-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathrun_cascade.py
393 lines (344 loc) · 16.9 KB
/
run_cascade.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
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
import logging
import os
import random
import sys
from dataclasses import dataclass, field
from typing import Optional
import numpy as np
from datasets import load_dataset, load_metric
import transformers
from transformers import (
AutoConfig,
AutoTokenizer,
EvalPrediction,
HfArgumentParser,
PretrainedConfig,
TrainingArguments,
default_data_collator,
set_seed, BertModel, AdamW,
)
from cascade_bert import CascadeBERTForSequenceClassification
# from transformers.trainer_utils import is_main_process
from trainer import CascadeBERTTrainer
from scipy.special import softmax
task_to_keys = {
"cola": ("sentence", None),
"mnli": ("premise", "hypothesis"),
"mrpc": ("sentence1", "sentence2"),
"qnli": ("question", "sentence"),
"qqp": ("question1", "question2"),
"rte": ("sentence1", "sentence2"),
"sst2": ("sentence", None),
"stsb": ("sentence1", "sentence2"),
"wnli": ("sentence1", "sentence2"),
}
logger = logging.getLogger(__name__)
@dataclass
class DataTrainingArguments:
"""
Arguments pertaining to what data we are going to input our model for training and eval.
Using `HfArgumentParser` we can turn this class
into argparse arguments to be able to specify them on
the command line.
"""
task_name: Optional[str] = field(
default=None,
metadata={"help": "The name of the task to train on: " + ", ".join(task_to_keys.keys())},
)
max_seq_length: int = field(
default=128,
metadata={
"help": "The maximum total input sequence length after tokenization. Sequences longer "
"than this will be truncated, sequences shorter will be padded."
},
)
overwrite_cache: bool = field(
default=False, metadata={"help": "Overwrite the cached preprocessed datasets or not."}
)
pad_to_max_length: bool = field(
default=True,
metadata={
"help": "Whether to pad all samples to `max_seq_length`. "
"If False, will pad the samples dynamically when batching to the maximum length in the batch."
},
)
train_file: Optional[str] = field(
default=None, metadata={"help": "A csv or a json file containing the training data."}
)
validation_file: Optional[str] = field(
default=None, metadata={"help": "A csv or a json file containing the validation data."}
)
def __post_init__(self):
if self.task_name is not None:
self.task_name = self.task_name.lower()
if self.task_name not in task_to_keys.keys():
raise ValueError("Unknown task, you should pick one in " + ",".join(task_to_keys.keys()))
elif self.train_file is None or self.validation_file is None:
raise ValueError("Need either a GLUE task or a training/validation file.")
else:
extension = self.train_file.split(".")[-1]
assert extension in ["csv", "json", "tsv"], "`train_file` should be a csv or a json file."
extension = self.validation_file.split(".")[-1]
assert extension in ["csv", "json"], "`validation_file` should be a csv or a json file."
@dataclass
class ModelArguments:
"""
Arguments pertaining to which model/config/tokenizer we are going to fine-tune from.
"""
cascade_model_name_or_path: str = field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models, split by ;"}
)
cascade_model_layers: str = field(
metadata={"help": "layer number of each cascading model , split by ;, e.g., 2;12"}
)
config_name: Optional[str] = field(
default=None, metadata={"help": "Pretrained config name or path if not the same as model_name"}
)
tokenizer_name: Optional[str] = field(
default=None, metadata={"help": "Pretrained tokenizer name or path if not the same as model_name"}
)
cache_dir: Optional[str] = field(
default=None, metadata={"help": "Where do you want to store the pretrained models downloaded from s3"}
)
use_fast_tokenizer: bool = field(
default=True,
metadata={"help": "Whether to use one of the fast tokenizer (backed by the tokenizers library) or not."},
)
saved_model_path: Optional[str] = field(
default="", metadata={"help": "Trained model path"}
)
infer_mode: Optional[str] = field(
default="small", metadata={"help": "infer mode: big, small, cascade"}
)
confidence_threshold: Optional[float] = field(
default=1.0, metadata={"help": "threshold for judging the easy examples"}
)
confidence_margin: Optional[float] = field(
default=0.3, metadata={"help": "confidence margin"}
)
dar_weight: Optional[float] = field(
default=0.5, metadata={"help": "difficulty-aware regularization weight"}
)
def main():
# See all possible arguments in src/transformers/training_args.py
# or by passing the --help flag to this script.
# We now keep distinct sets of args, for a cleaner separation of concerns.
parser = HfArgumentParser((ModelArguments, DataTrainingArguments, TrainingArguments))
if len(sys.argv) == 2 and sys.argv[1].endswith(".json"):
# If we pass only one argument to the script and it's the path to a json file,
# let's parse it to get our arguments.
model_args, data_args, training_args = parser.parse_json_file(json_file=os.path.abspath(sys.argv[1]))
else:
model_args, data_args, training_args = parser.parse_args_into_dataclasses()
if (
os.path.exists(training_args.output_dir)
and os.listdir(training_args.output_dir)
and training_args.do_train
and not training_args.overwrite_output_dir
):
raise ValueError(
f"Output directory ({training_args.output_dir}) already exists and is not empty. "
"Use --overwrite_output_dir to overcome."
)
# Setup logging
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
datefmt="%m/%d/%Y %H:%M:%S",
level=logging.INFO, # if logging.WARN # is_main_process(training_args.local_rank) else logging.WARN,
)
# Log on each process the small summary:
logger.warning(
f"Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu}"
+ f"distributed training: {bool(training_args.local_rank != -1)}, 16-bits training: {training_args.fp16}"
)
# Set the verbosity to info of the Transformers logger (on main process only):
#if is_main_process(training_args.local_rank):
# transformers.utils.logging.set_verbosity_info()
logger.info(f"Training/evaluation parameters {training_args}")
# Set seed before initializing model.
set_seed(training_args.seed)
eval_and_test_datasets = load_dataset("glue", data_args.task_name)
logger.info("Loading train data from %s" % data_args.train_file)
difficulty_train_datasets = load_dataset(
"json", data_files={"train": data_args.train_file})
# Labels
label_list = eval_and_test_datasets["train"].features["label"].names
num_labels = len(label_list)
# load complete models
cascade_model_names = model_args.cascade_model_name_or_path.split(";")
cascade_model_layers = [int(k) for k in model_args.cascade_model_layers.split(";")]
config = AutoConfig.from_pretrained(
model_args.config_name if model_args.config_name else cascade_model_names[-1],
num_labels=num_labels,
finetuning_task=data_args.task_name,
cache_dir=model_args.cache_dir,
)
tokenizer = AutoTokenizer.from_pretrained(
model_args.tokenizer_name if model_args.tokenizer_name else cascade_model_names[-1],
cache_dir=model_args.cache_dir,
use_fast=model_args.use_fast_tokenizer,
)
for i in range(1, len(cascade_model_layers)):
cascade_model_layers[i] += cascade_model_layers[i - 1]
print('Actual layer cost:', cascade_model_layers)
cascade_models = [BertModel.from_pretrained(
model_path,
from_tf=bool(".ckpt" in cascade_model_names[-1]),
config=config,
cache_dir=model_args.cache_dir,
) for model_path in cascade_model_names]
logger.info("Setting confidence margin to %.3f" % model_args.confidence_margin)
logger.info("Setting margin loss weight to %.3f" % model_args.dar_weight)
if model_args.saved_model_path != "":
model = CascadeBERTForSequenceClassification.from_pretrained(
model_args.saved_model_path,
from_tf=bool(".ckpt" in model_args.saved_model_path),
config=config,
cache_dir=model_args.cache_dir,
cascade_models=cascade_models,
confidence_margin=model_args.confidence_margin,
margin_loss_weight=model_args.dar_weight
)
else:
model = CascadeBERTForSequenceClassification(config,
cascade_models=cascade_models,
confidence_margin=model_args.confidence_margin,
margin_loss_weight=model_args.dar_weight
)
# setting up some configs
logger.info("Setting infer mode to: %s" % model_args.infer_mode)
model.set_infer_mode(model_args.infer_mode)
logger.info("Setting infer mode to: %.6f" % model_args.confidence_threshold)
model.set_threshold(model_args.confidence_threshold)
# Preprocessing the datasets
sentence1_key, sentence2_key = task_to_keys[data_args.task_name]
# Padding strategy
if data_args.pad_to_max_length:
padding = "max_length"
max_length = data_args.max_seq_length
else:
# We will pad later, dynamically at batch creation, to the max sequence length in each batch
padding = False
max_length = None
# Some models have set the order of the labels to use, so let's make sure we do use it.
label_to_id = {v: i for i, v in enumerate(label_list)}
# Formation Rule
difficulty_list = ["easy", "hard"]
difficulty_to_id = None
difficulty_to_id = {v: i for i, v in enumerate(difficulty_list)}
print(label_to_id)
def preprocess_function(examples):
# Tokenize the texts
args = (
(examples[sentence1_key],) if sentence2_key is None else (examples[sentence1_key], examples[sentence2_key])
)
result = tokenizer(*args, padding=padding, max_length=max_length, truncation=True)
if label_to_id is not None and "label" in examples:
result["label"] = [label_to_id[l] for l in examples["label"]]
if difficulty_to_id is not None and "difficulty" in examples:
result["difficulty_labels"] = [difficulty_to_id[d] for d in examples["difficulty"]]
return result
difficulty_train_datasets = difficulty_train_datasets.map(preprocess_function, batched=True,
load_from_cache_file=not data_args.overwrite_cache)
label_to_id = None # mimic original script
eval_and_test_datasets = eval_and_test_datasets.map(preprocess_function, batched=True,
load_from_cache_file=not data_args.overwrite_cache)
train_dataset = difficulty_train_datasets["train"]
eval_dataset = eval_and_test_datasets["validation_matched" if data_args.task_name == "mnli" else "validation"]
test_dataset = eval_and_test_datasets["test_matched" if data_args.task_name == "mnli" else "test"]
# Log a few random samples from the training set:
for index in random.sample(range(len(train_dataset)), 3):
logger.info(f"Sample {index} of the training set: {train_dataset[index]}.")
# Get the metric function
assert data_args.task_name is not None, "task name cannot be None"
metric = load_metric("glue", data_args.task_name)
# You can define your custom compute_metrics function. It takes an `EvalPrediction` object (a namedtuple with a
# predictions and label_ids field) and has to return a dictionary string to float.
def compute_metrics(p: EvalPrediction):
preds = p.predictions[0] if isinstance(p.predictions, tuple) else p.predictions
preds = np.argmax(preds, axis=1)
result = metric.compute(predictions=preds, references=p.label_ids)
if len(result) > 1:
result["combined_score"] = np.mean(list(result.values())).item()
return result
# Initialize our Trainer
trainer = CascadeBERTTrainer(
model=model,
args=training_args,
train_dataset=train_dataset,
eval_dataset=eval_dataset if training_args.do_eval else None,
compute_metrics=compute_metrics,
tokenizer=tokenizer,
# Data collator will default to DataCollatorWithPadding, so we change it if we already did the padding.
data_collator=default_data_collator if data_args.pad_to_max_length else None,
require_exit_distribution=True if model_args.infer_mode in ["cascade"] else False,
model_layer_num=[2, 14] if model_args.infer_mode != "proxy" else [2, 12],
)
# Training
if training_args.do_train:
trainer.train(
model_path=model_args.saevd_path if os.path.isdir(model_args.saved_model_path) else None
)
trainer.save_model() # Saves the tokenizer too for easy upload
# Evaluation
eval_results = {}
if training_args.do_eval:
logger.info("*** Evaluate ***")
# Loop to handle MNLI double evaluation (matched, mis-matched)
tasks = [data_args.task_name]
eval_datasets = [eval_dataset]
if data_args.task_name == "mnli":
tasks.append("mnli-mm")
eval_datasets.append(eval_and_test_datasets["validation_mismatched"])
for eval_dataset, task in zip(eval_datasets, tasks):
eval_result = trainer.evaluate(eval_dataset=eval_dataset)
speed_up = eval_result['eval_expected_acceleration'] if 'eval_expected_acceleration' in eval_result else 1.0
output_eval_file = os.path.join(training_args.output_dir, "eval_results_%s_TH%.6f_spd%.2f.txt" % (task,
model_args.confidence_threshold,
speed_up ))
if trainer.is_world_process_zero():
with open(output_eval_file, "w") as writer:
logger.info(f"***** Eval results {task} *****")
for key, value in eval_result.items():
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")
eval_prediction = trainer.predict(test_dataset=eval_dataset)
output_prob_file = os.path.join(training_args.output_dir, f"eval_prob_{task}.npy")
output_label_file = os.path.join(training_args.output_dir, f"eval_label_{task}.npy")
logits = eval_prediction.predictions
prob = softmax(logits, axis=-1)
label = eval_prediction.label_ids
# np.save(output_label_file, label)
# np.save(output_prob_file, prob)
eval_results.update(eval_result)
if training_args.do_predict:
logger.info("*** Test ***")
# Loop to handle MNLI double evaluation (matched, mis-matched)
tasks = [data_args.task_name]
test_datasets = [test_dataset]
if data_args.task_name == "mnli":
tasks.append("mnli-mm")
test_datasets.append(eval_and_test_datasets["test_mismatched"])
for test_dataset, task in zip(test_datasets, tasks):
# Removing the `label` columns because it contains -1 and Trainer won't like that.
test_dataset.remove_columns_("label")
prediction_ret = trainer.predict(test_dataset=test_dataset)
predictions = prediction_ret.predictions
speed_up = prediction_ret.metrics['eval_expected_acceleration']
predictions = np.argmax(predictions, axis=1)
output_test_file = os.path.join(training_args.output_dir, "test_results_%s_TH%.6f_spd%.2f.txt" % (task,
model_args.confidence_threshold,
speed_up))
if trainer.is_world_process_zero():
with open(output_test_file, "w") as writer:
logger.info(f"***** Test results {task} *****")
writer.write("index\tprediction\n")
for index, item in enumerate(predictions):
item = label_list[item]
writer.write(f"{index}\t{item}\n")
return eval_results
def _mp_fn(index):
# For xla_spawn (TPUs)
main()
if __name__ == "__main__":
main()