MultiTarget.py
6.15 KB
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
import logging
logger = logging.getLogger(__name__)
import tensorflow as tf
from transformers.modeling_tf_utils import get_initializer, input_processing, TFTokenClassificationLoss
from transformers.modeling_tf_outputs import TFTokenClassifierOutput
class TFMultiTargetTokenClassification(TFTokenClassificationLoss):
_keys_to_ignore_on_load_missing = [
r'dropout',
r'dot'
]
lm_layer_input_keys = (
'input_ids',
'attention_mask',
'token_type_ids',
'position_ids',
'head_mask',
'inputs_embeds',
'output_attentions',
'output_hidden_states',
'return_dict',
'training',
)
def get_lm_layer(self, config):
lm_layer_name = config.model_type if self.lm_layer_name is None else self.lm_layer_name
return self.lm_layer_class(config, name=lm_layer_name, **self.lm_layer_init_kwargs)
def __init__(self, config, *inputs, **kwargs):
logger.debug(f'config.model_type: {config.model_type}')
logger.debug(f'self.config_class: {self.config_class}')
categories = kwargs.pop('categories')
labels = kwargs.pop('labels')
super().__init__(config, *inputs, **kwargs)
self.categories, self.categories2d = [], []
for cat in categories:
if cat in labels:
self.categories.append(cat)
else:
self.categories2d.append(cat)
self.lm_layer = self.get_lm_layer(config)
logger.debug('self.lm_layer: {self.lm_layer}')
classifier_dropout = (
config.classifier_dropout if config.classifier_dropout is not None else config.hidden_dropout_prob
)
self.dropout = tf.keras.layers.Dropout(rate=classifier_dropout)
initializer_kwargs = {}
if getattr(config, 'initializer_range', None) is not None:
initializer_kwargs['initializer_range'] = config.initializer_range
self.classifiers = [
tf.keras.layers.Dense(
units=len(labels[cat]),
kernel_initializer=get_initializer(**initializer_kwargs),
name=f'classifier_{cat}',
) for cat in self.categories
]
logger.info(f'created {len(self.classifiers)} classifier(s): {", ".join(self.categories)}')
self.mappings2d = [
{
'head': tf.keras.layers.Dense(
units=128,
kernel_initializer=get_initializer(**initializer_kwargs),
name=f'head_mapping_{cat}'
),
'dependent' : tf.keras.layers.Dense(
units=128,
kernel_initializer=get_initializer(**initializer_kwargs),
name=f'dependent_mapping_{cat}',
)
} for cat in self.categories2d
]
logger.info(f'created {len(self.mappings2d)} head/dependent mapping(s): {", ".join(self.categories2d)}')
self.dot = tf.keras.layers.Dot(axes=(-1))
def call(
self,
input_ids=None,
attention_mask=None,
token_type_ids=None,
position_ids=None,
head_mask=None,
inputs_embeds=None,
output_attentions=None,
output_hidden_states=None,
return_dict=None,
labels=None,
training=False,
**kwargs,
):
inputs = input_processing(
func=self.call,
config=self.config,
input_ids=input_ids,
attention_mask=attention_mask,
token_type_ids=token_type_ids,
position_ids=position_ids,
head_mask=head_mask,
inputs_embeds=inputs_embeds,
output_attentions=output_attentions,
output_hidden_states=output_hidden_states,
return_dict=return_dict,
labels=labels,
training=training,
kwargs_call=kwargs,
)
kwargs = { key : inputs[key] for key in self.lm_layer_input_keys }
outputs = self.lm_layer(
#input_ids=inputs['input_ids'],
#attention_mask=inputs['attention_mask'],
#token_type_ids=inputs['token_type_ids'],
#position_ids=inputs['position_ids'],
#head_mask=inputs['head_mask'],
#inputs_embeds=inputs['inputs_embeds'],
#output_attentions=inputs['output_attentions'],
#output_hidden_states=inputs['output_hidden_states'],
#return_dict=inputs['return_dict'],
#training=inputs['training'],
**kwargs,
)
sequence_output = outputs[0]
sequence_output = self.dropout(inputs=sequence_output, training=inputs['training'])
logits = [classifier(inputs=sequence_output) for classifier in self.classifiers]
logits2d = []
for mapping in self.mappings2d:
heads = mapping['head'](inputs=sequence_output)
dependents = mapping['dependent'](inputs=sequence_output)
product = self.dot([heads, dependents])
logits2d.append(product)
keys = self.categories + self.categories2d
if len(keys) > 1:
return dict(zip(keys, logits + logits2d))
else:
logits = logits[0] if logits else logits2d[0]
return TFTokenClassifierOutput(logits=logits)
def serving_output(self, output):
return output
def build(self, input_shape=None):
if self.built:
return
self.built = True
if getattr(self, 'lm_layer', None) is not None:
with tf.name_scope(self.lm_layer.name):
self.lm_layer.build(None)
if getattr(self, 'classifiers', None) is not None:
for classifier in self.classifiers:
with tf.name_scope(classifier.name):
classifier.build([None, None, self.config.hidden_size])
if getattr(self, 'mappings2d', None) is not None:
for mapping2d in self.mappings2d:
for key in ['head', 'dependent']:
with tf.name_scope(mapping2d[key].name):
mapping2d[key].build([None, None, self.config.hidden_size])