File size: 12,116 Bytes
2d06dcc |
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 |
import numpy as np
import torch
import os
import csv
import sys
import logging
from transformers import BertTokenizer
from torch.utils.data import (DataLoader, RandomSampler, SequentialSampler, TensorDataset)
class BERT_Loader:
def __init__(self, args, base_attrs, logger_name = 'Detection'):
self.logger = logging.getLogger(logger_name)
self.train_examples, self.train_labeled_examples, self.train_unlabeled_examples = get_examples(args, base_attrs, 'train')
self.logger.info("Number of labeled training samples = %s", str(len(self.train_labeled_examples)))
self.logger.info("Number of unlabeled training samples = %s", str(len(self.train_unlabeled_examples)))
self.eval_examples = get_examples(args, base_attrs, 'eval')
self.logger.info("Number of evaluation samples = %s", str(len(self.eval_examples)))
self.test_examples = get_examples(args, base_attrs, 'test')
self.logger.info("Number of testing samples = %s", str(len(self.test_examples)))
self.base_attrs = base_attrs
self.init_loader(args)
def init_loader(self, args):
self.train_labeled_loader = get_loader(self.train_labeled_examples, args, self.base_attrs['label_list'], 'train_labeled', sampler_mode = 'random')
self.train_unlabeled_loader = get_loader(self.train_unlabeled_examples, args, self.base_attrs['label_list'], 'train_unlabeled', sampler_mode = 'sequential')
self.eval_loader = get_loader(self.eval_examples, args, self.base_attrs['label_list'], 'eval', sampler_mode = 'sequential')
self.test_loader = get_loader(self.test_examples, args, self.base_attrs['label_list'], 'test', sampler_mode = 'sequential')
self.num_train_examples = len(self.train_labeled_examples)
def get_examples(args, base_attrs, mode):
processor = DatasetProcessor()
ori_examples = processor.get_examples(base_attrs['data_dir'], mode)
if mode == 'train':
labeled_examples, unlabeled_examples = [], []
for example in ori_examples:
if (example.label in base_attrs['known_label_list']) and (np.random.uniform(0, 1) <= args.labeled_ratio):
labeled_examples.append(example)
else:
example.label = base_attrs['unseen_label']
unlabeled_examples.append(example)
return ori_examples, labeled_examples, unlabeled_examples
elif mode == 'eval':
examples = []
for example in ori_examples:
if (example.label in base_attrs['known_label_list']):
examples.append(example)
return examples
elif mode == 'test':
examples = []
for example in ori_examples:
if (example.label in base_attrs['label_list']) and (example.label != base_attrs['unseen_label']):
examples.append(example)
else:
example.label = base_attrs['unseen_label']
examples.append(example)
return examples
def get_loader(examples, args, label_list, mode, sampler_mode = 'sequential'):
tokenizer = BertTokenizer.from_pretrained(args.bert_model, do_lower_case=True)
features = convert_examples_to_features(examples, label_list, args.max_seq_length, tokenizer)
input_ids = torch.tensor([f.input_ids for f in features], dtype=torch.long)
input_mask = torch.tensor([f.input_mask for f in features], dtype=torch.long)
segment_ids = torch.tensor([f.segment_ids for f in features], dtype=torch.long)
if mode == 'train_unlabeled':
label_ids = torch.tensor([-1 for f in features], dtype=torch.long)
else:
label_ids = torch.tensor([f.label_id for f in features], dtype=torch.long)
datatensor = TensorDataset(input_ids, input_mask, segment_ids, label_ids)
if sampler_mode == 'random':
sampler = RandomSampler(datatensor)
elif sampler_mode == 'sequential':
sampler = SequentialSampler(datatensor)
if mode == 'train_labeled':
dataloader = DataLoader(datatensor, sampler = sampler, batch_size = args.train_batch_size)
else:
if mode == 'train_unlabeled':
dataloader = DataLoader(datatensor, sampler=sampler, batch_size = args.train_batch_size)
elif mode == 'eval':
dataloader = DataLoader(datatensor, sampler=sampler, batch_size = args.eval_batch_size)
elif mode == 'test':
dataloader = DataLoader(datatensor, sampler=sampler, batch_size = args.test_batch_size)
return dataloader
class InputExample(object):
"""A single training/test example for simple sequence classification."""
def __init__(self, guid, text_a, text_b=None, label=None):
"""Constructs a InputExample.
Args:
guid: Unique id for the example.
text_a: string. The untokenized text of the first sequence. For single
sequence tasks, only this sequence must be specified.
text_b: (Optional) string. The untokenized text of the second sequence.
Only must be specified for sequence pair tasks.
label: (Optional) string. The label of the example. This should be
specified for train and dev examples, but not for test examples.
"""
self.guid = guid
self.text_a = text_a
self.text_b = text_b
self.label = label
class InputFeatures(object):
"""A single set of features of data."""
def __init__(self, input_ids, input_mask, segment_ids, label_id):
self.input_ids = input_ids
self.input_mask = input_mask
self.segment_ids = segment_ids
self.label_id = label_id
class DataProcessor(object):
"""Base class for data converters for sequence classification data sets."""
def get_train_examples(self, data_dir):
"""Gets a collection of `InputExample`s for the train set."""
raise NotImplementedError()
def get_dev_examples(self, data_dir):
"""Gets a collection of `InputExample`s for the dev set."""
raise NotImplementedError()
def get_labels(self):
"""Gets the list of labels for this data set."""
raise NotImplementedError()
@classmethod
def _read_tsv(cls, input_file, quotechar=None):
"""Reads a tab separated value file."""
with open(input_file, "r") as f:
reader = csv.reader(f, delimiter="\t", quotechar=quotechar)
lines = []
for line in reader:
if sys.version_info[0] == 2:
line = list(unicode(cell, 'utf-8') for cell in line)
lines.append(line)
return lines
class DatasetProcessor(DataProcessor):
def get_examples(self, data_dir, mode):
if mode == 'train':
return self._create_examples(
self._read_tsv(os.path.join(data_dir, "train.tsv")), "train")
elif mode == 'eval':
return self._create_examples(
self._read_tsv(os.path.join(data_dir, "dev.tsv")), "train")
elif mode == 'test':
return self._create_examples(
self._read_tsv(os.path.join(data_dir, "test.tsv")), "test")
def _create_examples(self, lines, set_type):
"""Creates examples for the training and dev sets."""
examples = []
for (i, line) in enumerate(lines):
if i == 0:
continue
if len(line) != 2:
continue
guid = "%s-%s" % (set_type, i)
text_a = line[0]
label = line[1]
examples.append(
InputExample(guid=guid, text_a=text_a, text_b=None, label=label))
return examples
def convert_examples_to_features(examples, label_list, max_seq_length, tokenizer):
"""Loads a data file into a list of `InputBatch`s."""
label_map = {}
if label_list is not None:
for i, label in enumerate(label_list):
label_map[label] = i
features = []
for (ex_index, example) in enumerate(examples):
tokens_a = tokenizer.tokenize(example.text_a)
tokens_b = None
if example.text_b:
tokens_b = tokenizer.tokenize(example.text_b)
# Modifies `tokens_a` and `tokens_b` in place so that the total
# length is less than the specified length.
# Account for [CLS], [SEP], [SEP] with "- 3"
_truncate_seq_pair(tokens_a, tokens_b, max_seq_length - 3)
else:
# Account for [CLS] and [SEP] with "- 2"
if len(tokens_a) > max_seq_length - 2:
tokens_a = tokens_a[:(max_seq_length - 2)]
# The convention in BERT is:
# (a) For sequence pairs:
# tokens: [CLS] is this jack ##son ##ville ? [SEP] no it is not . [SEP]
# type_ids: 0 0 0 0 0 0 0 0 1 1 1 1 1 1
# (b) For single sequences:
# tokens: [CLS] the dog is hairy . [SEP]
# type_ids: 0 0 0 0 0 0 0
#
# Where "type_ids" are used to indicate whether this is the first
# sequence or the second sequence. The embedding vectors for `type=0` and
# `type=1` were learned during pre-training and are added to the wordpiece
# embedding vector (and position vector). This is not *strictly* necessary
# since the [SEP] token unambigiously separates the sequences, but it makes
# it easier for the model to learn the concept of sequences.
#
# For classification tasks, the first vector (corresponding to [CLS]) is
# used as as the "sentence vector". Note that this only makes sense because
# the entire model is fine-tuned.
tokens = ["[CLS]"] + tokens_a + ["[SEP]"]
segment_ids = [0] * len(tokens)
if tokens_b:
tokens += tokens_b + ["[SEP]"]
segment_ids += [1] * (len(tokens_b) + 1)
input_ids = tokenizer.convert_tokens_to_ids(tokens)
# The mask has 1 for real tokens and 0 for padding tokens. Only real
# tokens are attended to.
input_mask = [1] * len(input_ids)
# Zero-pad up to the sequence length.
padding = [0] * (max_seq_length - len(input_ids))
input_ids += padding
input_mask += padding
segment_ids += padding
assert len(input_ids) == max_seq_length
assert len(input_mask) == max_seq_length
assert len(segment_ids) == max_seq_length
label_id = label_map[example.label] if label_list is not None else None
# if ex_index < 5:
# logger.info("*** Example ***")
# logger.info("guid: %s" % (example.guid))
# logger.info("tokens: %s" % " ".join(
# [str(x) for x in tokens]))
# logger.info("input_ids: %s" % " ".join([str(x) for x in input_ids]))
# logger.info("input_mask: %s" % " ".join([str(x) for x in input_mask]))
# logger.info(
# "segment_ids: %s" % " ".join([str(x) for x in segment_ids]))
# logger.info("label: %s (id = %d)" % (example.label, label_id))
features.append(
InputFeatures(input_ids=input_ids,
input_mask=input_mask,
segment_ids=segment_ids,
label_id=label_id))
return features
def _truncate_seq_pair(tokens_a, tokens_b, max_length):
"""Truncates a sequence pair in place to the maximum length."""
# This is a simple heuristic which will always truncate the longer sequence
# one token at a time. This makes more sense than truncating an equal percent
# of tokens from each, since if one sequence is very short then each token
# that's truncated likely contains more information than a longer sequence.
while True:
total_length = len(tokens_a) + len(tokens_b)
if total_length <= max_length:
break
if len(tokens_a) > len(tokens_b):
tokens_a.pop(0) # For dialogue context
else:
tokens_b.pop()
|