutXLMForSequenceClassification(LayoutXLMPretrainedModel):
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMForSequenceClassification, self).__init__(config)
self.num_classes = config.num_labels
self.layoutxlm = LayoutXLMModel(config)
self.dropout = nn.Dropout(config.hidden_dropout_prob)
self.classifier = nn.Linear(config.hidden_size * 3, self.num_classes)
[文档] def resize_position_embeddings(self, new_num_position_embeddings):
"""
Resizes position embeddings of the model if `new_num_position_embeddings != config["max_position_embeddings"]`.
Arguments:
new_num_position_embeddings (`int`):
The number of new position embedding matrix. If position embeddings are learned, increasing the size
will add newly initialized vectors at the end, whereas reducing the size will remove vectors from the
end.
"""
self.layoutxlm.resize_position_embeddings(new_num_position_embeddings)
[文档] def forward(
self,
input_ids=None,
bbox=None,
image=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
head_mask=None,
labels=None,
):
input_shape = paddle.shape(input_ids)
visual_shape = list(input_shape)
visual_shape[1] = (
self.layoutxlm.config.image_feature_pool_shape[0] * self.layoutxlm.config.image_feature_pool_shape[1]
)
visual_bbox = self.layoutxlm._calc_visual_bbox(
self.layoutxlm.config.image_feature_pool_shape, bbox, visual_shape
)
visual_position_ids = paddle.arange(0, visual_shape[1]).expand([input_shape[0], visual_shape[1]])
initial_image_embeddings = self.layoutxlm._calc_img_embeddings(
image=image,
bbox=visual_bbox,
position_ids=visual_position_ids,
)
outputs = self.layoutxlm(
input_ids=input_ids,
bbox=bbox,
image=image,
attention_mask=attention_mask,
token_type_ids=token_type_ids,
position_ids=position_ids,
head_mask=head_mask,
)
seq_length = input_ids.shape[1]
# sequence out and image out
sequence_output, final_image_embeddings = outputs[0][:, :seq_length], outputs[0][:, seq_length:]
cls_final_output = sequence_output[:, 0, :]
# average-pool the visual embeddings
pooled_initial_image_embeddings = initial_image_embeddings.mean(axis=1)
pooled_final_image_embeddings = final_image_embeddings.mean(axis=1)
# concatenate with cls_final_output
sequence_output = paddle.concat(
[cls_final_output, pooled_initial_image_embeddings, pooled_final_image_embeddings], axis=1
)
sequence_output = self.dropout(sequence_output)
logits = self.classifier(sequence_output)
outputs = (logits,)
if labels is not None:
loss_fct = nn.CrossEntropyLoss()
loss = loss_fct(
logits.reshape([-1, self.num_classes]),
labels.reshape(
[
-1,
]
),
)
outputs = (loss,) + outputs
return outputs
class LayoutXLMPredictionHead(Layer):
"""
Bert Model with a `language modeling` head on top for CLM fine-tuning.
"""
def __init__(self, hidden_size, vocab_size, activation, embedding_weights=None):
super(LayoutXLMPredictionHead, self).__init__()
self.transform = nn.Linear(hidden_size, hidden_size)
self.activation = getattr(nn.functional, activation)
self.layer_norm = nn.LayerNorm(hidden_size)
self.decoder_weight = (
self.create_parameter(shape=[vocab_size, hidden_size], dtype=self.transform.weight.dtype, is_bias=False)
if embedding_weights is None
else embedding_weights
)
self.decoder_bias = self.create_parameter(shape=[vocab_size], dtype=self.decoder_weight.dtype, is_bias=True)
def forward(self, hidden_states, masked_positions=None):
if masked_positions is not None:
hidden_states = paddle.reshape(hidden_states, [-1, hidden_states.shape[-1]])
hidden_states = paddle.tensor.gather(hidden_states, masked_positions)
# gather masked tokens might be more quick
hidden_states = self.transform(hidden_states)
hidden_states = self.activation(hidden_states)
hidden_states = self.layer_norm(hidden_states)
hidden_states = paddle.tensor.matmul(hidden_states, self.decoder_weight, transpose_y=True) + self.decoder_bias
return hidden_states
class LayoutXLMPretrainingHeads(Layer):
def __init__(self, hidden_size, vocab_size, activation, embedding_weights=None):
super(LayoutXLMPretrainingHeads, self).__init__()
self.predictions = LayoutXLMPredictionHead(hidden_size, vocab_size, activation, embedding_weights)
def forward(self, sequence_output, masked_positions=None):
prediction_scores = self.predictions(sequence_output, masked_positions)
return prediction_scores
[文档]class LayoutXLMForPretraining(LayoutXLMPretrainedModel):
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMForPretraining, self).__init__(config)
self.layoutxlm = LayoutXLMModel(config)
self.cls = LayoutXLMPretrainingHeads(
config.hidden_size,
config.vocab_size,
config.hidden_act,
embedding_weights=self.layoutxlm.embeddings.word_embeddings.weight,
)
[文档] def resize_position_embeddings(self, new_num_position_embeddings):
"""
Resizes position embeddings of the model if `new_num_position_embeddings != config["max_position_embeddings"]`.
Arguments:
new_num_position_embeddings (`int`):
The number of new position embedding matrix. If position embeddings are learned, increasing the size
will add newly initialized vectors at the end, whereas reducing the size will remove vectors from the
end.
"""
self.layoutxlm.resize_position_embeddings(new_num_position_embeddings)
[文档] def forward(
self,
input_ids=None,
bbox=None,
image=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
head_mask=None,
masked_positions=None,
):
outputs = self.layoutxlm(
input_ids=input_ids,
bbox=bbox,
image=image,
attention_mask=attention_mask,
token_type_ids=token_type_ids,
position_ids=position_ids,
head_mask=head_mask,
)
sequence_output = outputs[0]
prediction_scores = self.cls(sequence_output, masked_positions)
return prediction_scores
class BiaffineAttention(nn.Layer):
"""Implements a biaffine attention operator for binary relation classification."""
def __init__(self, in_features, out_features):
super(BiaffineAttention, self).__init__()
self.in_features = in_features
self.out_features = out_features
self.bilinear = nn.Bilinear(in_features, in_features, out_features, bias_attr=False)
self.linear = nn.Linear(2 * in_features, out_features)
def forward(self, x_1, x_2):
return self.bilinear(x_1, x_2) + self.linear(paddle.concat((x_1, x_2), axis=-1))
class REDecoder(nn.Layer):
def __init__(self, hidden_size=768, hidden_dropout_prob=0.1):
super(REDecoder, self).__init__()
self.entity_emb = nn.Embedding(3, hidden_size)
projection = nn.Sequential(
nn.Linear(hidden_size * 2, hidden_size),
nn.ReLU(),
nn.Dropout(hidden_dropout_prob),
nn.Linear(hidden_size, hidden_size // 2),
nn.ReLU(),
nn.Dropout(hidden_dropout_prob),
)
self.ffnn_head = copy.deepcopy(projection)
self.ffnn_tail = copy.deepcopy(projection)
self.rel_classifier = BiaffineAttention(hidden_size // 2, 2)
self.loss_fct = CrossEntropyLoss()
def build_relation(self, relations, entities):
batch_size, max_seq_len = paddle.shape(entities)[:2]
new_relations = paddle.full(
shape=[batch_size, max_seq_len * max_seq_len, 3], fill_value=-1, dtype=relations.dtype
)
for b in range(batch_size):
if entities[b, 0, 0] <= 2:
entitie_new = paddle.full(shape=[512, 3], fill_value=-1, dtype=entities.dtype)
entitie_new[0, :] = 2
entitie_new[1:3, 0] = 0 # start
entitie_new[1:3, 1] = 1 # end
entitie_new[1:3, 2] = 0 # label
entities[b] = entitie_new
entitie_label = entities[b, 1 : entities[b, 0, 2] + 1, 2]
all_possible_relations1 = paddle.arange(0, entities[b, 0, 2], dtype=entities.dtype)
all_possible_relations1 = all_possible_relations1[entitie_label == 1]
all_possible_relations2 = paddle.arange(0, entities[b, 0, 2], dtype=entities.dtype)
all_possible_relations2 = all_possible_relations2[entitie_label == 2]
all_possible_relations = paddle.stack(
paddle.meshgrid(all_possible_relations1, all_possible_relations2), axis=2
).reshape([-1, 2])
if len(all_possible_relations) == 0:
all_possible_relations = paddle.full_like(all_possible_relations, fill_value=-1, dtype=entities.dtype)
all_possible_relations[0, 0] = 0
all_possible_relations[0, 1] = 1
relation_head = relations[b, 1 : relations[b, 0, 0] + 1, 0]
relation_tail = relations[b, 1 : relations[b, 0, 1] + 1, 1]
positive_relations = paddle.stack([relation_head, relation_tail], axis=1)
all_possible_relations_repeat = all_possible_relations.unsqueeze(axis=1).tile(
[1, len(positive_relations), 1]
)
positive_relations_repeat = positive_relations.unsqueeze(axis=0).tile([len(all_possible_relations), 1, 1])
mask = paddle.all(all_possible_relations_repeat == positive_relations_repeat, axis=2)
negative_mask = paddle.any(mask, axis=1) is False
negative_relations = all_possible_relations[negative_mask]
positive_mask = paddle.any(mask, axis=0) is True
positive_relations = positive_relations[positive_mask]
if negative_mask.sum() > 0:
reordered_relations = paddle.concat([positive_relations, negative_relations])
else:
reordered_relations = positive_relations
relation_per_doc_label = paddle.zeros([len(reordered_relations), 1], dtype=reordered_relations.dtype)
relation_per_doc_label[: len(positive_relations)] = 1
relation_per_doc = paddle.concat([reordered_relations, relation_per_doc_label], axis=1)
assert len(relation_per_doc[:, 0]) != 0
new_relations[b, 0] = paddle.shape(relation_per_doc)[0].astype(new_relations.dtype)
new_relations[b, 1 : len(relation_per_doc) + 1] = relation_per_doc
# new_relations.append(relation_per_doc)
return new_relations, entities
def get_predicted_relations(self, logits, relations, entities):
pred_relations = []
for i, pred_label in enumerate(logits.argmax(-1)):
if pred_label != 1:
continue
rel = paddle.full(shape=[7, 2], fill_value=-1, dtype=relations.dtype)
rel[0, 0] = relations[:, 0][i]
rel[1, 0] = entities[:, 0][relations[:, 0][i] + 1]
rel[1, 1] = entities[:, 1][relations[:, 0][i] + 1]
rel[2, 0] = entities[:, 2][relations[:, 0][i] + 1]
rel[3, 0] = relations[:, 1][i]
rel[4, 0] = entities[:, 0][relations[:, 1][i] + 1]
rel[4, 1] = entities[:, 1][relations[:, 1][i] + 1]
rel[5, 0] = entities[:, 2][relations[:, 1][i] + 1]
rel[6, 0] = 1
pred_relations.append(rel)
return pred_relations
def forward(self, hidden_states, entities, relations):
batch_size, max_length, _ = paddle.shape(entities)
relations, entities = self.build_relation(relations, entities)
loss = 0
all_pred_relations = paddle.full(
shape=[batch_size, max_length * max_length, 7, 2], fill_value=-1, dtype=entities.dtype
)
for b in range(batch_size):
relation = relations[b, 1 : relations[b, 0, 0] + 1]
head_entities = relation[:, 0]
tail_entities = relation[:, 1]
relation_labels = relation[:, 2]
entities_start_index = paddle.to_tensor(entities[b, 1 : entities[b, 0, 0] + 1, 0])
entities_labels = paddle.to_tensor(entities[b, 1 : entities[b, 0, 2] + 1, 2])
head_index = entities_start_index[head_entities]
head_label = entities_labels[head_entities]
head_label_repr = self.entity_emb(head_label)
tail_index = entities_start_index[tail_entities]
tail_label = entities_labels[tail_entities]
tail_label_repr = self.entity_emb(tail_label)
tmp_hidden_states = hidden_states[b][head_index]
if len(tmp_hidden_states.shape) == 1:
tmp_hidden_states = paddle.unsqueeze(tmp_hidden_states, axis=0)
head_repr = paddle.concat((tmp_hidden_states, head_label_repr), axis=-1)
tmp_hidden_states = hidden_states[b][tail_index]
if len(tmp_hidden_states.shape) == 1:
tmp_hidden_states = paddle.unsqueeze(tmp_hidden_states, axis=0)
tail_repr = paddle.concat((tmp_hidden_states, tail_label_repr), axis=-1)
heads = self.ffnn_head(head_repr)
tails = self.ffnn_tail(tail_repr)
logits = self.rel_classifier(heads, tails)
loss += self.loss_fct(logits, relation_labels)
pred_relations = self.get_predicted_relations(logits, relation, entities[b])
if len(pred_relations) > 0:
pred_relations = paddle.stack(pred_relations)
all_pred_relations[b, 0, :, :] = paddle.shape(pred_relations)[0].astype(all_pred_relations.dtype)
all_pred_relations[b, 1 : len(pred_relations) + 1, :, :] = pred_relations
return loss, all_pred_relations
[文档]class LayoutXLMForQuestionAnswering(LayoutXLMPretrainedModel):
def __init__(self, config: LayoutXLMConfig):
super(LayoutXLMForQuestionAnswering, self).__init__(config)
self.num_classes = config.num_labels
self.layoutxlm = LayoutXLMModel(config)
self.has_visual_segment_embedding = config.has_visual_segment_embedding
self.dropout = nn.Dropout(config.hidden_dropout_prob)
self.qa_outputs = nn.Linear(config.hidden_size, self.num_classes)
[文档] def forward(
self,
input_ids=None,
bbox=None,
image=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
head_mask=None,
start_positions=None,
end_positions=None,
):
# In LayoutXLM the type vocab size is 1
token_type_ids = paddle.zeros_like(input_ids)
outputs = self.layoutxlm(
input_ids=input_ids,
token_type_ids=token_type_ids,
bbox=bbox,
image=image,
attention_mask=attention_mask,
position_ids=position_ids,
head_mask=head_mask,
)
seq_length = input_ids.shape[1]
# sequence out and image out
sequence_output = outputs[0][:, :seq_length]
sequence_output = self.dropout(sequence_output)
if token_type_ids is not None:
span_mask = -token_type_ids * 1e8
else:
span_mask = 0
logits = self.qa_outputs(sequence_output)
start_logits, end_logits = paddle.split(logits, num_or_sections=2, axis=-1)
start_logits = start_logits.squeeze(-1) + span_mask
end_logits = end_logits.squeeze(-1) + span_mask
outputs = (start_logits, end_logits) + outputs[2:]
total_loss = None
if start_positions is not None and end_positions is not None:
# If we are on multi-GPU, split add a dimension
if len(start_positions.shape) > 1:
start_positions = start_positions.squeeze(-1)
if len(end_positions.shape) > 1:
end_positions = end_positions.squeeze(-1)
# Sometimes the start/end positions are outside our model inputs, we ignore these terms
ignored_index = start_logits.shape[1]
start_positions = start_positions.clip(0, ignored_index)
end_positions = end_positions.clip(0, ignored_index)
loss_fct = nn.CrossEntropyLoss(ignore_index=ignored_index)
start_loss = loss_fct(start_logits, start_positions)
end_loss = loss_fct(end_logits, end_positions)
total_loss = (start_loss + end_loss) / 2
if not total_loss:
return outputs
else:
outputs = (total_loss,) + outputs
return outputs
