-
Notifications
You must be signed in to change notification settings - Fork 6
/
finetune_on_phonemebert.py
283 lines (252 loc) · 10.4 KB
/
finetune_on_phonemebert.py
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
import json
import os
import torch
import torch.nn as nn
from datasets import load_metric
from transformers import DataCollatorWithPadding
from transformers import AutoTokenizer, AutoModel
from transformers import Trainer, TrainingArguments, TrainerCallback
from transformers.trainer_callback import EarlyStoppingCallback
import numpy as np
import argparse
from dataset import PhonemeBERTDataset, separate_phonemebert_test_set
from collator import DataCollatorWithPaddingMLM
from loss_functions import SupervisedContrastiveLoss, KLWithSoftLabelLoss
metric = load_metric("accuracy")
def compute_metrics(eval_pred):
logits, labels = eval_pred
predictions = np.argmax(logits, axis=-1)
metrics = metric.compute(
predictions=predictions,
references=labels
)
return {
"accuracy": metrics["accuracy"],
}
class Net(torch.nn.Module):
def __init__(self, args, num_labels):
super(Net, self).__init__()
self.args = args
self.bert = AutoModel.from_pretrained(args.model_name_or_path)
self.bert.config.type_vocab_size = 2
self.bert.embeddings.token_type_embeddings = nn.Embedding(2, self.bert.config.hidden_size)
self.bert.embeddings.token_type_embeddings.weight.data.normal_(mean=0.0, std=self.bert.config.initializer_range)
self.bert.config.hidden_dropout_prob = args.dropout
self.mlp = nn.Linear(self.bert.config.hidden_size, self.bert.config.hidden_size)
self.clf_head = nn.Linear(self.bert.config.hidden_size, num_labels)
def forward(self, **inputs):
label = inputs.pop('labels')
pseudo_label = inputs.pop('pseudo_label')
bert_output = self.bert(**inputs)
# last_hidden = torch.mean(bert_output.last_hidden_state, dim=1)
last_hidden = bert_output.last_hidden_state[:, 0]
logits = self.clf_head(self.mlp(last_hidden))
""" Calculate Loss """
ce_loss_fn = nn.CrossEntropyLoss(reduction='mean')
loss = ce_loss_fn(logits, label)
if self.args.use_pseudo:
kd_loss_fn = KLWithSoftLabelLoss(self.args.pseudo_label_temperature, self.args.pseudo_weight)
pseudo_loss = kd_loss_fn(logits, pseudo_label)
loss += pseudo_loss
if self.args.use_contrastive:
contrastive_loss_fn = SupervisedContrastiveLoss(temperature=self.args.contrastive_temperature)
if self.args.use_pseudo:
loss += self.args.contrastive_weight * contrastive_loss_fn(
last_hidden, label, soft_labels=pseudo_label)
else:
loss += self.args.contrastive_weight * contrastive_loss_fn(last_hidden, label)
return loss, logits
class ContrastiveTrainer(Trainer):
def compute_loss(self, model, inputs, return_outputs=False):
loss, logits = model(**inputs)
outputs = {'logits': logits}
return loss if not return_outputs else (loss, outputs)
def save_prediction(args, pred, test_dataset):
logits, labels = pred[0], pred[1]
predictions = np.argmax(logits, axis=-1)
with open(os.path.join(args.output_dir, "output.npy"), 'wb') as f:
output = [predictions, labels]
np.save(f, output)
with open(os.path.join(args.output_dir, "test_data.json"), "w") as f:
test_data = {
"text": test_dataset.text,
"phoneme_text": test_dataset.phoneme_text,
"golden": test_dataset.golden,
"id2label": test_dataset.id2label,
# "golden_phoneme": test_dataset.golden_phoneme,
}
json.dump(test_data, f)
class UpdatePseudoLabelCallback(TrainerCallback):
def __init__(self, trainer, warmup=0) -> None:
super().__init__()
self._trainer = trainer
self.warmup = warmup
def on_epoch_end(self, args, state, control, **kwargs):
pred = self._trainer.predict(test_dataset=self._trainer.train_dataset)
print("\ntrain metric: ", pred[2])
if state.epoch > self.warmup:
percent = max(5 * state.epoch, 30)
self._trainer.train_dataset.update_pseudo_label(pred, 100-percent, verbose=True)
def parse_args():
parser = argparse.ArgumentParser()
parser.add_argument(
"--dataset_dir",
default='../Phoneme-BERT/phomeme-bert-data/downstream-datasets/atis',
type=str, help="dataset directory"
)
parser.add_argument(
"--model_name_or_path", default='roberta-base', type=str, help="model to finetune"
)
parser.add_argument(
"--tokenizer_name", default='roberta-base', type=str, help="model to finetune"
)
parser.add_argument(
"--output_dir", default='runs/finetune', type=str, help="dir to save finetuned model"
)
parser.add_argument(
"--log_dir", default='logs/', type=str, help="dir to save finetuned model"
)
parser.add_argument(
"--log_name", default='finetune', type=str, help="dir to save finetuned model"
)
parser.add_argument(
"--seed", default=42, type=int, help="seed"
)
parser.add_argument(
"-n", default=5, type=int, help="num to run & average"
)
parser.add_argument(
"--max_epoch", default=20, type=int, help="total number of epoch"
)
parser.add_argument(
"--train_bsize", default=64, type=int, help="training batch size"
)
parser.add_argument(
"--eval_bsize", default=64, type=int, help="evaluation batch size"
)
parser.add_argument(
"--patience", default=3, type=int, help="early stopping patience"
)
parser.add_argument(
"--train_golden", action='store_true', help="train on golden transcript"
)
parser.add_argument(
"--eval_golden", action='store_true', help="eval on golden transcript"
)
parser.add_argument(
"--use_phoneme", action='store_true', help="use phoneme + text sequence"
)
parser.add_argument(
"--input_mask_ratio", default=0, type=float, help="mlm ratio when training"
)
parser.add_argument(
"--dropout", default=0.1, type=float, help="model hidden dropout"
)
parser.add_argument(
"--save_predict", action='store_true', help="save prediction & test text"
)
parser.add_argument(
"--use_contrastive", action='store_true', help="supervised contrastive objective"
)
parser.add_argument(
"--contrastive_temperature", default=0.2, type=float, help="contrastive temperature"
)
parser.add_argument(
"--contrastive_weight", default=0.1, type=float, help="contrastive loss weight vs classification"
)
parser.add_argument(
"--use_pseudo", action='store_true', help="train from pseudo label"
)
parser.add_argument(
"--pseudo_label_temperature", default=5, type=float, help="contrastive temperature"
)
parser.add_argument(
"--pseudo_weight", default=10, type=float, help="contrastive loss weight vs classification"
)
args = parser.parse_args()
return args
if __name__ == "__main__":
args = parse_args()
# Dataset
print('reading dataset')
tokenizer = AutoTokenizer.from_pretrained(args.tokenizer_name)
train_dir = os.path.join(args.dataset_dir, 'train')
eval_dir = os.path.join(args.dataset_dir, 'valid')
test_dir = os.path.join(args.dataset_dir, 'test')
train_dataset = PhonemeBERTDataset(
tokenizer, train_dir, use_golden=args.train_golden,
use_phoneme=args.use_phoneme
)
eval_dataset = PhonemeBERTDataset(
tokenizer, eval_dir, use_golden=args.eval_golden,
use_phoneme=args.use_phoneme,
id2label=train_dataset.id2label
)
test_dataset = PhonemeBERTDataset(
tokenizer, test_dir, use_golden=args.eval_golden,
use_phoneme=args.use_phoneme,
id2label=train_dataset.id2label
)
test_datasets = [test_dataset] + separate_phonemebert_test_set(test_dataset)
data_collator = DataCollatorWithPaddingMLM(
tokenizer=tokenizer,
mlm=args.input_mask_ratio > 0,
mlm_probability=args.input_mask_ratio
)
all_preds = []
for n in range(args.n):
print('start training: {}'.format(n))
model = Net(args, len(train_dataset.id2label))
# Train model
training_args = TrainingArguments(
output_dir=args.output_dir,
overwrite_output_dir=True,
evaluation_strategy="epoch",
logging_strategy="epoch",
save_strategy="epoch",
load_best_model_at_end=True,
metric_for_best_model="eval_accuracy",
num_train_epochs=args.max_epoch,
per_device_train_batch_size=args.train_bsize,
per_device_eval_batch_size=args.eval_bsize,
weight_decay=0.01, # strength of weight decay
seed=args.seed + n,
)
trainer = ContrastiveTrainer(
model=model,
args=training_args,
data_collator=data_collator,
tokenizer=tokenizer,
train_dataset=train_dataset,
eval_dataset=eval_dataset,
compute_metrics=compute_metrics,
callbacks=[EarlyStoppingCallback(early_stopping_patience=args.patience)],
)
if args.use_pseudo:
trainer.add_callback(UpdatePseudoLabelCallback(trainer))
trainer.train()
test_data_collator = DataCollatorWithPadding(tokenizer=tokenizer)
trainer.data_collator = test_data_collator
keys = ['test_accuracy']
preds = []
for i, test_dataset in enumerate(test_datasets):
pred = trainer.predict(test_dataset=test_dataset)
pred = {k: pred[2][k] for k in keys}
preds.append(pred)
all_preds.append(preds)
predictions = {}
for preds in all_preds:
for i, pred in enumerate(preds):
for k, v in pred.items():
key = k+'-{}'.format(i)
predictions[key] = predictions.get(key, []) + [np.round(v, 4)]
os.makedirs(args.log_dir, exist_ok=True)
logfile = os.path.join(args.log_dir, "{}.log".format(args.log_name))
with open(logfile, 'w') as f:
f.write("{:>30}\t{:>8}\t{:>8}\t{}\n".format('metric', 'mean', 'std', 'values'))
print("\n{:>30}\t{:>8}\t{:>8}\t{}".format('metric', 'mean', 'std', 'values'))
for k, v in predictions.items():
mean = np.round(np.mean(v), 4)
std = np.round(np.std(v), 4)
print("{:>30}\t{:>8}\t{:>8}\t{}".format(k, mean, std, v))
f.write("{:>30}\t{:>8}\t{:>8}\t{}\n".format(k, mean, std, v))