def bertForQuestionAnswering(*args, **kwargs):
"""
BertForQuestionAnswering is a fine-tuning model that includes BertModel
with a token-level classifiers on top of the full sequence of last hidden
states. Note that the classification head is only initialized
and has to be trained.
Example:
# Load the tokenizer
>>> import torch
>>> tokenizer = torch.hub.load('huggingface/pytorch-pretrained-BERT', 'bertTokenizer', 'bert-base-cased', do_basic_tokenize=False)
# Prepare tokenized input
>>> text = "[CLS] Who was Jim Henson ? [SEP] Jim Henson was a puppeteer [SEP]"
>>> tokenized_text = tokenizer.tokenize(text)
>>> indexed_tokens = tokenizer.convert_tokens_to_ids(tokenized_text)
>>> segments_ids = [0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1]
>>> tokens_tensor = torch.tensor([indexed_tokens])
>>> segments_tensors = torch.tensor([segments_ids])
# Load bertForQuestionAnswering
>>> model = torch.hub.load('huggingface/pytorch-pretrained-BERT', 'bertForQuestionAnswering', 'bert-base-cased')
>>> model.eval()
# Predict the start and end positions logits
>>> with torch.no_grad():
start_logits, end_logits = model(tokens_tensor, segments_tensors)
# Or get the total loss which is the sum of the CrossEntropy loss for the start and end token positions
>>> start_positions, end_positions = torch.tensor([12]), torch.tensor([14])
# set model.train() before if training this loss
>>> multiple_choice_loss = model(tokens_tensor, segments_tensors, start_positions=start_positions, end_positions=end_positions)
"""
model = BertForQuestionAnswering.from_pretrained(*args, **kwargs)
return model
def bertForQuestionAnswering(*args, **kwargs):
"""
BertForQuestionAnswering is a fine-tuning model that includes BertModel
with a token-level classifiers on top of the full sequence of last hidden
states.
"""
model = BertForQuestionAnswering.from_pretrained(*args, **kwargs)
return model
def __init__(self):
self.log = {}
self.device = torch.device(
"cuda" if torch.cuda.is_available() else "cpu")
self.tokenizer = BertTokenizer.from_pretrained(
args.bert_model, do_lower_case=args.do_lower_case)
self.model = BertForQuestionAnswering.from_pretrained(args.bert_model)
self.model.to(self.device)
self.model.eval()
self.args = args
def __init__(self,
token_makers,
pretrained_model_name=None,
answer_maxlen=30):
super(BertForQA, self).__init__(token_makers)
self.bert = True # for optimizer's model parameters
self.answer_maxlen = answer_maxlen
self.model = BertForQuestionAnswering.from_pretrained(
pretrained_model_name, cache_dir=str(CachePath.ROOT))
self.criterion = nn.CrossEntropyLoss()
def __init__(self, word_vectors, char_vectors, hidden_size, drop_prob):
super(Embedding, self).__init__()
self.drop_prob = drop_prob
# self.embed = nn.Embedding.from_pretrained(word_vectors)
# This is actually BERT
self.embed = BertForQuestionAnswering.from_pretrained(
'bert-large-uncased')
for name, param in self.embed.named_parameters():
param.requires_grad = False
self.embed_char = CNNEmbeddings(char_vectors=char_vectors,
embed_size=64) # added_flag
# self.proj = nn.Linear(word_vectors.size(1), hidden_size, bias=False)
self.proj = nn.Linear(64 + 1024, hidden_size,
bias=False) # added_flag hardcoded
self.hwy = HighwayEncoder(2, hidden_size)
def main():
parser = argparse.ArgumentParser()
## Required parameters
parser.add_argument(
"--bert_model",
default=None,
type=str,
required=True,
help="Bert pre-trained model selected in the list: bert-base-uncased, "
"bert-large-uncased, bert-base-cased, bert-base-multilingual, bert-base-chinese."
)
parser.add_argument(
"--output_dir",
default=None,
type=str,
required=True,
help=
"The output directory where the model checkpoints and predictions will be written."
)
## Other parameters
parser.add_argument("--bin_path", default=None, type=str, help="bin path")
parser.add_argument(
"--predict_file",
default=None,
type=str,
help="SQuAD json for predictions. E.g., dev-v1.1.json or test-v1.1.json"
)
parser.add_argument(
"--max_seq_length",
default=384,
type=int,
help=
"The maximum total input sequence length after WordPiece tokenization. Sequences "
"longer than this will be truncated, and sequences shorter than this will be padded."
)
parser.add_argument(
"--doc_stride",
default=128,
type=int,
help=
"When splitting up a long document into chunks, how much stride to take between chunks."
)
parser.add_argument(
"--max_query_length",
default=64,
type=int,
help=
"The maximum number of tokens for the question. Questions longer than this will "
"be truncated to this length.")
parser.add_argument("--predict_batch_size",
default=8,
type=int,
help="Total batch size for predictions.")
parser.add_argument("--learning_rate",
default=5e-5,
type=float,
help="The initial learning rate for Adam.")
parser.add_argument("--num_train_epochs",
default=3.0,
type=float,
help="Total number of training epochs to perform.")
parser.add_argument(
"--warmup_proportion",
default=0.1,
type=float,
help=
"Proportion of training to perform linear learning rate warmup for. E.g., 0.1 = 10% "
"of training.")
parser.add_argument(
"--n_best_size",
default=5,
type=int,
help=
"The total number of n-best predictions to generate in the nbest_predictions.json "
"output file.")
parser.add_argument(
"--max_answer_length",
default=500,
type=int,
help=
"The maximum length of an answer that can be generated. This is needed because the start "
"and end predictions are not conditioned on one another.")
parser.add_argument(
"--verbose_logging",
default=False,
action='store_true',
help=
"If true, all of the warnings related to data processing will be printed. "
"A number of warnings are expected for a normal SQuAD evaluation.")
parser.add_argument("--no_cuda",
default=False,
action='store_true',
help="Whether not to use CUDA when available")
parser.add_argument('--seed',
type=int,
default=42,
help="random seed for initialization")
parser.add_argument(
'--gradient_accumulation_steps',
type=int,
default=1,
help=
"Number of updates steps to accumulate before performing a backward/update pass."
)
parser.add_argument(
"--do_lower_case",
default=True,
action='store_true',
help=
"Whether to lower case the input text. True for uncased models, False for cased models."
)
parser.add_argument("--local_rank",
type=int,
default=-1,
help="local_rank for distributed training on gpus")
parser.add_argument(
'--fp16',
default=False,
action='store_true',
help="Whether to use 16-bit float precision instead of 32-bit")
parser.add_argument(
'--loss_scale',
type=float,
default=0,
help=
"Loss scaling to improve fp16 numeric stability. Only used when fp16 set to True.\n"
"0 (default value): dynamic loss scaling.\n"
"Positive power of 2: static loss scaling value.\n")
args = parser.parse_args()
if args.local_rank == -1 or args.no_cuda:
device = torch.device("cuda" if torch.cuda.is_available()
and not args.no_cuda else "cpu")
n_gpu = torch.cuda.device_count()
else:
torch.cuda.set_device(args.local_rank)
device = torch.device("cuda", args.local_rank)
n_gpu = 1
# Initializes the distributed backend which will take care of sychronizing nodes/GPUs
torch.distributed.init_process_group(backend='nccl')
logger.info(
"device: {} n_gpu: {}, distributed training: {}, 16-bits training: {}".
format(device, n_gpu, bool(args.local_rank != -1), args.fp16))
if args.gradient_accumulation_steps < 1:
raise ValueError(
"Invalid gradient_accumulation_steps parameter: {}, should be >= 1"
.format(args.gradient_accumulation_steps))
# args.train_batch_size = int(args.train_batch_size / args.gradient_accumulation_steps)
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
if n_gpu > 0:
torch.cuda.manual_seed_all(args.seed)
# if not args.do_train and not args.do_predict:
# raise ValueError("At least one of `do_train` or `do_predict` must be True.")
# if args.do_train:
# if not args.train_file:
# raise ValueError(
# "If `do_train` is True, then `train_file` must be specified.")
# if args.do_predict:
# if not args.predict_file:
# raise ValueError(
# "If `do_predict` is True, then `predict_file` must be specified.")
if os.path.exists(args.output_dir) and os.listdir(args.output_dir):
raise ValueError(
"Output directory () already exists and is not empty.")
os.makedirs(args.output_dir, exist_ok=True)
tokenizer = BertTokenizer.from_pretrained(args.bert_model)
# Save a trained model
# model_to_save = model.module if hasattr(model, 'module') else model # Only save the model it-self
# output_model_file = os.path.join(args.output_dir, "pytorch_model.bin")
# torch.save(model_to_save.state_dict(), output_model_file)
# Load a trained model that you have fine-tuned
model_state_dict = torch.load(args.bin_path)
model = BertForQuestionAnswering.from_pretrained(
args.bert_model, state_dict=model_state_dict)
model.to(device)
eval_examples = read_squad_examples(input_file=args.predict_file,
is_training=False)
eval_features = convert_examples_to_features(
examples=eval_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=False)
logger.info("***** Running predictions *****")
logger.info(" Num orig examples = %d", len(eval_examples))
logger.info(" Num split examples = %d", len(eval_features))
logger.info(" Batch size = %d", args.predict_batch_size)
all_input_ids = torch.tensor([f.input_ids for f in eval_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in eval_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in eval_features],
dtype=torch.long)
all_example_index = torch.arange(all_input_ids.size(0), dtype=torch.long)
eval_data = TensorDataset(all_input_ids, all_input_mask, all_segment_ids,
all_example_index)
# Run prediction for full data
eval_sampler = SequentialSampler(eval_data)
eval_dataloader = DataLoader(eval_data,
sampler=eval_sampler,
batch_size=args.predict_batch_size)
model.eval()
all_results = []
logger.info("Start evaluating")
for input_ids, input_mask, segment_ids, example_indices in tqdm(
eval_dataloader, desc="Evaluating"):
if len(all_results) % 1000 == 0:
logger.info("Processing example: %d" % (len(all_results)))
input_ids = input_ids.to(device)
input_mask = input_mask.to(device)
segment_ids = segment_ids.to(device)
with torch.no_grad():
batch_start_logits, batch_end_logits = model(
input_ids, segment_ids, input_mask)
for i, example_index in enumerate(example_indices):
start_logits = batch_start_logits[i].detach().cpu().tolist()
end_logits = batch_end_logits[i].detach().cpu().tolist()
eval_feature = eval_features[example_index.item()]
unique_id = int(eval_feature.unique_id)
all_results.append(
RawResult(unique_id=unique_id,
start_logits=start_logits,
end_logits=end_logits))
output_prediction_file = os.path.join(args.output_dir, "predictions.json")
output_nbest_file = os.path.join(args.output_dir, "nbest_predictions.json")
write_predictions(eval_examples, eval_features, all_results,
args.n_best_size, args.max_answer_length,
args.do_lower_case, output_prediction_file,
output_nbest_file, args.verbose_logging)
def main():
args = set_config()
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
tokenizer = BertTokenizer.from_pretrained(args.bert_model,
do_lower_case=args.do_lower_case)
# Prepare model
model = BertForQuestionAnswering2.from_pretrained(args.bert_model)
# model = BertForQuestionAnswering1.from_pretrained(args.bert_model)
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
# output_model_file = os.path.join(args.output_dir, "pytorch_model_{}.bin".format(0))
# model_state_dict = torch.load(output_model_file)
# model = BertForQuestionAnswering2.from_pretrained(args.bert_model, state_dict=model_state_dict)
model.to(device)
optimizer = torch.optim.Adam(model.parameters(), lr=args.learning_rate)
global_step = 0
if args.do_train:
# load train data
train_examples, train_features = get_train_feature(
args, args.do_train, tokenizer)
train_data = DataIteratorPack(train_features,
train_examples,
args.train_batch_size,
device,
sent_limit=None,
sequential=False)
# train_example_dict = example_dict(train_examples)
# load dev data
eval_examples, eval_features = get_train_feature(
args, not args.do_train, tokenizer)
logger.info("***** Running training *****")
logger.info(" Num split examples = %d", len(train_features))
logger.info(" Batch size = %d", args.train_batch_size)
total_train_loss = 0
VERBOSE_STEP = 100
grad_accumulate_step = 1
for epoch in trange(int(args.num_train_epochs), desc="Epoch"):
model.train()
# learning rate decay
# if epoch > 1:
# args.learning_rate = args.learning_rate * args.decay
# for param_group in optimizer.param_groups:
# param_group['lr'] = args.learning_rate
# print('lr = {}'.format(args.learning_rate))
for step, batch in enumerate(train_data):
# batch = tuple(t.to(device) for t in batch) # multi-gpu does scattering it-self
input_ids = batch["context_idxs"]
input_mask = batch["context_mask"]
segment_ids = batch["segment_idxs"]
start_positions = batch["y1"]
end_positions = batch["y2"]
q_types = batch["q_type"]
loss = model(input_ids, segment_ids, input_mask,
start_positions, end_positions, q_types)
if args.gradient_accumulation_steps > 1:
loss = loss / args.gradient_accumulation_steps
loss.backward()
if (global_step + 1) % grad_accumulate_step == 0:
optimizer.step()
optimizer.zero_grad()
total_train_loss += loss
global_step += 1
if global_step % VERBOSE_STEP == 0:
print("-- In Epoch{}: ".format(epoch))
print("Avg-LOSS: {}/batch/step: {}".format(
total_train_loss / VERBOSE_STEP,
global_step / VERBOSE_STEP))
total_train_loss = 0
# Save a trained model
model_to_save = model.module if hasattr(
model, 'module') else model # Only save the model it-self
output_model_file = os.path.join(
args.output_dir, "pytorch_model_{}.bin".format(epoch))
if global_step % VERBOSE_STEP == 0:
if args.do_train:
torch.save(model_to_save.state_dict(),
output_model_file)
# Load a trained model that you have fine-tuned
# model_state_dict = torch.load(output_model_file)
# model = BertForQuestionAnswering.from_pretrained(args.bert_model, state_dict=model_state_dict)
else:
model = BertForQuestionAnswering.from_pretrained(
args.bert_model)
model.to(device)
train_data.refresh()
if args.do_predict:
eval_examples_dict = example_dict(eval_examples)
# eval_features_dict = example_dict(eval_features)
logger.info("***** Running predictions *****")
logger.info(" Num split examples = %d", len(eval_features))
logger.info(" Batch size = %d", args.predict_batch_size)
all_input_ids = torch.tensor(
[f.doc_input_ids for f in eval_features], dtype=torch.long)
all_input_mask = torch.tensor(
[f.doc_input_mask for f in eval_features],
dtype=torch.long)
all_segment_ids = torch.tensor(
[f.doc_segment_ids for f in eval_features],
dtype=torch.long)
all_example_index = torch.arange(all_input_ids.size(0),
dtype=torch.long)
eval_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_example_index)
eval_sampler = SequentialSampler(eval_data)
eval_dataloader = DataLoader(
eval_data,
sampler=eval_sampler,
batch_size=args.predict_batch_size)
# eval_dataloader = DataIteratorPack(eval_features, eval_examples, args.train_batch_size, device, sent_limit=None,
# sequential=False)
model.eval()
all_results = []
logger.info("Start evaluating")
for input_ids, input_mask, segment_ids, example_indices in tqdm(
eval_dataloader, desc="Evaluating"):
if len(all_results) % 1000 == 0:
logger.info("Processing example: %d" %
(len(all_results)))
input_ids = input_ids.to(device)
input_mask = input_mask.to(device)
segment_ids = segment_ids.to(device)
with torch.no_grad():
batch_start_logits, batch_end_logits, batch_types = model(
input_ids, segment_ids, input_mask)
for i, example_index in enumerate(example_indices):
start_logits = [
batch_start_logits[j][i].detach().cpu().tolist()
for j in range(len((batch_start_logits)))
]
end_logits = [
batch_end_logits[j][i].detach().cpu().tolist()
for j in range(len((batch_end_logits)))
]
eval_feature = eval_features[example_index.item()]
unique_id = int(eval_feature.unique_id)
types = batch_types[i].detach().cpu().tolist()
all_results.append(
RawResult(unique_id=unique_id,
start_logits=start_logits,
end_logits=end_logits,
types=types))
output_prediction_file = os.path.join(
args.output_dir, "predictions_{}.json".format(epoch))
output_nbest_file = os.path.join(
args.output_dir, "nbest_predictions_{}.json".format(epoch))
output_null_log_odds_file = os.path.join(
args.output_dir, "null_odds_{}.json".format(epoch))
all_predictions = write_predictions(
eval_examples, eval_features, all_results,
args.n_best_size, args.max_answer_length,
args.do_lower_case, output_prediction_file,
output_nbest_file, output_null_log_odds_file,
args.verbose_logging, args.version_2_with_negative,
args.null_score_diff_threshold)
metrics = evaluate(eval_examples_dict, all_predictions)
print('epoch {:3d} | EM {:.4f} | F1 {:.4f}'.format(
epoch, metrics['exact_match'], metrics['f1']))
def BertSquad(file="",
mode='predict',
bert_model="bert-base-uncased",
output='./output'):
parser = {}
parser["bert_model"] = bert_model
parser["output_dir"] = output
parser["train_file"] = file
parser["predict_file"] = file
parser["max_seq_length"] = 384
parser["doc_stride"] = 128
parser["max_query_length"] = 64
parser["do_train"] = mode == 'train'
parser["do_predict"] = mode == 'predict'
parser["train_batch_size"] = 32
parser["predict_batch_size"] = 8
parser["learning_rate"] = 5e-5
parser["num_train_epochs"] = 3.0
parser["warmup_proportion"] = 0.1
parser["n_best_size"] = 20
parser["max_answer_length"] = 30
parser["verbose_logging"] = False
parser["no_cuda"] = False
parser['seed'] = 42
parser['gradient_accumulation_steps'] = 1
parser["do_lower_case"] = ('uncased' in bert_model)
parser["local_rank"] = -1
parser['fp16'] = False
parser['overwrite_output_dir'] = False
parser['loss_scale'] = 0
parser['version_2_with_negative'] = False
parser['null_score_diff_threshold'] = 0.0
parser['server_ip'] = ''
parser['server_port'] = ''
args = AttrDict.AttrDict(parser)
print(args)
if args.server_ip and args.server_port:
# Distant debugging - see https://code.visualstudio.com/docs/python/debugging#_attach-to-a-local-script
import ptvsd
print("Waiting for debugger attach")
ptvsd.enable_attach(address=(args.server_ip, args.server_port),
redirect_output=True)
ptvsd.wait_for_attach()
if args.local_rank == -1 or args.no_cuda:
device = torch.device("cuda" if torch.cuda.is_available()
and not args.no_cuda else "cpu")
n_gpu = torch.cuda.device_count()
else:
torch.cuda.set_device(args.local_rank)
device = torch.device("cuda", args.local_rank)
n_gpu = 1
# Initializes the distributed backend which will take care of sychronizing nodes/GPUs
torch.distributed.init_process_group(backend='nccl')
logging.basicConfig(
format='%(asctime)s - %(levelname)s - %(name)s - %(message)s',
datefmt='%m/%d/%Y %H:%M:%S',
level=logging.INFO if args.local_rank in [-1, 0] else logging.WARN)
logger.info(
"device: {} n_gpu: {}, distributed training: {}, 16-bits training: {}".
format(device, n_gpu, bool(args.local_rank != -1), args.fp16))
if args.gradient_accumulation_steps < 1:
raise ValueError(
"Invalid gradient_accumulation_steps parameter: {}, should be >= 1"
.format(args.gradient_accumulation_steps))
args.train_batch_size = args.train_batch_size // args.gradient_accumulation_steps
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
if n_gpu > 0:
torch.cuda.manual_seed_all(args.seed)
if not args.do_train and not args.do_predict:
raise ValueError(
"At least one of `do_train` or `do_predict` must be True.")
if args.do_train:
if not args.train_file:
raise ValueError(
"If `do_train` is True, then `train_file` must be specified.")
if args.do_predict:
if not args.predict_file:
raise ValueError(
"If `do_predict` is True, then `predict_file` must be specified."
)
if os.path.exists(args.output_dir) and os.listdir(
args.output_dir
) and args.do_train and not args.overwrite_output_dir:
raise ValueError(
"Output directory () already exists and is not empty.")
if not os.path.exists(args.output_dir):
os.makedirs(args.output_dir)
if args.local_rank not in [-1, 0]:
torch.distributed.barrier(
) # Make sure only the first process in distributed training will download model & vocab
tokenizer = BertTokenizer.from_pretrained(args.bert_model,
do_lower_case=args.do_lower_case)
model = BertForQuestionAnswering.from_pretrained(args.bert_model)
if args.local_rank == 0:
torch.distributed.barrier()
if args.fp16:
model.half()
model.to(device)
if args.local_rank != -1:
model = torch.nn.parallel.DistributedDataParallel(
model,
device_ids=[args.local_rank],
output_device=args.local_rank,
find_unused_parameters=True)
elif n_gpu > 1:
model = torch.nn.DataParallel(model)
if args.do_train:
if args.local_rank in [-1, 0]:
tb_writer = SummaryWriter()
# Prepare data loader
train_examples = read_squad_examples(
input_file=args.train_file,
is_training=True,
version_2_with_negative=args.version_2_with_negative)
cached_train_features_file = args.train_file + '_{0}_{1}_{2}_{3}'.format(
list(filter(None, args.bert_model.split('/'))).pop(),
str(args.max_seq_length), str(args.doc_stride),
str(args.max_query_length))
try:
with open(cached_train_features_file, "rb") as reader:
train_features = pickle.load(reader)
except:
train_features = convert_examples_to_features(
examples=train_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=True)
if args.local_rank == -1 or torch.distributed.get_rank() == 0:
logger.info(" Saving train features into cached file %s",
cached_train_features_file)
with open(cached_train_features_file, "wb") as writer:
pickle.dump(train_features, writer)
all_input_ids = torch.tensor([f.input_ids for f in train_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in train_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in train_features],
dtype=torch.long)
all_start_positions = torch.tensor(
[f.start_position for f in train_features], dtype=torch.long)
all_end_positions = torch.tensor(
[f.end_position for f in train_features], dtype=torch.long)
train_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_start_positions,
all_end_positions)
if args.local_rank == -1:
train_sampler = RandomSampler(train_data)
else:
train_sampler = DistributedSampler(train_data)
train_dataloader = DataLoader(train_data,
sampler=train_sampler,
batch_size=args.train_batch_size)
num_train_optimization_steps = len(
train_dataloader
) // args.gradient_accumulation_steps * args.num_train_epochs
# if args.local_rank != -1:
# num_train_optimization_steps = num_train_optimization_steps // torch.distributed.get_world_size()
# Prepare optimizer
param_optimizer = list(model.named_parameters())
# hack to remove pooler, which is not used
# thus it produce None grad that break apex
param_optimizer = [n for n in param_optimizer if 'pooler' not in n[0]]
no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [{
'params': [
p for n, p in param_optimizer
if not any(nd in n for nd in no_decay)
],
'weight_decay':
0.01
}, {
'params':
[p for n, p in param_optimizer if any(nd in n for nd in no_decay)],
'weight_decay':
0.0
}]
if args.fp16:
try:
from apex.optimizers import FP16_Optimizer
from apex.optimizers import FusedAdam
except ImportError:
raise ImportError(
"Please install apex from https://www.github.com/nvidia/apex to use distributed and fp16 training."
)
optimizer = FusedAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
bias_correction=False,
max_grad_norm=1.0)
if args.loss_scale == 0:
optimizer = FP16_Optimizer(optimizer, dynamic_loss_scale=True)
else:
optimizer = FP16_Optimizer(optimizer,
static_loss_scale=args.loss_scale)
warmup_linear = WarmupLinearSchedule(
warmup=args.warmup_proportion,
t_total=num_train_optimization_steps)
else:
optimizer = BertAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
warmup=args.warmup_proportion,
t_total=num_train_optimization_steps)
global_step = 0
logger.info("***** Running training *****")
logger.info(" Num orig examples = %d", len(train_examples))
logger.info(" Num split examples = %d", len(train_features))
logger.info(" Batch size = %d", args.train_batch_size)
logger.info(" Num steps = %d", num_train_optimization_steps)
model.train()
for epoch in trange(int(args.num_train_epochs), desc="Epoch"):
for step, batch in enumerate(
tqdm(train_dataloader,
desc="Iteration",
disable=args.local_rank not in [-1, 0])):
if n_gpu == 1:
batch = tuple(
t.to(device)
for t in batch) # multi-gpu does scattering it-self
input_ids, input_mask, segment_ids, start_positions, end_positions = batch
loss = model(input_ids, segment_ids, input_mask,
start_positions, end_positions)
if n_gpu > 1:
loss = loss.mean() # mean() to average on multi-gpu.
if args.gradient_accumulation_steps > 1:
loss = loss / args.gradient_accumulation_steps
if args.fp16:
optimizer.backward(loss)
else:
loss.backward()
if (step + 1) % args.gradient_accumulation_steps == 0:
if args.fp16:
# modify learning rate with special warm up BERT uses
# if args.fp16 is False, BertAdam is used and handles this automatically
lr_this_step = args.learning_rate * warmup_linear.get_lr(
global_step, args.warmup_proportion)
for param_group in optimizer.param_groups:
param_group['lr'] = lr_this_step
optimizer.step()
optimizer.zero_grad()
global_step += 1
if args.local_rank in [-1, 0]:
tb_writer.add_scalar('lr',
optimizer.get_lr()[0],
global_step)
tb_writer.add_scalar('loss', loss.item(), global_step)
if args.do_train and (args.local_rank == -1
or torch.distributed.get_rank() == 0):
# Save a trained model, configuration and tokenizer
model_to_save = model.module if hasattr(
model, 'module') else model # Only save the model it-self
# If we save using the predefined names, we can load using `from_pretrained`
output_model_file = os.path.join(args.output_dir, WEIGHTS_NAME)
output_config_file = os.path.join(args.output_dir, CONFIG_NAME)
torch.save(model_to_save.state_dict(), output_model_file)
model_to_save.config.to_json_file(output_config_file)
tokenizer.save_vocabulary(args.output_dir)
# Load a trained model and vocabulary that you have fine-tuned
model = BertForQuestionAnswering.from_pretrained(args.output_dir)
tokenizer = BertTokenizer.from_pretrained(
args.output_dir, do_lower_case=args.do_lower_case)
# Good practice: save your training arguments together with the trained model
output_args_file = os.path.join(args.output_dir, 'training_args.bin')
torch.save(args, output_args_file)
else:
# Load a trained model and vocabulary that you have fine-tuned
model = BertForQuestionAnswering.from_pretrained(args.output_dir)
tokenizer = BertTokenizer.from_pretrained(
args.output_dir, do_lower_case=args.do_lower_case)
model.to(device)
if args.do_predict and (args.local_rank == -1
or torch.distributed.get_rank() == 0):
eval_examples = read_squad_examples(
input_file=args.predict_file,
is_training=False,
version_2_with_negative=args.version_2_with_negative)
eval_features = convert_examples_to_features(
examples=eval_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=False)
logger.info("***** Running predictions *****")
logger.info(" Num orig examples = %d", len(eval_examples))
logger.info(" Num split examples = %d", len(eval_features))
logger.info(" Batch size = %d", args.predict_batch_size)
all_input_ids = torch.tensor([f.input_ids for f in eval_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in eval_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in eval_features],
dtype=torch.long)
all_example_index = torch.arange(all_input_ids.size(0),
dtype=torch.long)
eval_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_example_index)
# Run prediction for full data
eval_sampler = SequentialSampler(eval_data)
eval_dataloader = DataLoader(eval_data,
sampler=eval_sampler,
batch_size=args.predict_batch_size)
model.eval()
all_results = []
logger.info("Start evaluating")
for input_ids, input_mask, segment_ids, example_indices in tqdm(
eval_dataloader,
desc="Evaluating",
disable=args.local_rank not in [-1, 0]):
if len(all_results) % 1000 == 0:
logger.info("Processing example: %d" % (len(all_results)))
input_ids = input_ids.to(device)
input_mask = input_mask.to(device)
segment_ids = segment_ids.to(device)
with torch.no_grad():
batch_start_logits, batch_end_logits = model(
input_ids, segment_ids, input_mask)
for i, example_index in enumerate(example_indices):
start_logits = batch_start_logits[i].detach().cpu().tolist()
end_logits = batch_end_logits[i].detach().cpu().tolist()
eval_feature = eval_features[example_index.item()]
unique_id = int(eval_feature.unique_id)
all_results.append(
RawResult(unique_id=unique_id,
start_logits=start_logits,
end_logits=end_logits))
output_prediction_file = os.path.join(args.output_dir,
"predictions.json")
output_nbest_file = os.path.join(args.output_dir,
"nbest_predictions.json")
output_null_log_odds_file = os.path.join(args.output_dir,
"null_odds.json")
write_predictions(eval_examples, eval_features, all_results,
args.n_best_size, args.max_answer_length,
args.do_lower_case, output_prediction_file,
output_nbest_file, output_null_log_odds_file,
args.verbose_logging, args.version_2_with_negative,
args.null_score_diff_threshold)
def main():
parser = argparse.ArgumentParser()
## Required parameters
parser.add_argument(
"--bert_model",
default=None,
type=str,
required=True,
help="Bert pre-trained model selected in the list: bert-base-uncased, "
"bert-large-uncased, bert-base-cased, bert-base-multilingual, bert-base-chinese."
)
parser.add_argument(
"--output_dir",
default=None,
type=str,
required=True,
help=
"The output directory where the model checkpoints and predictions will be written."
)
## Other parameters
parser.add_argument("--train_file",
default=None,
type=str,
help="SQuAD json for training. E.g., train-v1.1.json")
parser.add_argument(
"--predict_file",
default=None,
type=str,
help="SQuAD json for predictions. E.g., dev-v1.1.json or test-v1.1.json"
)
parser.add_argument(
"--max_seq_length",
default=384,
type=int,
help=
"The maximum total input sequence length after WordPiece tokenization. Sequences "
"longer than this will be truncated, and sequences shorter than this will be padded."
)
parser.add_argument(
"--doc_stride",
default=128,
type=int,
help=
"When splitting up a long document into chunks, how much stride to take between chunks."
)
parser.add_argument(
"--max_query_length",
default=64,
type=int,
help=
"The maximum number of tokens for the question. Questions longer than this will "
"be truncated to this length.")
parser.add_argument("--do_train",
default=False,
action='store_true',
help="Whether to run training.")
parser.add_argument("--do_predict",
default=False,
action='store_true',
help="Whether to run eval on the dev set.")
parser.add_argument("--train_batch_size",
default=32,
type=int,
help="Total batch size for training.")
parser.add_argument("--predict_batch_size",
default=8,
type=int,
help="Total batch size for predictions.")
parser.add_argument("--learning_rate",
default=5e-5,
type=float,
help="The initial learning rate for Adam.")
parser.add_argument("--num_train_epochs",
default=3.0,
type=float,
help="Total number of training epochs to perform.")
parser.add_argument(
"--warmup_proportion",
default=0.1,
type=float,
help=
"Proportion of training to perform linear learning rate warmup for. E.g., 0.1 = 10% "
"of training.")
parser.add_argument(
"--n_best_size",
default=20,
type=int,
help=
"The total number of n-best predictions to generate in the nbest_predictions.json "
"output file.")
parser.add_argument(
"--max_answer_length",
default=30,
type=int,
help=
"The maximum length of an answer that can be generated. This is needed because the start "
"and end predictions are not conditioned on one another.")
parser.add_argument(
"--verbose_logging",
default=False,
action='store_true',
help=
"If true, all of the warnings related to data processing will be printed. "
"A number of warnings are expected for a normal SQuAD evaluation.")
parser.add_argument("--no_cuda",
default=False,
action='store_true',
help="Whether not to use CUDA when available")
parser.add_argument('--seed',
type=int,
default=42,
help="random seed for initialization")
parser.add_argument(
'--gradient_accumulation_steps',
type=int,
default=1,
help=
"Number of updates steps to accumulate before performing a backward/update pass."
)
parser.add_argument(
"--do_lower_case",
default=True,
action='store_true',
help=
"Whether to lower case the input text. True for uncased models, False for cased models."
)
parser.add_argument("--local_rank",
type=int,
default=-1,
help="local_rank for distributed training on gpus")
parser.add_argument(
'--fp16',
default=False,
action='store_true',
help="Whether to use 16-bit float precision instead of 32-bit")
parser.add_argument(
'--loss_scale',
type=float,
default=0,
help=
"Loss scaling to improve fp16 numeric stability. Only used when fp16 set to True.\n"
"0 (default value): dynamic loss scaling.\n"
"Positive power of 2: static loss scaling value.\n")
args = parser.parse_args()
if args.local_rank == -1 or args.no_cuda:
device = torch.device("cuda" if torch.cuda.is_available()
and not args.no_cuda else "cpu")
# device = torch.device("cuda:1")
n_gpu = torch.cuda.device_count()
# n_gpu = 1
else:
torch.cuda.set_device(args.local_rank)
device = torch.device("cuda", args.local_rank)
n_gpu = 1
# Initializes the distributed backend which will take care of sychronizing nodes/GPUs
torch.distributed.init_process_group(backend='nccl')
logger.info(
"device: {} n_gpu: {}, distributed training: {}, 16-bits training: {}".
format(device, n_gpu, bool(args.local_rank != -1), args.fp16))
if args.gradient_accumulation_steps < 1:
raise ValueError(
"Invalid gradient_accumulation_steps parameter: {}, should be >= 1"
.format(args.gradient_accumulation_steps))
args.train_batch_size = int(args.train_batch_size /
args.gradient_accumulation_steps)
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
if n_gpu > 0:
torch.cuda.manual_seed_all(args.seed)
if not args.do_train and not args.do_predict:
raise ValueError(
"At least one of `do_train` or `do_predict` must be True.")
if args.do_train:
if not args.train_file:
raise ValueError(
"If `do_train` is True, then `train_file` must be specified.")
if args.do_predict:
if not args.predict_file:
raise ValueError(
"If `do_predict` is True, then `predict_file` must be specified."
)
if os.path.exists(args.output_dir) and os.listdir(args.output_dir):
raise ValueError(
"Output directory () already exists and is not empty.")
os.makedirs(args.output_dir, exist_ok=True)
tokenizer = BertTokenizer.from_pretrained(args.bert_model)
train_examples = None
num_train_steps = None
if args.do_train:
train_examples = read_squad_examples(input_file=args.train_file,
is_training=True)
num_train_steps = int(
len(train_examples) / args.train_batch_size /
args.gradient_accumulation_steps * args.num_train_epochs)
# Prepare model
model = BertForQuestionAnswering.from_pretrained(
args.bert_model,
cache_dir=PYTORCH_PRETRAINED_BERT_CACHE /
'distributed_{}'.format(args.local_rank))
if args.fp16:
model.half()
model.to(device)
if args.local_rank != -1:
try:
from apex.parallel import DistributedDataParallel as DDP
except ImportError:
raise ImportError(
"Please install apex from https://www.github.com/nvidia/apex to use distributed and fp16 training."
)
model = DDP(model)
elif n_gpu > 1:
model = torch.nn.DataParallel(model)
# Prepare optimizer
param_optimizer = list(model.named_parameters())
# hack to remove pooler, which is not used
# thus it produce None grad that break apex
param_optimizer = [n for n in param_optimizer if 'pooler' not in n[0]]
no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [{
'params':
[p for n, p in param_optimizer if not any(nd in n for nd in no_decay)],
'weight_decay':
0.01
}, {
'params':
[p for n, p in param_optimizer if any(nd in n for nd in no_decay)],
'weight_decay':
0.0
}]
t_total = num_train_steps
if args.local_rank != -1:
t_total = t_total // torch.distributed.get_world_size()
if args.fp16:
try:
from apex.optimizers import FP16_Optimizer
from apex.optimizers import FusedAdam
except ImportError:
raise ImportError(
"Please install apex from https://www.github.com/nvidia/apex to use distributed and fp16 training."
)
optimizer = FusedAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
bias_correction=False,
max_grad_norm=1.0)
if args.loss_scale == 0:
optimizer = FP16_Optimizer(optimizer, dynamic_loss_scale=True)
else:
optimizer = FP16_Optimizer(optimizer,
static_loss_scale=args.loss_scale)
else:
optimizer = BertAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
warmup=args.warmup_proportion,
t_total=t_total)
global_step = 0
if args.do_train:
cached_train_features_file = args.train_file + '_{1}_{2}_{3}'.format(
args.bert_model, str(args.max_seq_length), str(args.doc_stride),
str(args.max_query_length))
train_features = None
try:
with open(cached_train_features_file, "rb") as reader:
train_features = pickle.load(reader)
except:
train_features = convert_examples_to_features(
examples=train_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=True)
if args.local_rank == -1 or torch.distributed.get_rank() == 0:
logger.info(" Saving train features into cached file %s",
cached_train_features_file)
with open(cached_train_features_file, "wb") as writer:
pickle.dump(train_features, writer)
logger.info("***** Running training *****")
logger.info(" Num orig examples = %d", len(train_examples))
logger.info(" Num split examples = %d", len(train_features))
logger.info(" Batch size = %d", args.train_batch_size)
logger.info(" Num steps = %d", num_train_steps)
all_input_ids = torch.tensor([f.input_ids for f in train_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in train_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in train_features],
dtype=torch.long)
all_start_positions = torch.tensor(
[f.start_position for f in train_features], dtype=torch.long)
all_end_positions = torch.tensor(
[f.end_position for f in train_features], dtype=torch.long)
train_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_start_positions,
all_end_positions)
if args.local_rank == -1:
train_sampler = RandomSampler(train_data)
else:
train_sampler = DistributedSampler(train_data)
train_dataloader = DataLoader(train_data,
sampler=train_sampler,
batch_size=args.train_batch_size)
model.train()
for _ in trange(int(args.num_train_epochs), desc="Epoch"):
for step, batch in enumerate(
tqdm(train_dataloader, desc="Iteration")):
if n_gpu == 1:
batch = tuple(
t.to(device)
for t in batch) # multi-gpu does scattering it-self
input_ids, input_mask, segment_ids, start_positions, end_positions = batch
loss = model(input_ids, segment_ids, input_mask,
start_positions, end_positions)
if n_gpu > 1:
loss = loss.mean() # mean() to average on multi-gpu.
if args.gradient_accumulation_steps > 1:
loss = loss / args.gradient_accumulation_steps
print(loss, end=" \r")
if args.fp16:
optimizer.backward(loss)
else:
loss.backward()
if (step + 1) % args.gradient_accumulation_steps == 0:
# modify learning rate with special warm up BERT uses
lr_this_step = args.learning_rate * warmup_linear(
global_step / t_total, args.warmup_proportion)
for param_group in optimizer.param_groups:
param_group['lr'] = lr_this_step
optimizer.step()
optimizer.zero_grad()
global_step += 1
# Save a trained model
model_to_save = model.module if hasattr(
model, 'module') else model # Only save the model it-self
output_model_file = os.path.join(args.output_dir, "pytorch_model.bin")
torch.save(model_to_save.state_dict(), output_model_file)
# Load a trained model that you have fine-tuned
model_state_dict = torch.load(output_model_file)
model = BertForQuestionAnswering.from_pretrained(
args.bert_model, state_dict=model_state_dict)
model.to(device)
if args.do_predict and (args.local_rank == -1
or torch.distributed.get_rank() == 0):
eval_examples = read_squad_examples(input_file=args.predict_file,
is_training=False)
eval_features = convert_examples_to_features(
examples=eval_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=False)
logger.info("***** Running predictions *****")
logger.info(" Num orig examples = %d", len(eval_examples))
logger.info(" Num split examples = %d", len(eval_features))
logger.info(" Batch size = %d", args.predict_batch_size)
all_input_ids = torch.tensor([f.input_ids for f in eval_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in eval_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in eval_features],
dtype=torch.long)
all_example_index = torch.arange(all_input_ids.size(0),
dtype=torch.long)
eval_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_example_index)
# Run prediction for full data
eval_sampler = SequentialSampler(eval_data)
eval_dataloader = DataLoader(eval_data,
sampler=eval_sampler,
batch_size=args.predict_batch_size)
model.eval()
all_results = []
logger.info("Start evaluating")
for input_ids, input_mask, segment_ids, example_indices in tqdm(
eval_dataloader, desc="Evaluating"):
if len(all_results) % 1000 == 0:
logger.info("Processing example: %d" % (len(all_results)))
input_ids = input_ids.to(device)
input_mask = input_mask.to(device)
segment_ids = segment_ids.to(device)
with torch.no_grad():
batch_start_logits, batch_end_logits = model(
input_ids, segment_ids, input_mask)
for i, example_index in enumerate(example_indices):
start_logits = batch_start_logits[i].detach().cpu().tolist()
end_logits = batch_end_logits[i].detach().cpu().tolist()
eval_feature = eval_features[example_index.item()]
unique_id = int(eval_feature.unique_id)
all_results.append(
RawResult(unique_id=unique_id,
start_logits=start_logits,
end_logits=end_logits))
output_prediction_file = os.path.join(args.output_dir,
"predictions.json")
output_nbest_file = os.path.join(args.output_dir,
"nbest_predictions.json")
write_predictions(eval_examples, eval_features, all_results,
args.n_best_size, args.max_answer_length,
args.do_lower_case, output_prediction_file,
output_nbest_file, args.verbose_logging)
def main():
parser = argparse.ArgumentParser()
## Required parameters
parser.add_argument("--bert_model", default=None, type=str, required=True,
help="Bert pre-trained model selected in the list: bert-base-uncased, "
"bert-large-uncased, bert-base-cased, bert-large-cased, bert-base-multilingual-uncased, "
"bert-base-multilingual-cased, bert-base-chinese.")
parser.add_argument("--output_dir", default=None, type=str, required=True,
help="The output directory where the model checkpoints and predictions will be written.")
## Other parameters
parser.add_argument("--train_file", default=None, type=str, help="SQuAD json for training. E.g., train-v1.1.json")
parser.add_argument("--predict_file", default=None, type=str,
help="SQuAD json for predictions. E.g., dev-v1.1.json or test-v1.1.json")
parser.add_argument("--max_seq_length", default=384, type=int,
help="The maximum total input sequence length after WordPiece tokenization. Sequences "
"longer than this will be truncated, and sequences shorter than this will be padded.")
parser.add_argument("--doc_stride", default=128, type=int,
help="When splitting up a long document into chunks, how much stride to take between chunks.")
parser.add_argument("--max_query_length", default=64, type=int,
help="The maximum number of tokens for the question. Questions longer than this will "
"be truncated to this length.")
parser.add_argument("--do_train", action='store_true', help="Whether to run training.")
parser.add_argument("--do_predict", action='store_true', help="Whether to run eval on the dev set.")
parser.add_argument("--train_batch_size", default=32, type=int, help="Total batch size for training.")
parser.add_argument("--predict_batch_size", default=8, type=int, help="Total batch size for predictions.")
parser.add_argument("--learning_rate", default=5e-5, type=float, help="The initial learning rate for Adam.")
parser.add_argument("--num_train_epochs", default=3.0, type=float,
help="Total number of training epochs to perform.")
parser.add_argument("--warmup_proportion", default=0.1, type=float,
help="Proportion of training to perform linear learning rate warmup for. E.g., 0.1 = 10%% "
"of training.")
parser.add_argument("--n_best_size", default=20, type=int,
help="The total number of n-best predictions to generate in the nbest_predictions.json "
"output file.")
parser.add_argument("--max_answer_length", default=30, type=int,
help="The maximum length of an answer that can be generated. This is needed because the start "
"and end predictions are not conditioned on one another.")
parser.add_argument("--verbose_logging", action='store_true',
help="If true, all of the warnings related to data processing will be printed. "
"A number of warnings are expected for a normal SQuAD evaluation.")
parser.add_argument("--no_cuda",
action='store_true',
help="Whether not to use CUDA when available")
parser.add_argument('--seed',
type=int,
default=42,
help="random seed for initialization")
parser.add_argument('--gradient_accumulation_steps',
type=int,
default=1,
help="Number of updates steps to accumulate before performing a backward/update pass.")
parser.add_argument("--do_lower_case",
action='store_true',
help="Whether to lower case the input text. True for uncased models, False for cased models.")
parser.add_argument("--local_rank",
type=int,
default=-1,
help="local_rank for distributed training on gpus")
parser.add_argument('--fp16',
action='store_true',
help="Whether to use 16-bit float precision instead of 32-bit")
parser.add_argument('--overwrite_output_dir',
action='store_true',
help="Overwrite the content of the output directory")
parser.add_argument('--loss_scale',
type=float, default=0,
help="Loss scaling to improve fp16 numeric stability. Only used when fp16 set to True.\n"
"0 (default value): dynamic loss scaling.\n"
"Positive power of 2: static loss scaling value.\n")
parser.add_argument('--version_2_with_negative',
action='store_true',
help='If true, the SQuAD examples contain some that do not have an answer.')
parser.add_argument('--null_score_diff_threshold',
type=float, default=0.0,
help="If null_score - best_non_null is greater than the threshold predict null.")
parser.add_argument('--server_ip', type=str, default='', help="Can be used for distant debugging.")
parser.add_argument('--server_port', type=str, default='', help="Can be used for distant debugging.")
args = parser.parse_args()
print(args)
if args.server_ip and args.server_port:
# Distant debugging - see https://code.visualstudio.com/docs/python/debugging#_attach-to-a-local-script
import ptvsd
print("Waiting for debugger attach")
ptvsd.enable_attach(address=(args.server_ip, args.server_port), redirect_output=True)
ptvsd.wait_for_attach()
if args.local_rank == -1 or args.no_cuda:
device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
n_gpu = torch.cuda.device_count()
else:
torch.cuda.set_device(args.local_rank)
device = torch.device("cuda", args.local_rank)
n_gpu = 1
# Initializes the distributed backend which will take care of sychronizing nodes/GPUs
torch.distributed.init_process_group(backend='nccl')
logging.basicConfig(format = '%(asctime)s - %(levelname)s - %(name)s - %(message)s',
datefmt = '%m/%d/%Y %H:%M:%S',
level = logging.INFO if args.local_rank in [-1, 0] else logging.WARN)
logger.info("device: {} n_gpu: {}, distributed training: {}, 16-bits training: {}".format(
device, n_gpu, bool(args.local_rank != -1), args.fp16))
if args.gradient_accumulation_steps < 1:
raise ValueError("Invalid gradient_accumulation_steps parameter: {}, should be >= 1".format(
args.gradient_accumulation_steps))
args.train_batch_size = args.train_batch_size // args.gradient_accumulation_steps
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
if n_gpu > 0:
torch.cuda.manual_seed_all(args.seed)
if not args.do_train and not args.do_predict:
raise ValueError("At least one of `do_train` or `do_predict` must be True.")
if args.do_train:
if not args.train_file:
raise ValueError(
"If `do_train` is True, then `train_file` must be specified.")
if args.do_predict:
if not args.predict_file:
raise ValueError(
"If `do_predict` is True, then `predict_file` must be specified.")
if os.path.exists(args.output_dir) and os.listdir(args.output_dir) and args.do_train and not args.overwrite_output_dir:
raise ValueError("Output directory () already exists and is not empty.")
if not os.path.exists(args.output_dir):
os.makedirs(args.output_dir)
if args.local_rank not in [-1, 0]:
torch.distributed.barrier() # Make sure only the first process in distributed training will download model & vocab
tokenizer = BertTokenizer.from_pretrained(args.bert_model, do_lower_case=args.do_lower_case)
model = BertForQuestionAnswering.from_pretrained(args.bert_model)
if args.local_rank == 0:
torch.distributed.barrier()
if args.fp16:
model.half()
model.to(device)
if args.local_rank != -1:
model = torch.nn.parallel.DistributedDataParallel(model,
device_ids=[args.local_rank],
output_device=args.local_rank,
find_unused_parameters=True)
elif n_gpu > 1:
model = torch.nn.DataParallel(model)
if args.do_train:
if args.local_rank in [-1, 0]:
tb_writer = SummaryWriter()
# Prepare data loader
train_examples = read_squad_examples(
input_file=args.train_file, is_training=True, version_2_with_negative=args.version_2_with_negative)
cached_train_features_file = args.train_file+'_{0}_{1}_{2}_{3}'.format(
list(filter(None, args.bert_model.split('/'))).pop(), str(args.max_seq_length), str(args.doc_stride), str(args.max_query_length))
try:
with open(cached_train_features_file, "rb") as reader:
train_features = pickle.load(reader)
except:
train_features = convert_examples_to_features(
examples=train_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=True)
if args.local_rank == -1 or torch.distributed.get_rank() == 0:
logger.info(" Saving train features into cached file %s", cached_train_features_file)
with open(cached_train_features_file, "wb") as writer:
pickle.dump(train_features, writer)
all_input_ids = torch.tensor([f.input_ids for f in train_features], dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in train_features], dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in train_features], dtype=torch.long)
all_start_positions = torch.tensor([f.start_position for f in train_features], dtype=torch.long)
all_end_positions = torch.tensor([f.end_position for f in train_features], dtype=torch.long)
train_data = TensorDataset(all_input_ids, all_input_mask, all_segment_ids,
all_start_positions, all_end_positions)
if args.local_rank == -1:
train_sampler = RandomSampler(train_data)
else:
train_sampler = DistributedSampler(train_data)
train_dataloader = DataLoader(train_data, sampler=train_sampler, batch_size=args.train_batch_size)
num_train_optimization_steps = len(train_dataloader) // args.gradient_accumulation_steps * args.num_train_epochs
# if args.local_rank != -1:
# num_train_optimization_steps = num_train_optimization_steps // torch.distributed.get_world_size()
# Prepare optimizer
param_optimizer = list(model.named_parameters())
# hack to remove pooler, which is not used
# thus it produce None grad that break apex
param_optimizer = [n for n in param_optimizer if 'pooler' not in n[0]]
no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [
{'params': [p for n, p in param_optimizer if not any(nd in n for nd in no_decay)], 'weight_decay': 0.01},
{'params': [p for n, p in param_optimizer if any(nd in n for nd in no_decay)], 'weight_decay': 0.0}
]
if args.fp16:
try:
from apex.optimizers import FP16_Optimizer
from apex.optimizers import FusedAdam
except ImportError:
raise ImportError("Please install apex from https://www.github.com/nvidia/apex to use distributed and fp16 training.")
optimizer = FusedAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
bias_correction=False,
max_grad_norm=1.0)
if args.loss_scale == 0:
optimizer = FP16_Optimizer(optimizer, dynamic_loss_scale=True)
else:
optimizer = FP16_Optimizer(optimizer, static_loss_scale=args.loss_scale)
warmup_linear = WarmupLinearSchedule(warmup=args.warmup_proportion,
t_total=num_train_optimization_steps)
else:
optimizer = BertAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
warmup=args.warmup_proportion,
t_total=num_train_optimization_steps)
global_step = 0
logger.info("***** Running training *****")
logger.info(" Num orig examples = %d", len(train_examples))
logger.info(" Num split examples = %d", len(train_features))
logger.info(" Batch size = %d", args.train_batch_size)
logger.info(" Num steps = %d", num_train_optimization_steps)
model.train()
for epoch in trange(int(args.num_train_epochs), desc="Epoch"):
for step, batch in enumerate(tqdm(train_dataloader, desc="Iteration", disable=args.local_rank not in [-1, 0])):
if n_gpu == 1:
batch = tuple(t.to(device) for t in batch) # multi-gpu does scattering it-self
input_ids, input_mask, segment_ids, start_positions, end_positions = batch
loss = model(input_ids, segment_ids, input_mask, start_positions, end_positions)
if n_gpu > 1:
loss = loss.mean() # mean() to average on multi-gpu.
if args.gradient_accumulation_steps > 1:
loss = loss / args.gradient_accumulation_steps
if args.fp16:
optimizer.backward(loss)
else:
loss.backward()
if (step + 1) % args.gradient_accumulation_steps == 0:
if args.fp16:
# modify learning rate with special warm up BERT uses
# if args.fp16 is False, BertAdam is used and handles this automatically
lr_this_step = args.learning_rate * warmup_linear.get_lr(global_step, args.warmup_proportion)
for param_group in optimizer.param_groups:
param_group['lr'] = lr_this_step
optimizer.step()
optimizer.zero_grad()
global_step += 1
if args.local_rank in [-1, 0]:
tb_writer.add_scalar('lr', optimizer.get_lr()[0], global_step)
tb_writer.add_scalar('loss', loss.item(), global_step)
if args.do_train and (args.local_rank == -1 or torch.distributed.get_rank() == 0):
# Save a trained model, configuration and tokenizer
model_to_save = model.module if hasattr(model, 'module') else model # Only save the model it-self
# If we save using the predefined names, we can load using `from_pretrained`
output_model_file = os.path.join(args.output_dir, WEIGHTS_NAME)
output_config_file = os.path.join(args.output_dir, CONFIG_NAME)
torch.save(model_to_save.state_dict(), output_model_file)
model_to_save.config.to_json_file(output_config_file)
tokenizer.save_vocabulary(args.output_dir)
# Load a trained model and vocabulary that you have fine-tuned
model = BertForQuestionAnswering.from_pretrained(args.output_dir)
tokenizer = BertTokenizer.from_pretrained(args.output_dir, do_lower_case=args.do_lower_case)
# Good practice: save your training arguments together with the trained model
output_args_file = os.path.join(args.output_dir, 'training_args.bin')
torch.save(args, output_args_file)
else:
model = BertForQuestionAnswering.from_pretrained(args.bert_model)
model.to(device)
if args.do_predict and (args.local_rank == -1 or torch.distributed.get_rank() == 0):
eval_examples = read_squad_examples(
input_file=args.predict_file, is_training=False, version_2_with_negative=args.version_2_with_negative)
eval_features = convert_examples_to_features(
examples=eval_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=False)
logger.info("***** Running predictions *****")
logger.info(" Num orig examples = %d", len(eval_examples))
logger.info(" Num split examples = %d", len(eval_features))
logger.info(" Batch size = %d", args.predict_batch_size)
all_input_ids = torch.tensor([f.input_ids for f in eval_features], dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in eval_features], dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in eval_features], dtype=torch.long)
all_example_index = torch.arange(all_input_ids.size(0), dtype=torch.long)
eval_data = TensorDataset(all_input_ids, all_input_mask, all_segment_ids, all_example_index)
# Run prediction for full data
eval_sampler = SequentialSampler(eval_data)
eval_dataloader = DataLoader(eval_data, sampler=eval_sampler, batch_size=args.predict_batch_size)
model.eval()
all_results = []
logger.info("Start evaluating")
for input_ids, input_mask, segment_ids, example_indices in tqdm(eval_dataloader, desc="Evaluating", disable=args.local_rank not in [-1, 0]):
if len(all_results) % 1000 == 0:
logger.info("Processing example: %d" % (len(all_results)))
input_ids = input_ids.to(device)
input_mask = input_mask.to(device)
segment_ids = segment_ids.to(device)
with torch.no_grad():
batch_start_logits, batch_end_logits = model(input_ids, segment_ids, input_mask)
for i, example_index in enumerate(example_indices):
start_logits = batch_start_logits[i].detach().cpu().tolist()
end_logits = batch_end_logits[i].detach().cpu().tolist()
eval_feature = eval_features[example_index.item()]
unique_id = int(eval_feature.unique_id)
all_results.append(RawResult(unique_id=unique_id,
start_logits=start_logits,
end_logits=end_logits))
output_prediction_file = os.path.join(args.output_dir, "predictions.json")
output_nbest_file = os.path.join(args.output_dir, "nbest_predictions.json")
output_null_log_odds_file = os.path.join(args.output_dir, "null_odds.json")
write_predictions(eval_examples, eval_features, all_results,
args.n_best_size, args.max_answer_length,
args.do_lower_case, output_prediction_file,
output_nbest_file, output_null_log_odds_file, args.verbose_logging,
args.version_2_with_negative, args.null_score_diff_threshold)
def main(args):
device = torch.device(
"cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
n_gpu = torch.cuda.device_count()
logger.info("device: {}, n_gpu: {}, 16-bits training: {}".format(
device, n_gpu, args.fp16))
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
if n_gpu > 0:
torch.cuda.manual_seed_all(args.seed)
if args.gradient_accumulation_steps < 1:
raise ValueError(
"Invalid gradient_accumulation_steps parameter: {}, should be >= 1"
.format(args.gradient_accumulation_steps))
args.train_batch_size = args.train_batch_size // args.gradient_accumulation_steps
if not args.do_train and not args.do_eval:
raise ValueError(
"At least one of `do_train` or `do_eval` must be True.")
if args.do_train:
assert (args.train_file is not None) and (args.dev_file is not None)
if args.eval_test:
assert args.test_file is not None
else:
assert args.dev_file is not None
if not os.path.exists(args.output_dir):
os.makedirs(args.output_dir)
if args.do_train:
logger.addHandler(
logging.FileHandler(os.path.join(args.output_dir, "train.log"),
'w'))
else:
logger.addHandler(
logging.FileHandler(os.path.join(args.output_dir, "eval.log"),
'w'))
logger.info(args)
tokenizer = BertTokenizer.from_pretrained(args.model,
do_lower_case=args.do_lower_case)
# read query templates
query_templates = read_query_templates(normal_file=args.normal_file,
des_file=args.des_file)
if args.do_train or (not args.eval_test):
eval_examples = read_ace_examples(input_file=args.dev_file,
is_training=False)
gold_examples = read_ace_examples(input_file=args.gold_file,
is_training=False)
eval_features = convert_examples_to_features(
examples=eval_examples,
tokenizer=tokenizer,
query_templates=query_templates,
nth_query=args.nth_query,
is_training=False)
logger.info("***** Dev *****")
logger.info(" Num orig examples = %d", len(eval_examples))
logger.info(" Num split examples = %d", len(eval_features))
logger.info(" Batch size = %d", args.eval_batch_size)
all_input_ids = torch.tensor([f.input_ids for f in eval_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in eval_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in eval_features],
dtype=torch.long)
all_if_trigger_ids = torch.tensor(
[f.if_trigger_ids for f in eval_features], dtype=torch.long)
all_example_index = torch.arange(all_input_ids.size(0),
dtype=torch.long)
eval_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_if_trigger_ids,
all_example_index)
eval_dataloader = DataLoader(eval_data,
batch_size=args.eval_batch_size)
if args.do_train:
train_examples = read_ace_examples(input_file=args.train_file,
is_training=True)
train_features = convert_examples_to_features(
examples=train_examples,
tokenizer=tokenizer,
query_templates=query_templates,
nth_query=args.nth_query,
is_training=True)
if args.train_mode == 'sorted' or args.train_mode == 'random_sorted':
train_features = sorted(train_features,
key=lambda f: np.sum(f.input_mask))
else:
random.shuffle(train_features)
all_input_ids = torch.tensor([f.input_ids for f in train_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in train_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in train_features],
dtype=torch.long)
all_if_trigger_ids = torch.tensor(
[f.if_trigger_ids for f in train_features], dtype=torch.long)
all_start_positions = torch.tensor(
[f.start_position for f in train_features], dtype=torch.long)
all_end_positions = torch.tensor(
[f.end_position for f in train_features], dtype=torch.long)
train_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_if_trigger_ids,
all_start_positions, all_end_positions)
train_dataloader = DataLoader(train_data,
batch_size=args.train_batch_size)
train_batches = [batch for batch in train_dataloader]
num_train_optimization_steps = \
len(train_dataloader) // args.gradient_accumulation_steps * args.num_train_epochs
logger.info("***** Train *****")
logger.info(" Num orig examples = %d", len(train_examples))
logger.info(" Num split examples = %d", len(train_features))
logger.info(" Batch size = %d", args.train_batch_size)
logger.info(" Num steps = %d", num_train_optimization_steps)
eval_step = max(1, len(train_batches) // args.eval_per_epoch)
best_result = None
lrs = [args.learning_rate] if args.learning_rate else \
[1e-6, 2e-6, 3e-6, 5e-6, 1e-5, 2e-5, 3e-5, 5e-5]
for lr in lrs:
if not args.add_if_trigger_embedding:
model = BertForQuestionAnswering.from_pretrained(
args.model, cache_dir=PYTORCH_PRETRAINED_BERT_CACHE)
else:
model = BertForQuestionAnswering_withIfTriggerEmbedding.from_pretrained(
args.model, cache_dir=PYTORCH_PRETRAINED_BERT_CACHE)
if args.fp16:
model.half()
model.to(device)
if n_gpu > 1:
model = torch.nn.DataParallel(model)
param_optimizer = list(model.named_parameters())
param_optimizer = [
n for n in param_optimizer if 'pooler' not in n[0]
]
no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [{
'params': [
p for n, p in param_optimizer
if not any(nd in n for nd in no_decay)
],
'weight_decay':
0.01
}, {
'params': [
p for n, p in param_optimizer
if any(nd in n for nd in no_decay)
],
'weight_decay':
0.0
}]
optimizer = BertAdam(optimizer_grouped_parameters,
lr=lr,
warmup=args.warmup_proportion,
t_total=num_train_optimization_steps)
tr_loss = 0
nb_tr_examples = 0
nb_tr_steps = 0
global_step = 0
start_time = time.time()
for epoch in range(int(args.num_train_epochs)):
model.train()
logger.info("Start epoch #{} (lr = {})...".format(epoch, lr))
if args.train_mode == 'random' or args.train_mode == 'random_sorted':
random.shuffle(train_batches)
for step, batch in enumerate(train_batches):
if n_gpu == 1:
batch = tuple(t.to(device) for t in batch)
input_ids, input_mask, segment_ids, if_trigger_ids, start_positions, end_positions = batch
if not args.add_if_trigger_embedding:
loss = model(input_ids, segment_ids, input_mask,
start_positions, end_positions)
else:
loss = model(input_ids, segment_ids, if_trigger_ids,
input_mask, start_positions,
end_positions)
if n_gpu > 1:
loss = loss.mean()
if args.gradient_accumulation_steps > 1:
loss = loss / args.gradient_accumulation_steps
tr_loss += loss.item()
nb_tr_examples += input_ids.size(0)
nb_tr_steps += 1
loss.backward()
if (step + 1) % args.gradient_accumulation_steps == 0:
optimizer.step()
optimizer.zero_grad()
global_step += 1
if (step + 1) % eval_step == 0 or step == 0:
save_model = False
if args.do_eval:
# result, _, _ = evaluate(args, model, device, eval_dataset, eval_dataloader, eval_examples, eval_features)
result, preds = evaluate(args, model, device,
eval_dataloader,
eval_examples,
gold_examples,
eval_features)
# import ipdb; ipdb.set_trace()
model.train()
result['global_step'] = global_step
result['epoch'] = epoch
result['learning_rate'] = lr
result['batch_size'] = args.train_batch_size
if (best_result is
None) or (result[args.eval_metric] >
best_result[args.eval_metric]):
best_result = result
save_model = True
logger.info(
'Epoch: {}, Step: {} / {}, used_time = {:.2f}s, loss = {:.6f}'
.format(epoch, step + 1,
len(train_batches),
time.time() - start_time,
tr_loss / nb_tr_steps))
logger.info(
"!!! Best dev %s (lr=%s, epoch=%d): p_c: %.2f, r_c: %.2f, f1_c: %.2f, p_i: %.2f, r_i: %.2f, f1_i: %.2f, best_na_thresh: %.5f"
%
# logger.info("!!! Best dev %s (lr=%s, epoch=%d): p_c: %.2f, r_c: %.2f, f1_c: %.2f, best_na_thresh: %.10f" %
# (args.eval_metric, str(lr), epoch, result["prec_c"], result["recall_c"], result["f1_c"], result["best_na_thresh"]))
(args.eval_metric, str(lr), epoch,
result["prec_c"], result["recall_c"],
result["f1_c"], result["prec_i"],
result["recall_i"], result["f1_i"],
result["best_na_thresh"]))
else:
save_model = True
if (int(args.num_train_epochs) - epoch < 3 and
(step + 1) / len(train_batches) > 0.7
) or step == 0:
save_model = True
else:
save_model = False
if save_model:
model_to_save = model.module if hasattr(
model, 'module') else model
subdir = os.path.join(
args.output_dir,
"epoch{epoch}-step{step}".format(epoch=epoch,
step=step))
if not os.path.exists(subdir):
os.makedirs(subdir)
output_model_file = os.path.join(
subdir, WEIGHTS_NAME)
output_config_file = os.path.join(
subdir, CONFIG_NAME)
torch.save(model_to_save.state_dict(),
output_model_file)
model_to_save.config.to_json_file(
output_config_file)
tokenizer.save_vocabulary(subdir)
if best_result:
with open(
os.path.join(args.output_dir,
"eval_results.txt"),
"w") as writer:
for key in sorted(best_result.keys()):
writer.write(
"%s = %s\n" %
(key, str(best_result[key])))
if args.do_eval:
if args.eval_test:
eval_examples = read_ace_examples(input_file=args.test_file,
is_training=False)
gold_examples = read_ace_examples(input_file=args.gold_file,
is_training=False)
eval_features = convert_examples_to_features(
examples=eval_examples,
tokenizer=tokenizer,
query_templates=query_templates,
nth_query=args.nth_query,
is_training=False)
logger.info("***** Test *****")
logger.info(" Num orig examples = %d", len(eval_examples))
logger.info(" Num split examples = %d", len(eval_features))
logger.info(" Batch size = %d", args.eval_batch_size)
all_input_ids = torch.tensor([f.input_ids for f in eval_features],
dtype=torch.long)
all_input_mask = torch.tensor(
[f.input_mask for f in eval_features], dtype=torch.long)
all_segment_ids = torch.tensor(
[f.segment_ids for f in eval_features], dtype=torch.long)
all_if_trigger_ids = torch.tensor(
[f.if_trigger_ids for f in eval_features], dtype=torch.long)
all_example_index = torch.arange(all_input_ids.size(0),
dtype=torch.long)
eval_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_if_trigger_ids,
all_example_index)
eval_dataloader = DataLoader(eval_data,
batch_size=args.eval_batch_size)
if not args.add_if_trigger_embedding:
model = BertForQuestionAnswering.from_pretrained(args.model_dir)
else:
model = BertForQuestionAnswering_withIfTriggerEmbedding.from_pretrained(
args.model_dir)
if args.fp16:
model.half()
model.to(device)
result, preds = evaluate(args,
model,
device,
eval_dataloader,
eval_examples,
gold_examples,
eval_features,
pred_only=True)
with open(os.path.join(args.model_dir, "test_results.txt"),
"w") as writer:
for key in result:
writer.write("%s = %s\n" % (key, str(result[key])))
with open(os.path.join(args.model_dir, "arg_predictions.json"),
"w") as writer:
for key in preds:
writer.write(json.dumps(preds[key], default=int) + "\n")
def main(args):
device = torch.device(
"cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
n_gpu = torch.cuda.device_count()
logger.info("device: {}, n_gpu: {}, 16-bits training: {}".format(
device, n_gpu, args.fp16))
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
if n_gpu > 0:
torch.cuda.manual_seed_all(args.seed)
if args.gradient_accumulation_steps < 1:
raise ValueError(
"Invalid gradient_accumulation_steps parameter: {}, should be >= 1"
.format(args.gradient_accumulation_steps))
args.train_batch_size = \
args.train_batch_size // args.gradient_accumulation_steps
if not args.do_train and not args.do_eval:
raise ValueError(
"At least one of `do_train` or `do_eval` must be True.")
if args.do_train:
assert (args.train_file is not None) and (args.dev_file is not None)
if args.eval_test:
assert args.test_file is not None
else:
assert args.dev_file is not None
if not os.path.exists(args.output_dir):
os.makedirs(args.output_dir)
if args.do_train:
logger.addHandler(
logging.FileHandler(os.path.join(args.output_dir, "train.log"),
'w'))
else:
logger.addHandler(
logging.FileHandler(os.path.join(args.output_dir, "eval.log"),
'w'))
logger.info(args)
tokenizer = BertTokenizer.from_pretrained(args.model,
do_lower_case=args.do_lower_case)
if args.do_train or (not args.eval_test):
with gzip.GzipFile(args.dev_file, 'r') as reader:
content = reader.read().decode('utf-8').strip().split('\n')[1:]
eval_dataset = [json.loads(line) for line in content]
eval_examples = read_mrqa_examples(input_file=args.dev_file,
is_training=False)
eval_features = convert_examples_to_features(
examples=eval_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=False)
logger.info("***** Dev *****")
logger.info(" Num orig examples = %d", len(eval_examples))
logger.info(" Num split examples = %d", len(eval_features))
logger.info(" Batch size = %d", args.eval_batch_size)
all_input_ids = torch.tensor([f.input_ids for f in eval_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in eval_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in eval_features],
dtype=torch.long)
all_example_index = torch.arange(all_input_ids.size(0),
dtype=torch.long)
eval_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_example_index)
eval_dataloader = DataLoader(eval_data,
batch_size=args.eval_batch_size)
if args.do_train:
train_examples = read_mrqa_examples(input_file=args.train_file,
is_training=True)
train_features = convert_examples_to_features(
examples=train_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=True)
if args.train_mode == 'sorted' or args.train_mode == 'random_sorted':
train_features = sorted(train_features,
key=lambda f: np.sum(f.input_mask))
else:
random.shuffle(train_features)
all_input_ids = torch.tensor([f.input_ids for f in train_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in train_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in train_features],
dtype=torch.long)
all_start_positions = torch.tensor(
[f.start_position for f in train_features], dtype=torch.long)
all_end_positions = torch.tensor(
[f.end_position for f in train_features], dtype=torch.long)
train_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_start_positions,
all_end_positions)
train_dataloader = DataLoader(train_data,
batch_size=args.train_batch_size)
train_batches = [batch for batch in train_dataloader]
num_train_optimization_steps = \
len(train_dataloader) // args.gradient_accumulation_steps * args.num_train_epochs
logger.info("***** Train *****")
logger.info(" Num orig examples = %d", len(train_examples))
logger.info(" Num split examples = %d", len(train_features))
logger.info(" Batch size = %d", args.train_batch_size)
logger.info(" Num steps = %d", num_train_optimization_steps)
eval_step = max(1, len(train_batches) // args.eval_per_epoch)
best_result = None
lrs = [args.learning_rate] if args.learning_rate else [
1e-6, 2e-6, 3e-6, 5e-6, 1e-5, 2e-5, 3e-5, 5e-5
]
for lr in lrs:
model = BertForQuestionAnswering.from_pretrained(
args.model, cache_dir=PYTORCH_PRETRAINED_BERT_CACHE)
if args.fp16:
model.half()
model.to(device)
if n_gpu > 1:
model = torch.nn.DataParallel(model)
param_optimizer = list(model.named_parameters())
param_optimizer = [
n for n in param_optimizer if 'pooler' not in n[0]
]
no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [{
'params': [
p for n, p in param_optimizer
if not any(nd in n for nd in no_decay)
],
'weight_decay':
0.01
}, {
'params': [
p for n, p in param_optimizer
if any(nd in n for nd in no_decay)
],
'weight_decay':
0.0
}]
if args.fp16:
try:
from apex.optimizers import FP16_Optimizer
from apex.optimizers import FusedAdam
except ImportError:
raise ImportError(
"Please install apex from https://www.github.com/nvidia/apex"
"to use distributed and fp16 training.")
optimizer = FusedAdam(optimizer_grouped_parameters,
lr=lr,
bias_correction=False,
max_grad_norm=1.0)
if args.loss_scale == 0:
optimizer = FP16_Optimizer(optimizer,
dynamic_loss_scale=True)
else:
optimizer = FP16_Optimizer(
optimizer, static_loss_scale=args.loss_scale)
else:
optimizer = BertAdam(optimizer_grouped_parameters,
lr=lr,
warmup=args.warmup_proportion,
t_total=num_train_optimization_steps)
tr_loss = 0
nb_tr_examples = 0
nb_tr_steps = 0
global_step = 0
start_time = time.time()
for epoch in range(int(args.num_train_epochs)):
model.train()
logger.info("Start epoch #{} (lr = {})...".format(epoch, lr))
for step, batch in enumerate(train_batches):
if n_gpu == 1:
batch = tuple(t.to(device) for t in batch)
input_ids, input_mask, segment_ids, start_positions, end_positions = batch
loss = model(input_ids, segment_ids, input_mask,
start_positions, end_positions)
if n_gpu > 1:
loss = loss.mean()
if args.gradient_accumulation_steps > 1:
loss = loss / args.gradient_accumulation_steps
tr_loss += loss.item()
nb_tr_examples += input_ids.size(0)
nb_tr_steps += 1
if args.fp16:
optimizer.backward(loss)
else:
loss.backward()
if (step + 1) % args.gradient_accumulation_steps == 0:
if args.fp16:
lr_this_step = lr * \
warmup_linear(global_step/num_train_optimization_steps, args.warmup_proportion)
for param_group in optimizer.param_groups:
param_group['lr'] = lr_this_step
optimizer.step()
optimizer.zero_grad()
global_step += 1
if (step + 1) % eval_step == 0:
logger.info(
'Epoch: {}, Step: {} / {}, used_time = {:.2f}s, loss = {:.6f}'
.format(epoch, step + 1, len(train_dataloader),
time.time() - start_time,
tr_loss / nb_tr_steps))
save_model = False
if args.do_eval:
result, _, _ = \
evaluate(args, model, device, eval_dataset,
eval_dataloader, eval_examples, eval_features)
model.train()
result['global_step'] = global_step
result['epoch'] = epoch
result['learning_rate'] = lr
result['batch_size'] = args.train_batch_size
if (best_result is
None) or (result[args.eval_metric] >
best_result[args.eval_metric]):
best_result = result
save_model = True
logger.info(
"!!! Best dev %s (lr=%s, epoch=%d): %.2f" %
(args.eval_metric, str(lr), epoch,
result[args.eval_metric]))
else:
save_model = True
if save_model:
model_to_save = model.module if hasattr(
model, 'module') else model
output_model_file = os.path.join(
args.output_dir, WEIGHTS_NAME)
output_config_file = os.path.join(
args.output_dir, CONFIG_NAME)
torch.save(model_to_save.state_dict(),
output_model_file)
model_to_save.config.to_json_file(
output_config_file)
tokenizer.save_vocabulary(args.output_dir)
if best_result:
with open(
os.path.join(args.output_dir,
EVAL_FILE),
"w") as writer:
for key in sorted(best_result.keys()):
writer.write(
"%s = %s\n" %
(key, str(best_result[key])))
if args.do_eval:
if args.eval_test:
with gzip.GzipFile(args.test_file, 'r') as reader:
content = reader.read().decode('utf-8').strip().split('\n')[1:]
eval_dataset = [json.loads(line) for line in content]
eval_examples = read_mrqa_examples(input_file=args.test_file,
is_training=False)
eval_features = convert_examples_to_features(
examples=eval_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=False)
logger.info("***** Test *****")
logger.info(" Num orig examples = %d", len(eval_examples))
logger.info(" Num split examples = %d", len(eval_features))
logger.info(" Batch size = %d", args.eval_batch_size)
all_input_ids = torch.tensor([f.input_ids for f in eval_features],
dtype=torch.long)
all_input_mask = torch.tensor(
[f.input_mask for f in eval_features], dtype=torch.long)
all_segment_ids = torch.tensor(
[f.segment_ids for f in eval_features], dtype=torch.long)
all_example_index = torch.arange(all_input_ids.size(0),
dtype=torch.long)
eval_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_example_index)
eval_dataloader = DataLoader(eval_data,
batch_size=args.eval_batch_size)
model = BertForQuestionAnswering.from_pretrained(args.output_dir)
if args.fp16:
model.half()
model.to(device)
result, preds, nbest_preds = \
evaluate(args, model, device, eval_dataset,
eval_dataloader, eval_examples, eval_features)
with open(os.path.join(args.output_dir, PRED_FILE), "w") as writer:
writer.write(json.dumps(preds, indent=4) + "\n")
with open(os.path.join(args.output_dir, TEST_FILE), "w") as writer:
for key in sorted(result.keys()):
writer.write("%s = %s\n" % (key, str(result[key])))
def load(self):
parser = argparse.ArgumentParser()
## Required parameters
parser.add_argument(
"--bert_model",
default=None,
type=str,
required=True,
help=
"Bert pre-trained model selected in the list: bert-base-uncased, "
"bert-large-uncased, bert-base-cased, bert-base-multilingual, bert-base-chinese."
)
parser.add_argument(
"--output_dir",
default=None,
type=str,
required=True,
help=
"The output directory where the model checkpoints will be written."
)
## Other parameters
parser.add_argument(
"--train_file",
default=None,
type=str,
help="SQuAD json for training. E.g., train-v1.1.json")
parser.add_argument(
"--predict_file",
default=None,
type=str,
help=
"SQuAD json for predictions. E.g., dev-v1.1.json or test-v1.1.json"
)
parser.add_argument(
"--max_seq_length",
default=384,
type=int,
help=
"The maximum total input sequence length after WordPiece tokenization. Sequences "
"longer than this will be truncated, and sequences shorter than this will be padded."
)
parser.add_argument(
"--doc_stride",
default=128,
type=int,
help=
"When splitting up a long document into chunks, how much stride to take between chunks."
)
parser.add_argument(
"--max_query_length",
default=64,
type=int,
help=
"The maximum number of tokens for the question. Questions longer than this will "
"be truncated to this length.")
parser.add_argument('--do_lower_case',
action="store_true",
default=False,
help="Lowercase the input")
parser.add_argument("--do_train",
default=False,
action='store_true',
help="Whether to run training.")
parser.add_argument("--do_predict",
default=False,
action='store_true',
help="Whether to run eval on the dev set.")
parser.add_argument("--train_batch_size",
default=32,
type=int,
help="Total batch size for training.")
parser.add_argument("--predict_batch_size",
default=8,
type=int,
help="Total batch size for predictions.")
parser.add_argument("--learning_rate",
default=5e-5,
type=float,
help="The initial learning rate for Adam.")
parser.add_argument("--num_train_epochs",
default=3.0,
type=float,
help="Total number of training epochs to perform.")
parser.add_argument(
"--warmup_proportion",
default=0.1,
type=float,
help=
"Proportion of training to perform linear learning rate warmup for. E.g., 0.1 = 10% "
"of training.")
parser.add_argument(
"--n_best_size",
default=20,
type=int,
help=
"The total number of n-best predictions to generate in the nbest_predictions.json "
"output file.")
parser.add_argument(
"--max_answer_length",
default=30,
type=int,
help=
"The maximum length of an answer that can be generated. This is needed because the start "
"and end predictions are not conditioned on one another.")
parser.add_argument(
"--verbose_logging",
default=False,
action='store_true',
help=
"If true, all of the warnings related to data processing will be printed. "
"A number of warnings are expected for a normal SQuAD evaluation.")
parser.add_argument("--no_cuda",
default=False,
action='store_true',
help="Whether not to use CUDA when available")
parser.add_argument('--seed',
type=int,
default=42,
help="random seed for initialization")
parser.add_argument(
'--gradient_accumulation_steps',
type=int,
default=1,
help=
"Number of updates steps to accumulate before performing a backward/update pass."
)
parser.add_argument("--local_rank",
type=int,
default=-1,
help="local_rank for distributed training on gpus")
parser.add_argument(
'--optimize_on_cpu',
default=False,
action='store_true',
help=
"Whether to perform optimization and keep the optimizer averages on CPU"
)
parser.add_argument(
'--fp16',
default=False,
action='store_true',
help="Whether to use 16-bit float precision instead of 32-bit")
parser.add_argument(
'--loss_scale',
type=float,
default=128,
help=
'Loss scaling, positive power of 2 values can improve fp16 convergence.'
)
self.args = parser.parse_args([
'--bert_model=bert-base-cased', '--output_dir=./output',
'--do_predict', '--predict_file=./squad/dev-v1.1.json'
])
if self.args.local_rank == -1 or self.args.no_cuda:
self.device = torch.device("cuda" if torch.cuda.is_available()
and not self.args.no_cuda else "cpu")
n_gpu = torch.cuda.device_count()
else:
self.device = torch.device("cuda", self.args.local_rank)
n_gpu = 1
# Initializes the distributed backend which will take care of sychronizing nodes/GPUs
torch.distributed.init_process_group(backend='nccl')
if self.args.fp16:
logger.info(
"16-bits training currently not supported in distributed training"
)
self.args.fp16 = False # (see https://github.com/pytorch/pytorch/pull/13496)
logger.info(
"device: {} n_gpu: {}, distributed training: {}, 16-bits trainiing: {}"
.format(self.device, n_gpu, bool(self.args.local_rank != -1),
self.args.fp16))
if self.args.gradient_accumulation_steps < 1:
raise ValueError(
"Invalid gradient_accumulation_steps parameter: {}, should be >= 1"
.format(self.args.gradient_accumulation_steps))
self.args.train_batch_size = int(self.args.train_batch_size /
self.args.gradient_accumulation_steps)
random.seed(self.args.seed)
np.random.seed(self.args.seed)
torch.manual_seed(self.args.seed)
if n_gpu > 0:
torch.cuda.manual_seed_all(self.args.seed)
if not self.args.do_train and not self.args.do_predict:
raise ValueError(
"At least one of `do_train` or `do_predict` must be True.")
if self.args.do_train:
if not self.args.train_file:
raise ValueError(
"If `do_train` is True, then `train_file` must be specified."
)
if self.args.do_predict:
if not self.args.predict_file:
raise ValueError(
"If `do_predict` is True, then `predict_file` must be specified."
)
if os.path.exists(self.args.output_dir) and os.listdir(
self.args.output_dir):
raise ValueError(
"Output directory () already exists and is not empty.")
os.makedirs(self.args.output_dir, exist_ok=True)
self.tokenizer = BertTokenizer.from_pretrained(tokenizer_path)
train_examples = None
num_train_steps = None
if self.args.do_train:
train_examples = read_squad_examples(
input_file=self.args.train_file, is_training=True)
num_train_steps = int(
len(train_examples) / self.args.train_batch_size /
self.args.gradient_accumulation_steps *
self.args.num_train_epochs)
# Prepare model
self.model = BertForQuestionAnswering.from_pretrained(
pretrained_path,
cache_dir=PYTORCH_PRETRAINED_BERT_CACHE /
'distributed_{}'.format(self.args.local_rank))
self.model.bert.load_state_dict(
torch.load(finetuned_path, map_location='cpu'))
if self.args.fp16:
self.model.half()
self.model.to(self.device)
if self.args.local_rank != -1:
self.model = torch.nn.parallel.DistributedDataParallel(
self.model,
device_ids=[self.args.local_rank],
output_device=self.args.local_rank)
elif n_gpu > 1:
self.model = torch.nn.DataParallel(model)
self.model.eval()
def __init__(self):
# Hyperparameters
self.BERT_MODEL = "bert-base-uncased"
self.OUTPUT_DIR = "bert-model"
self.TRAIN_FILE = ""
self.PREDICT_FILE = "squad/test-pred.json"
self.MAX_SEQ_LENGTH = 384
self.DOC_STRIDE = 128
self.MAX_QUERY_LENGTH = 64
self.DO_TRAIN = False
self.DO_PREDICT = True
self.TRAIN_BATCH_SIZE = 12
self.PREDICT_BATCH_SIZE = 8
self.LEARNING_RATE = 3e-5
self.NUM_TRAIN_EPOCHS = 2.0
self.WARMUP_PROPORTION = 0.1
self.N_BEST_SIZE = 20
self.MAX_ANSWER_LENGTH = 30
self.VERBOSE_LOGGING = False
self.NO_CUDA = False
self.SEED = 42
self.GRADIENT_ACCUMULATION_STEPS = 1
self.DO_LOWER_CASE = True
self.LOCAL_RANK = -1
self.FP16 = False
self.LOSS_SCALE = 0
self.VERSION_2_WITH_NEGATIVE = True
self.NULL_SCORE_DIFF_THRESHOLD = 0.0
if self.LOCAL_RANK == -1 or self.NO_CUDA:
self.device = torch.device("cuda" if torch.cuda.is_available()
and not self.NO_CUDA else "cpu")
n_gpu = torch.cuda.device_count()
else:
torch.cuda.set_device(self.LOCAL_RANK)
self.device = torch.device("cuda", self.LOCAL_RANK)
n_gpu = 1
# Initializes the distributed backend which will take care of sychronizing nodes/GPUs
torch.distributed.init_process_group(backend='nccl')
logger.info(
"device: {} n_gpu: {}, distributed training: {}, 16-bits training: {}"
.format(self.device, n_gpu, bool(self.LOCAL_RANK != -1),
self.FP16))
if self.GRADIENT_ACCUMULATION_STEPS < 1:
raise ValueError(
"Invalid gradient_accumulation_steps parameter: {}, should be >= 1"
.format(self.GRADIENT_ACCUMULATION_STEPS))
self.TRAIN_BATCH_SIZE = self.TRAIN_BATCH_SIZE // self.GRADIENT_ACCUMULATION_STEPS
random.seed(self.SEED)
np.random.seed(self.SEED)
torch.manual_seed(self.SEED)
if n_gpu > 0:
torch.cuda.manual_seed_all(self.SEED)
if not self.DO_TRAIN and not self.DO_PREDICT:
raise ValueError(
"At least one of `do_train` or `do_predict` must be True.")
if self.DO_TRAIN:
if not self.TRAIN_FILE:
raise ValueError(
"If `do_train` is True, then `train_file` must be specified."
)
if self.DO_PREDICT:
if not self.PREDICT_FILE:
raise ValueError(
"If `do_predict` is True, then `predict_file` must be specified."
)
if os.path.exists(self.OUTPUT_DIR) and os.listdir(
self.OUTPUT_DIR) and self.DO_TRAIN:
raise ValueError(
"Output directory () already exists and is not empty.")
if not os.path.exists(self.OUTPUT_DIR):
os.makedirs(self.OUTPUT_DIR)
self.tokenizer = BertTokenizer.from_pretrained(
self.BERT_MODEL, do_lower_case=self.DO_LOWER_CASE)
train_examples = None
num_train_optimization_steps = None
# Prepare model
self.model = BertForQuestionAnswering.from_pretrained(
self.BERT_MODEL,
cache_dir=os.path.join(str(PYTORCH_PRETRAINED_BERT_CACHE),
'distributed_{}'.format(self.LOCAL_RANK)))
if self.FP16:
self.model.half()
self.model.to(self.device)
if self.LOCAL_RANK != -1:
try:
from apex.parallel import DistributedDataParallel as DDP
except ImportError:
raise ImportError(
"Please install apex from https://www.github.com/nvidia/apex to use distributed and fp16 training."
)
self.model = DDP(self.model)
elif n_gpu > 1:
self.model = torch.nn.DataParallel(self.model)
# Prepare optimizer
param_optimizer = list(self.model.named_parameters())
# hack to remove pooler, which is not used
# thus it produce None grad that break apex
param_optimizer = [n for n in param_optimizer if 'pooler' not in n[0]]
no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [{
'params': [
p for n, p in param_optimizer
if not any(nd in n for nd in no_decay)
],
'weight_decay':
0.01
}, {
'params':
[p for n, p in param_optimizer if any(nd in n for nd in no_decay)],
'weight_decay':
0.0
}]
if self.FP16:
try:
from apex.optimizers import FP16_Optimizer
from apex.optimizers import FusedAdam
except ImportError:
raise ImportError(
"Please install apex from https://www.github.com/nvidia/apex to use distributed and fp16 training."
)
optimizer = FusedAdam(optimizer_grouped_parameters,
lr=self.LEARNING_RATE,
bias_correction=False,
max_grad_norm=1.0)
if self.LOSS_SCALE == 0:
optimizer = self.FP16_Optimizer(optimizer,
dynamic_loss_scale=True)
else:
optimizer = self.FP16_Optimizer(
optimizer, static_loss_scale=self.LOSS_SCALE)
else:
optimizer = BertAdam(optimizer_grouped_parameters,
lr=self.LEARNING_RATE,
warmup=self.WARMUP_PROPORTION,
t_total=num_train_optimization_steps)
# self.model = BertForQuestionAnswering.from_pretrained(self.BERT_MODEL)
output_model_file = os.path.join(self.OUTPUT_DIR, WEIGHTS_NAME)
output_config_file = os.path.join(self.OUTPUT_DIR, CONFIG_NAME)
# Load a trained model and config that you have fine-tuned
config = BertConfig(output_config_file)
self.model = BertForQuestionAnswering(config)
if torch.cuda.is_available():
self.model.load_state_dict(torch.load(output_model_file))
else:
self.model.load_state_dict(
torch.load(output_model_file, map_location='cpu'))
self.model.to(self.device)
print('\n*** QA MODULE READY [1/3] ***\n')
import math
import os
import random
import pickle
from tqdm import tqdm, trange
import numpy as np
import torch
from torch.utils.data import TensorDataset, DataLoader, RandomSampler, SequentialSampler
from torch.utils.data.distributed import DistributedSampler
from pytorch_pretrained_bert.tokenization import whitespace_tokenize, BasicTokenizer, BertTokenizer
from pytorch_pretrained_bert.modeling import BertForQuestionAnswering
from pytorch_pretrained_bert.optimization import BertAdam, warmup_linear
from pytorch_pretrained_bert.file_utils import PYTORCH_PRETRAINED_BERT_CACHE
model = BertForQuestionAnswering.from_pretrained("bert-base-uncased")
device = "cuda:0"
model.train(device)
param_optimizer = list(model.named_parameters())
param_optimizer = [n for n in param_optimizer if 'pooler' not in n[0]]
no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [
{'params': [p for n, p in param_optimizer if not any(
nd in n for nd in no_decay)], 'weight_decay': 0.01},
{'params': [p for n, p in param_optimizer if any(nd in n for nd in no_decay)], 'weight_decay': 0.0}
]
optimizer = torch.optim.Adam(optimizer_grouped_parameters, lr=0.0001)
print(optimizer_grouped_parameters)
def __init__(self):
self.bert_model = "bert-base-uncased"
self.pretrained_model = "./pretrained"
self.output_dir = "./tmp"
self.max_seq_length = 384
self.doc_stride = 128
self.max_query_length = 64
self.do_predict = True
self.predict_batch_size = 8
self.n_best_size = 20
self.max_query_length = 30
self.verbose_logging = False
self.no_cuda = False
self.seed = 42
self.fp16 = False
self.local_rank = -1
self.max_answer_length = 30
self.do_lower_case = True
if self.local_rank == -1 or self.no_cuda:
self.device = torch.device("cuda" if torch.cuda.is_available()
and not self.no_cuda else "cpu")
n_gpu = torch.cuda.device_count()
else:
self.device = torch.device("cuda", self.local_rank)
n_gpu = 1
# Initializes the distributed backend which will take care of sychronizing nodes/GPUs
torch.distributed.init_process_group(backend='nccl')
if self.fp16:
logger.info(
"16-bits training currently not supported in distributed training"
)
self.fp16 = False # (see https://github.com/pytorch/pytorch/pull/13496)
logger.info(
"device: {} n_gpu: {}, distributed training: {}, 16-bits trainiing: {}"
.format(self.device, n_gpu, bool(self.local_rank != -1),
self.fp16))
random.seed(self.seed)
np.random.seed(self.seed)
torch.manual_seed(self.seed)
if n_gpu > 0:
torch.cuda.manual_seed_all(self.seed)
if os.path.exists(self.output_dir) and os.listdir(self.output_dir):
raise ValueError(
"Output directory () already exists and is not empty.")
os.makedirs(self.output_dir, exist_ok=True)
self.tokenizer = BertTokenizer.from_pretrained(self.bert_model)
# Prepare model
self.model = BertForQuestionAnswering.from_pretrained(
self.pretrained_model,
cache_dir=PYTORCH_PRETRAINED_BERT_CACHE /
'distributed_{}'.format(self.local_rank))
if self.fp16:
self.model.half()
self.model.to(self.device)
if self.local_rank != -1:
self.model = torch.nn.parallel.DistributedDataParallel(
self.model,
device_ids=[self.local_rank],
output_device=self.local_rank)
elif n_gpu > 1:
self.model = torch.nn.DataParallel(self.model)
self.model.eval()
def main():
os.environ['NO_ETAG'] = '1'
BertTokenizer.from_pretrained(OPTS.bert_model, cache_dir=OPTS.cache_dir)
BertForQuestionAnswering.from_pretrained(OPTS.bert_model,
cache_dir=OPTS.cache_dir)
def main():
parser = argparse.ArgumentParser()
## Required parameters
parser.add_argument("--bert_model", default=None, type=str, required=True,
help="Bert pre-trained model selected in the list: bert-base-uncased, "
"bert-large-uncased, bert-base-cased, bert-base-multilingual, bert-base-chinese.")
parser.add_argument("--output_dir", default=None, type=str, required=True,
help="The output directory where the model checkpoints will be written.")
## Other parameters
parser.add_argument("--train_file", default=None, type=str, help="SQuAD json for training. E.g., train-v1.1.json")
parser.add_argument("--predict_file", default=None, type=str,
help="SQuAD json for predictions. E.g., dev-v1.1.json or test-v1.1.json")
parser.add_argument("--max_seq_length", default=384, type=int,
help="The maximum total input sequence length after WordPiece tokenization. Sequences "
"longer than this will be truncated, and sequences shorter than this will be padded.")
parser.add_argument("--doc_stride", default=128, type=int,
help="When splitting up a long document into chunks, how much stride to take between chunks.")
parser.add_argument("--max_query_length", default=64, type=int,
help="The maximum number of tokens for the question. Questions longer than this will "
"be truncated to this length.")
parser.add_argument("--do_train", default=False, action='store_true', help="Whether to run training.")
parser.add_argument("--do_predict", default=False, action='store_true', help="Whether to run eval on the dev set.")
parser.add_argument("--train_batch_size", default=32, type=int, help="Total batch size for training.")
parser.add_argument("--predict_batch_size", default=8, type=int, help="Total batch size for predictions.")
parser.add_argument("--learning_rate", default=5e-5, type=float, help="The initial learning rate for Adam.")
parser.add_argument("--num_train_epochs", default=3.0, type=float,
help="Total number of training epochs to perform.")
parser.add_argument("--warmup_proportion", default=0.1, type=float,
help="Proportion of training to perform linear learning rate warmup for. E.g., 0.1 = 10% "
"of training.")
parser.add_argument("--n_best_size", default=20, type=int,
help="The total number of n-best predictions to generate in the nbest_predictions.json "
"output file.")
parser.add_argument("--max_answer_length", default=30, type=int,
help="The maximum length of an answer that can be generated. This is needed because the start "
"and end predictions are not conditioned on one another.")
parser.add_argument("--verbose_logging", default=False, action='store_true',
help="If true, all of the warnings related to data processing will be printed. "
"A number of warnings are expected for a normal SQuAD evaluation.")
parser.add_argument('--seed',
type=int,
default=42,
help="random seed for initialization")
parser.add_argument('--init_gradient_accumulation_steps',
type=int,
default=1,
help="Initial number of updates steps to accumulate before performing a backward/update pass.")
parser.add_argument('--target_gradient_accumulation_steps',
type=int,
default=1,
help="Target number of updates steps to accumulate before performing a backward/update pass.")
parser.add_argument('--accumulation_warmup_proportion',default=0.2, type=float,
help="Proportion of training to ramp up gradient_accumulation_steps for. E.g., 0.1 = 10% ")
parser.add_argument("--do_lower_case",
default=True,
action='store_true',
help="Whether to lower case the input text. True for uncased models, False for cased models.")
parser.add_argument('--step_per_log',
type=int, default=100,
help='Number of updates steps to log metrics.')
parser.add_argument('--fp16',
default=False,
action='store_true',
help="Whether to use 16-bit float precision instead of 32-bit")
parser.add_argument('--loss_scale',
type=float, default=0,
help="Loss scaling to improve fp16 numeric stability. Only used when fp16 set to True.\n"
"0 (default value): dynamic loss scaling.\n"
"Positive power of 2: static loss scaling value.\n")
args = parser.parse_args()
# get the Azure ML run object
run = Run.get_context()
comm = DistributedCommunicator(accumulation_step=args.init_gradient_accumulation_steps)
rank = comm.rank
local_rank = comm.local_rank
world_size = comm.world_size
torch.cuda.set_device(local_rank)
device = torch.device("cuda", local_rank)
is_master = rank == 0
logger.info("world size: {}, local rank: {}, global rank: {}, fp16: {}".format(world_size, local_rank, rank, args.fp16))
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
if not args.do_train and not args.do_predict:
raise ValueError("At least one of `do_train` or `do_predict` must be True.")
if args.do_train:
if not args.train_file:
raise ValueError(
"If `do_train` is True, then `train_file` must be specified.")
if args.do_predict:
if not args.predict_file:
raise ValueError(
"If `do_predict` is True, then `predict_file` must be specified.")
if os.path.exists(args.output_dir) and os.listdir(args.output_dir):
raise ValueError("Output directory () already exists and is not empty.")
os.makedirs(args.output_dir, exist_ok=True)
output_model_file = os.path.join(args.output_dir, "pytorch_model.bin")
tokenizer = BertTokenizer.from_pretrained(args.bert_model)
# Prepare model
model = BertForQuestionAnswering.from_pretrained(args.bert_model,
cache_dir=PYTORCH_PRETRAINED_BERT_CACHE / 'distributed_{}'.format(local_rank))
if args.fp16:
model.half()
model.to(device)
comm.register_model(model, args.fp16)
if args.do_train:
train_examples = read_squad_examples(input_file=args.train_file, is_training=True)
num_train_steps = int(len(train_examples) / args.train_batch_size * args.num_train_epochs)
param_optimizer = list(model.named_parameters())
# hack to remove pooler, which is not used
# thus it produce None grad that break apex
param_optimizer = [n for n in param_optimizer if 'pooler' not in n[0]]
no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [
{'params': [p for n, p in param_optimizer if not any(nd in n for nd in no_decay)], 'weight_decay': 0.01},
{'params': [p for n, p in param_optimizer if any(nd in n for nd in no_decay)], 'weight_decay': 0.0}
]
t_total = num_train_steps // world_size
if args.fp16:
try:
from apex.optimizers import FP16_Optimizer
from apex.optimizers import FusedAdam
except ImportError:
raise ImportError("Please install apex from https://www.github.com/nvidia/apex to run this.")
optimizer = FusedAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
bias_correction=False,
max_grad_norm=1.0)
if args.loss_scale == 0:
optimizer = FP16_Optimizer(optimizer, dynamic_loss_scale=True)
else:
optimizer = FP16_Optimizer(optimizer, static_loss_scale=args.loss_scale)
else:
optimizer = BertAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
warmup=args.warmup_proportion,
t_total=t_total)
if is_master:
run.log('lr', np.float(args.learning_rate))
cached_train_features_file = args.train_file+'_{0}_{1}_{2}_{3}'.format(
list(filter(None, args.bert_model.split('/'))).pop(), str(args.max_seq_length), str(args.doc_stride), str(args.max_query_length))
train_features = None
try:
with open(cached_train_features_file, "rb") as reader:
train_features = pickle.load(reader)
except:
train_features = convert_examples_to_features(
examples=train_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=True)
if rank == 0:
logger.info(" Saving train features into cached file %s", cached_train_features_file)
with open(cached_train_features_file, "wb") as writer:
pickle.dump(train_features, writer)
logger.info("***** Running training *****")
logger.info(" Num orig examples = %d", len(train_examples))
logger.info(" Num split examples = %d", len(train_features))
logger.info(" Batch size = %d", args.train_batch_size)
logger.info(" Num steps = %d", num_train_steps)
all_input_ids = torch.tensor([f.input_ids for f in train_features], dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in train_features], dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in train_features], dtype=torch.long)
all_start_positions = torch.tensor([f.start_position for f in train_features], dtype=torch.long)
all_end_positions = torch.tensor([f.end_position for f in train_features], dtype=torch.long)
train_data = TensorDataset(all_input_ids, all_input_mask, all_segment_ids,
all_start_positions, all_end_positions)
if world_size > 1:
train_sampler = DistributedSampler(train_data,num_replicas=world_size, rank=rank)
else:
train_sampler = RandomSampler(train_data)
train_dataloader = DataLoader(train_data, sampler=train_sampler, batch_size=args.train_batch_size)
global_step, tr_loss = 0, 0
model.train()
for _ in trange(int(args.num_train_epochs), desc="Epoch"):
for _, batch in enumerate(tqdm(train_dataloader, desc="Iteration")):
batch = tuple(t.to(device) for t in batch)
input_ids, input_mask, segment_ids, start_positions, end_positions = batch
loss = model(input_ids, segment_ids, input_mask, start_positions, end_positions)
tr_loss += loss.item()
if args.fp16:
optimizer.backward(loss)
else:
loss.backward()
global_step += 1
if comm.synchronize():
lr_this_step = args.learning_rate * warmup_linear(global_step/t_total, args.warmup_proportion)
for param_group in optimizer.param_groups:
param_group['lr'] = lr_this_step
optimizer.step()
model.zero_grad()
comm.set_accumulation_step(adjust_gradient_accumulation_steps(
global_step/t_total, args.init_gradient_accumulation_steps,
args.target_gradient_accumulation_steps, args.accumulation_warmup_proportion))
if is_master and (global_step + 1) % args.step_per_log == 0:
run.log('train_loss', np.float(tr_loss / args.step_per_log))
tr_loss = 0
if is_master:
# Save a trained model
torch.save(model.state_dict(), output_model_file)
if args.do_predict and is_master:
# Load a trained model that you have fine-tuned
model_state_dict = torch.load(output_model_file)
model = BertForQuestionAnswering.from_pretrained(args.bert_model, state_dict=model_state_dict)
model.to(device)
eval_examples = read_squad_examples(
input_file=args.predict_file, is_training=False)
eval_features = convert_examples_to_features(
examples=eval_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=False)
logger.info("***** Running predictions *****")
logger.info(" Num orig examples = %d", len(eval_examples))
logger.info(" Num split examples = %d", len(eval_features))
logger.info(" Batch size = %d", args.predict_batch_size)
all_input_ids = torch.tensor([f.input_ids for f in eval_features], dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in eval_features], dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in eval_features], dtype=torch.long)
all_example_index = torch.arange(all_input_ids.size(0), dtype=torch.long)
eval_data = TensorDataset(all_input_ids, all_input_mask, all_segment_ids, all_example_index)
# Run prediction for full data
eval_sampler = SequentialSampler(eval_data)
eval_dataloader = DataLoader(eval_data, sampler=eval_sampler, batch_size=args.predict_batch_size)
model.eval()
all_results = []
logger.info("Start evaluating")
for input_ids, input_mask, segment_ids, example_indices in tqdm(eval_dataloader, desc="Evaluating"):
if len(all_results) % 1000 == 0:
logger.info("Processing example: %d" % (len(all_results)))
input_ids = input_ids.to(device)
input_mask = input_mask.to(device)
segment_ids = segment_ids.to(device)
with torch.no_grad():
batch_start_logits, batch_end_logits = model(input_ids, segment_ids, input_mask)
for i, example_index in enumerate(example_indices):
start_logits = batch_start_logits[i].detach().cpu().tolist()
end_logits = batch_end_logits[i].detach().cpu().tolist()
eval_feature = eval_features[example_index.item()]
unique_id = int(eval_feature.unique_id)
all_results.append(RawResult(unique_id=unique_id,
start_logits=start_logits,
end_logits=end_logits))
output_prediction_file = os.path.join(args.output_dir, "predictions.json")
output_nbest_file = os.path.join(args.output_dir, "nbest_predictions.json")
write_predictions(eval_examples, eval_features, all_results,
args.n_best_size, args.max_answer_length,
args.do_lower_case, output_prediction_file,
output_nbest_file, args.verbose_logging)
with open(args.predict_file) as predict_file:
dataset_json = json.load(predict_file)
dataset = dataset_json['data']
with open(output_prediction_file) as prediction_file:
predictions = json.load(prediction_file)
result = evaluate(dataset, predictions)
for key in result.keys():
logger.info(" %s = %s", key, str(result[key]))
run.log('exact_match', result['exact_match'])
run.log('f1', result['f1'])
def train(args):
args.train_batch_size = int(args.train_batch_size /
args.gradient_accumulation_steps)
tokenizer = BertTokenizer.from_pretrained(
modelconfig.MODEL_ARCHIVE_MAP[args.bert_model])
train_examples = data_utils.read_squad_examples(os.path.join(
args.data_dir, "train.json"),
is_training=True)
num_train_steps = int(
len(train_examples) / args.train_batch_size /
args.gradient_accumulation_steps) * args.num_train_epochs
train_features = data_utils.convert_examples_to_features(
train_examples,
tokenizer,
args.max_seq_length,
args.doc_stride,
args.max_query_length,
is_training=True)
logger.info("***** Running training *****")
logger.info(" Num orig examples = %d", len(train_examples))
logger.info(" Num split examples = %d", len(train_features))
logger.info(" Batch size = %d", args.train_batch_size)
logger.info(" Num steps = %d", num_train_steps)
all_input_ids = torch.tensor([f.input_ids for f in train_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in train_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in train_features],
dtype=torch.long)
all_start_positions = torch.tensor(
[f.start_position for f in train_features], dtype=torch.long)
all_end_positions = torch.tensor([f.end_position for f in train_features],
dtype=torch.long)
train_data = TensorDataset(all_input_ids, all_segment_ids, all_input_mask,
all_start_positions, all_end_positions)
train_sampler = RandomSampler(train_data)
train_dataloader = DataLoader(train_data,
sampler=train_sampler,
batch_size=args.train_batch_size)
# >>>>> validation
if args.do_valid:
valid_examples = data_utils.read_squad_examples(os.path.join(
args.data_dir, "dev.json"),
is_training=True)
valid_features = data_utils.convert_examples_to_features(
valid_examples,
tokenizer,
args.max_seq_length,
args.doc_stride,
args.max_query_length,
is_training=True)
valid_all_input_ids = torch.tensor(
[f.input_ids for f in valid_features], dtype=torch.long)
valid_all_segment_ids = torch.tensor(
[f.segment_ids for f in valid_features], dtype=torch.long)
valid_all_input_mask = torch.tensor(
[f.input_mask for f in valid_features], dtype=torch.long)
valid_all_start_positions = torch.tensor(
[f.start_position for f in valid_features], dtype=torch.long)
valid_all_end_positions = torch.tensor(
[f.end_position for f in valid_features], dtype=torch.long)
valid_data = TensorDataset(valid_all_input_ids, valid_all_segment_ids,
valid_all_input_mask,
valid_all_start_positions,
valid_all_end_positions)
logger.info("***** Running validations *****")
logger.info(" Num orig examples = %d", len(valid_examples))
logger.info(" Num split examples = %d", len(valid_features))
logger.info(" Batch size = %d", args.train_batch_size)
valid_sampler = SequentialSampler(valid_data)
valid_dataloader = DataLoader(valid_data,
sampler=valid_sampler,
batch_size=args.train_batch_size)
best_valid_loss = float('inf')
valid_losses = []
# <<<<< end of validation declaration
if not args.bert_model.endswith(".pt"):
model = BertForQuestionAnswering.from_pretrained(
modelconfig.MODEL_ARCHIVE_MAP[args.bert_model])
else:
model = torch.load(args.bert_model)
if args.fp16:
model.half()
model.cuda()
# Prepare optimizer
param_optimizer = [(k, v) for k, v in model.named_parameters()
if v.requires_grad == True]
param_optimizer = [n for n in param_optimizer if 'pooler' not in n[0]]
no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [{
'params':
[p for n, p in param_optimizer if not any(nd in n for nd in no_decay)],
'weight_decay':
0.01
}, {
'params':
[p for n, p in param_optimizer if any(nd in n for nd in no_decay)],
'weight_decay':
0.0
}]
t_total = num_train_steps
if args.fp16:
try:
from apex.optimizers import FP16_Optimizer
from apex.optimizers import FusedAdam
except ImportError:
raise ImportError(
"Please install apex from https://www.github.com/nvidia/apex to use distributed and fp16 training."
)
optimizer = FusedAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
bias_correction=False,
max_grad_norm=1.0)
if args.loss_scale == 0:
optimizer = FP16_Optimizer(optimizer, dynamic_loss_scale=True)
else:
optimizer = FP16_Optimizer(optimizer,
static_loss_scale=args.loss_scale)
else:
optimizer = BertAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
warmup=args.warmup_proportion,
t_total=t_total)
global_step = 0
model.train()
for _ in range(args.num_train_epochs):
for step, batch in enumerate(train_dataloader):
batch = tuple(t.cuda() for t in batch)
input_ids, segment_ids, input_mask, start_positions, end_positions = batch
loss = model(input_ids, segment_ids, input_mask, start_positions,
end_positions)
if args.gradient_accumulation_steps > 1:
loss = loss / args.gradient_accumulation_steps
if args.fp16:
optimizer.backward(loss)
else:
loss.backward()
if (step + 1) % args.gradient_accumulation_steps == 0:
# modify learning rate with special warm up BERT uses
lr_this_step = args.learning_rate * warmup_linear(
global_step / t_total, args.warmup_proportion)
for param_group in optimizer.param_groups:
param_group['lr'] = lr_this_step
optimizer.step()
optimizer.zero_grad()
global_step += 1
# >>>> perform validation at the end of each epoch .
if args.do_valid:
model.eval()
with torch.no_grad():
losses = []
valid_size = 0
for step, batch in enumerate(valid_dataloader):
batch = tuple(
t.cuda()
for t in batch) # multi-gpu does scattering it-self
input_ids, segment_ids, input_mask, start_positions, end_positions = batch
loss = model(input_ids, segment_ids, input_mask,
start_positions, end_positions)
losses.append(loss.data.item() * input_ids.size(0))
valid_size += input_ids.size(0)
valid_loss = sum(losses) / valid_size
logger.info("validation loss: %f", valid_loss)
valid_losses.append(valid_loss)
if valid_loss < best_valid_loss:
torch.save(model, os.path.join(args.output_dir, "model.pt"))
best_valid_loss = valid_loss
model.train()
if args.do_valid:
with open(os.path.join(args.output_dir, "valid.json"), "w") as fw:
json.dump({"valid_losses": valid_losses}, fw)
else:
torch.save(model, os.path.join(args.output_dir, "model.pt"))
def main():
args = set_config()
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
tokenizer = BertTokenizer.from_pretrained(args.bert_model,
do_lower_case=args.do_lower_case)
# Prepare model
model = BertForQuestionAnswering.from_pretrained(args.bert_model)
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
model.to(device)
optimizer = torch.optim.Adam(model.parameters(), lr=args.learning_rate)
global_step = 0
if args.do_train:
# load train data
train_examples, train_features = get_train_feature(
args, args.do_train, tokenizer)
for f in train_features:
print(len(f.doc_input_ids))
train_data = DataIteratorPack(train_features,
train_examples,
args.train_batch_size,
device,
sent_limit=25,
sequential=False)
# load dev data
eval_examples, eval_features = get_train_feature(
args, not args.do_train, tokenizer)
eval_data = DataIteratorPack(eval_features,
eval_examples,
args.predict_batch_size,
device,
sent_limit=25,
sequential=False)
logger.info("***** Running training *****")
logger.info(" Num split examples = %d", len(train_features))
logger.info(" Batch size = %d", args.train_batch_size)
total_train_loss = 0
VERBOSE_STEP = 100
grad_accumulate_step = 1
best_dev_F1 = None
for epoch in trange(int(args.num_train_epochs), desc="Epoch"):
model.train()
# learning rate decay
# if epoch > 1:
# args.learning_rate = args.learning_rate * args.decay
# for param_group in optimizer.param_groups:
# param_group['lr'] = args.learning_rate
# print('lr = {}'.format(args.learning_rate))
for step, batch in enumerate(train_data):
# batch = tuple(t.to(device) for t in batch) # multi-gpu does scattering it-self
input_ids = batch["context_idxs"]
input_mask = batch["context_mask"]
segment_ids = batch["segment_idxs"]
start_positions = batch["y1"]
end_positions = batch["y2"]
q_types = batch["q_type"]
loss = model(input_ids,
segment_ids,
input_mask,
start_positions,
end_positions,
q_types,
batch=batch)
if args.gradient_accumulation_steps > 1:
loss = loss / args.gradient_accumulation_steps
loss.backward()
if (global_step + 1) % grad_accumulate_step == 0:
optimizer.step()
optimizer.zero_grad()
total_train_loss += loss
global_step += 1
if global_step % VERBOSE_STEP == 0:
print("-- In Epoch{}: ".format(epoch))
print("Avg-LOSS: {}/batch/step: {}".format(
total_train_loss / VERBOSE_STEP,
global_step / VERBOSE_STEP))
total_train_loss = 0
# Save a trained model
model_to_save = model.module if hasattr(
model, 'module') else model # Only save the model it-self
train_data.refresh()
if args.do_predict:
eval_examples_dict = example_dict(eval_examples)
# eval_features_dict = example_dict(eval_features)
logger.info("***** Running predictions *****")
logger.info(" Num split examples = %d", len(eval_features))
logger.info(" Batch size = %d", args.predict_batch_size)
model.eval()
all_results = []
sp_dict = {}
logger.info("Start evaluating")
for step, batch in enumerate(eval_data):
# batch = tuple(t.to(device) for t in batch) # multi-gpu does scattering it-self
input_ids = batch["context_idxs"]
input_mask = batch["context_mask"]
segment_ids = batch["segment_idxs"]
if len(all_results) % 1000 == 0:
logger.info("Processing example: %d" %
(len(all_results)))
with torch.no_grad():
batch_start_logits, batch_end_logits, batch_types, sp = model(
input_ids, segment_ids, input_mask, batch=batch)
for i, example_index in enumerate(batch["ids"]):
start_logits = batch_start_logits[i].detach().cpu(
).tolist()
end_logits = batch_end_logits[i].detach().cpu().tolist(
)
# eval_feature = eval_features[example_index.item()]
unique_id = batch['unique_id'][i]
print(unique_id)
types = batch_types[i].detach().cpu().tolist()
all_results.append(
RawResult(unique_id=unique_id,
start_logits=start_logits,
end_logits=end_logits,
types=types))
predict_support_np = torch.sigmoid(
sp[:, :, 1]).data.cpu().numpy()
for i in range(predict_support_np.shape[0]):
cur_sp_pred = []
cur_id = batch['ids'][i]
for j in range(predict_support_np.shape[1]):
if j >= len(eval_examples_dict[cur_id].sent_names):
break
if predict_support_np[i, j] > args.sp_threshold:
cur_sp_pred.append(
eval_examples_dict[cur_id].sent_names[j])
sp_dict.update({cur_id: cur_sp_pred})
answer_dict = convert_to_tokens(eval_examples, eval_features,
all_results)
metrics = evaluate(eval_examples_dict, answer_dict)
print('hotpotqa epoch {:3d} | EM {:.4f} | F1 {:.4f}'.format(
epoch, metrics['exact_match'], metrics['f1']))
output_prediction_file = os.path.join(
args.output_dir, "predictions_{}.json".format(epoch))
output_nbest_file = os.path.join(
args.output_dir, "nbest_predictions_{}.json".format(epoch))
output_null_log_odds_file = os.path.join(
args.output_dir, "null_odds_{}.json".format(epoch))
all_predictions = write_predictions(
eval_examples, eval_features, all_results,
args.n_best_size, args.max_answer_length,
args.do_lower_case, output_prediction_file,
output_nbest_file, output_null_log_odds_file,
args.verbose_logging, args.version_2_with_negative,
args.null_score_diff_threshold)
metrics = evaluate(eval_examples_dict, all_predictions)
print('squad epoch {:3d} | EM {:.4f} | F1 {:.4f}'.format(
epoch, metrics['exact_match'], metrics['f1']))
dev_F1 = metrics['f1']
#learning rate decay
if best_dev_F1 is None or dev_F1 > best_dev_F1:
best_dev_F1 = dev_F1
output_model_file = os.path.join(args.output_dir,
"pytorch_model.bin")
torch.save(model_to_save.state_dict(), output_model_file)
cur_patience = 0
else:
cur_patience += 1
if cur_patience >= args.patience:
for param_group in optimizer.param_groups:
param_group['lr'] /= 2.0
if param_group['lr'] < 1e-6:
stop_train = True
break
cur_patience = 0
'Neopterin as an indicator of immune activation and prognosis in patients with gynecological malignancies.',
text_b=None,
label='1')
]
use_cuda = False
max_seq_length = 50
device = torch.device("cuda") if (use_cuda) else torch.device("cpu")
bert_model = 'bert-base-uncased'
cache_dir = '/home/dpappas/bert_cache/'
bert_tokenizer = BertTokenizer.from_pretrained(bert_model,
do_lower_case=True,
cache_dir=cache_dir)
bert_model = BertForQuestionAnswering.from_pretrained(
bert_model,
cache_dir=PYTORCH_PRETRAINED_BERT_CACHE /
'distributed_{}'.format(-1)).to(device)
eval_features = convert_examples_to_features(eval_examples, max_seq_length,
bert_tokenizer)
eval_feat = eval_features[0]
input_ids = torch.tensor([eval_feat.input_ids], dtype=torch.long).to(device)
input_mask = torch.tensor([eval_feat.input_mask], dtype=torch.long).to(device)
segment_ids = torch.tensor([eval_feat.segment_ids],
dtype=torch.long).to(device)
tokens = eval_feat.tokens
token_embeds, pooled_output = bert_model.bert(input_ids,
segment_ids,
input_mask,
output_all_encoded_layers=False)
def main(args=None):
if args is None:
args = model_utils.run_redundancy_span_get_local_args()
if args.local_rank == -1 or args.no_cuda:
device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
n_gpu = torch.cuda.device_count()
else:
torch.cuda.set_device(args.local_rank)
device = torch.device("cuda", args.local_rank)
n_gpu = 1
# Initializes the distributed backend which will take care of sychronizing nodes/GPUs
torch.distributed.init_process_group(backend='nccl')
# logger.info("device: {} n_gpu: {}, distributed training: {}, 16-bits training: {}".format(
# device, n_gpu, bool(args.local_rank != -1), args.fp16))
if args.gradient_accumulation_steps < 1:
raise ValueError("Invalid gradient_accumulation_steps parameter: {}, should be >= 1".format(args.gradient_accumulation_steps))
args.train_batch_size = int(args.train_batch_size / args.gradient_accumulation_steps)
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
if n_gpu > 0:
torch.cuda.manual_seed_all(args.seed)
if not args.do_train and not args.do_predict:
raise ValueError("At least one of `do_train` or `do_predict` must be True.")
if args.do_train:
if not args.train_file:
raise ValueError("If `do_train` is True, then `train_file` must be specified.")
if args.do_predict:
if not args.predict_file:
raise ValueError("If `do_predict` is True, then `predict_file` must be specified.")
if os.path.exists(args.output_dir) and os.listdir(args.output_dir):
raise ValueError("Output directory () already exists and is not empty.")
os.makedirs(args.output_dir, exist_ok=True)
tokenizer = BertTokenizer.from_pretrained(args.bert_model)
train_examples = None
num_train_steps = None
if args.do_train:
train_examples = read_many_examples(input_file=args.train_file, is_training=True)
num_train_steps = int(len(train_examples) / args.train_batch_size / args.gradient_accumulation_steps * args.num_train_epochs)
# Prepare model
"""=========fine-tune 1==========="""
model = BertForQuestionAnswering.from_pretrained(args.bert_model,
cache_dir=PYTORCH_PRETRAINED_BERT_CACHE / 'distributed_{}'.format(args.local_rank))
print (PYTORCH_PRETRAINED_BERT_CACHE / 'distributed_{}'.format(args.local_rank))
if args.fp16:
model.half()
model.to(device)
"""=============================="""
if args.local_rank != -1:
try:
from apex.parallel import DistributedDataParallel as DDP
except ImportError:
raise ImportError("Please install apex from https://www.github.com/nvidia/apex to use distributed and fp16 training.")
model = DDP(model)
elif n_gpu > 1:
model = torch.nn.DataParallel(model)
# Prepare optimizer
param_optimizer = list(model.named_parameters())
# hack to remove pooler, which is not used
# thus it produce None grad that break apex
param_optimizer = [n for n in param_optimizer if 'pooler' not in n[0]]
no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [
{'params': [p for n, p in param_optimizer if not any(nd in n for nd in no_decay)], 'weight_decay': 0.01},
{'params': [p for n, p in param_optimizer if any(nd in n for nd in no_decay)], 'weight_decay': 0.0}]
t_total = num_train_steps
if args.local_rank != -1:
t_total = t_total // torch.distributed.get_world_size()
if args.fp16:
try:
from apex.optimizers import FP16_Optimizer
from apex.optimizers import FusedAdam
except ImportError:
raise ImportError("Please install apex from https://www.github.com/nvidia/apex to use distributed and fp16 training.")
optimizer = FusedAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
bias_correction=False,
max_grad_norm=1.0)
if args.loss_scale == 0:
optimizer = FP16_Optimizer(optimizer, dynamic_loss_scale=True)
else:
optimizer = FP16_Optimizer(optimizer, static_loss_scale=args.loss_scale)
else:
optimizer = BertAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
warmup=args.warmup_proportion,
t_total=t_total)
max_eval_accuracy = 0
global_step = 0
if args.do_train:
cached_train_features_file = args.train_file + '_{0}_{1}_{2}_{3}'.format(
args.bert_model, str(args.max_seq_length), str(args.doc_stride), str(args.max_query_length))
train_features = None
try:
with open(cached_train_features_file, "rb") as reader:
train_features = pickle.load(reader)
except:
train_features = convert_examples_to_features(
examples=train_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=True)
if args.local_rank == -1 or torch.distributed.get_rank() == 0:
# logger.info(" Saving train features into cached file %s", cached_train_features_file)
with open(cached_train_features_file, "wb") as writer:
pickle.dump(train_features, writer)
all_input_ids = torch.tensor([f.input_ids for f in train_features], dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in train_features], dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in train_features], dtype=torch.long)
all_start_positions = torch.tensor([f.start_position for f in train_features], dtype=torch.long)
all_end_positions = torch.tensor([f.end_position for f in train_features], dtype=torch.long)
train_data = TensorDataset(all_input_ids, all_input_mask, all_segment_ids, all_start_positions, all_end_positions)
if args.local_rank == -1:
train_sampler = RandomSampler(train_data)
else:
train_sampler = DistributedSampler(train_data)
train_dataloader = DataLoader(train_data, sampler=train_sampler, batch_size=args.train_batch_size)
if args.do_predict and (args.local_rank == -1 or torch.distributed.get_rank() == 0):
eval_examples = read_many_examples(input_file=args.predict_file, is_training=False)
eval_features = convert_examples_to_features(
examples=eval_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=False)
all_input_ids = torch.tensor([f.input_ids for f in eval_features], dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in eval_features], dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in eval_features], dtype=torch.long)
# all_start_positions = torch.tensor([f.start_position for f in train_features], dtype=torch.long)
# all_end_positions = torch.tensor([f.end_position for f in train_features], dtype=torch.long)
all_example_index = torch.arange(all_input_ids.size(0), dtype=torch.long)
eval_data = TensorDataset(all_input_ids, all_input_mask, all_segment_ids, all_example_index)
# Run prediction for full data
eval_sampler = SequentialSampler(eval_data)
eval_dataloader = DataLoader(eval_data, sampler=eval_sampler, batch_size=args.predict_batch_size)
for _ in trange(int(args.num_train_epochs), desc="Epoch"):
model.train()
for step, batch in enumerate(tqdm(train_dataloader, desc="Iteration")):
if n_gpu == 1:
batch = tuple(t.to(device) for t in batch) # multi-gpu does scattering it-self
input_ids, input_mask, segment_ids, start_positions, end_positions = batch
loss = model(input_ids=input_ids, token_type_ids=segment_ids,
attention_mask=input_mask, start_positions=start_positions, end_positions=end_positions)
if n_gpu > 1:
loss = loss.mean() # mean() to average on multi-gpu.
if args.gradient_accumulation_steps > 1:
loss = loss / args.gradient_accumulation_steps
if args.fp16:
optimizer.backward(loss)
else:
loss.backward(retain_graph=True)
if (step + 1) % args.gradient_accumulation_steps == 0:
# modify learning rate with special warm up BERT uses
lr_this_step = args.learning_rate * span_utils.warmup_linear(global_step / t_total, args.warmup_proportion)
for param_group in optimizer.param_groups:
param_group['lr'] = lr_this_step
optimizer.step()
optimizer.zero_grad()
global_step += 1
if args.do_predict and (args.local_rank == -1 or torch.distributed.get_rank() == 0):
model.eval()
all_results = []
# logger.info("Start evaluating")
for input_ids, input_mask, segment_ids, example_indices, in tqdm(eval_dataloader, desc="Evaluating"):
# if len(all_results) % 1000 == 0: logger.info("Processing example: %d" % (len(all_results)))
input_ids = input_ids.to(device)
input_mask = input_mask.to(device)
segment_ids = segment_ids.to(device)
with torch.no_grad():
batch_start_logits, batch_end_logits = model(input_ids=input_ids, token_type_ids=segment_ids, attention_mask=input_mask)
for i, example_index in enumerate(example_indices):
start_logits = batch_start_logits[i].detach().cpu().tolist()
end_logits = batch_end_logits[i].detach().cpu().tolist()
eval_feature = eval_features[example_index.item()]
# unique_id = int(eval_feature.unique_id)
all_results.append(RawResult(unique_id=int(eval_feature.unique_id), start_logits=start_logits, end_logits=end_logits))
eval_accuracy,instance_num, all_predictions = write_predictions(eval_examples, eval_features,
all_results, args.n_best_size, args.max_answer_length,
args.do_lower_case, args.verbose_logging)
if max_eval_accuracy < eval_accuracy:
max_eval_accuracy = eval_accuracy
# Save a trained model
model_to_save = model.module if hasattr(model, 'module') else model # Only save the model it-self
output_model_file = os.path.join(args.output_dir, "pytorch_model.bin")
torch.save(model_to_save.state_dict(), output_model_file)
output_prediction_file = os.path.join(args.output_dir, "predictions.json")
output_nbest_file = os.path.join(args.output_dir, "nbest_predictions.json")
with open(output_prediction_file, "w") as writer:
writer.write(json.dumps(all_predictions, indent=4) + "\n")
# with open(output_nbest_file, "w") as writer:
# writer.write(json.dumps(all_nbest_json, indent=4) + "\n")
result_json = collections.OrderedDict()
result_json['result'] = eval_accuracy
result_json['eval_accuracy'] = eval_accuracy
result_json['instance_num'] = instance_num
with open(output_nbest_file, "w") as writer:
writer.write(json.dumps(result_json, indent=4) + "\n")
def main():
parser = argparse.ArgumentParser()
## Required parameters
parser.add_argument(
"--bert_model",
default=None,
type=str,
required=True,
help="Bert pre-trained model selected in the list: bert-base-uncased, "
"bert-large-uncased, bert-base-cased, bert-base-multilingual, bert-base-chinese."
)
parser.add_argument(
"--output_dir",
default=None,
type=str,
required=True,
help="The output directory where the model checkpoints will be written."
)
## Other parameters
parser.add_argument("--train_file",
default=None,
type=str,
help="SQuAD json for training. E.g., train-v1.1.json")
parser.add_argument(
"--predict_file",
default=None,
type=str,
help="SQuAD json for predictions. E.g., dev-v1.1.json or test-v1.1.json"
)
parser.add_argument(
"--max_seq_length",
default=384,
type=int,
help=
"The maximum total input sequence length after WordPiece tokenization. Sequences "
"longer than this will be truncated, and sequences shorter than this will be padded."
)
parser.add_argument(
"--doc_stride",
default=128,
type=int,
help=
"When splitting up a long document into chunks, how much stride to take between chunks."
)
parser.add_argument(
"--max_query_length",
default=64,
type=int,
help=
"The maximum number of tokens for the question. Questions longer than this will "
"be truncated to this length.")
parser.add_argument("--do_train",
default=False,
action='store_true',
help="Whether to run training.")
parser.add_argument("--do_predict",
default=False,
action='store_true',
help="Whether to run eval on the dev set.")
parser.add_argument("--train_batch_size",
default=32,
type=int,
help="Total batch size for training.")
parser.add_argument("--predict_batch_size",
default=8,
type=int,
help="Total batch size for predictions.")
parser.add_argument("--learning_rate",
default=5e-5,
type=float,
help="The initial learning rate for Adam.")
parser.add_argument("--num_train_epochs",
default=3.0,
type=float,
help="Total number of training epochs to perform.")
parser.add_argument(
"--warmup_proportion",
default=0.1,
type=float,
help=
"Proportion of training to perform linear learning rate warmup for. E.g., 0.1 = 10% "
"of training.")
parser.add_argument(
"--n_best_size",
default=20,
type=int,
help=
"The total number of n-best predictions to generate in the nbest_predictions.json "
"output file.")
parser.add_argument(
"--max_answer_length",
default=30,
type=int,
help=
"The maximum length of an answer that can be generated. This is needed because the start "
"and end predictions are not conditioned on one another.")
parser.add_argument(
"--verbose_logging",
default=False,
action='store_true',
help=
"If true, all of the warnings related to data processing will be printed. "
"A number of warnings are expected for a normal SQuAD evaluation.")
parser.add_argument("--no_cuda",
default=False,
action='store_true',
help="Whether not to use CUDA when available")
parser.add_argument('--seed',
type=int,
default=42,
help="random seed for initialization")
parser.add_argument(
'--gradient_accumulation_steps',
type=int,
default=1,
help=
"Number of updates steps to accumulate before performing a backward/update pass."
)
parser.add_argument(
"--do_lower_case",
default=True,
action='store_true',
help=
"Whether to lower case the input text. True for uncased models, False for cased models."
)
parser.add_argument("--local_rank",
type=int,
default=-1,
help="local_rank for distributed training on gpus")
parser.add_argument(
'--optimize_on_cpu',
default=False,
action='store_true',
help=
"Whether to perform optimization and keep the optimizer averages on CPU"
)
parser.add_argument(
'--fp16',
default=False,
action='store_true',
help="Whether to use 16-bit float precision instead of 32-bit")
parser.add_argument(
'--loss_scale',
type=float,
default=128,
help=
'Loss scaling, positive power of 2 values can improve fp16 convergence.'
)
args = parser.parse_args()
eval_period = 1200
if args.local_rank == -1 or args.no_cuda:
device = torch.device("cuda" if torch.cuda.is_available()
and not args.no_cuda else "cpu")
n_gpu = torch.cuda.device_count()
else:
device = torch.device("cuda", args.local_rank)
n_gpu = 1
# Initializes the distributed backend which will take care of sychronizing nodes/GPUs
torch.distributed.init_process_group(backend='nccl')
if args.fp16:
logger.info(
"16-bits training currently not supported in distributed training"
)
args.fp16 = False # (see https://github.com/pytorch/pytorch/pull/13496)
logger.info(
"device: {} n_gpu: {}, distributed training: {}, 16-bits trainiing: {}"
.format(device, n_gpu, bool(args.local_rank != -1), args.fp16))
if args.gradient_accumulation_steps < 1:
raise ValueError(
"Invalid gradient_accumulation_steps parameter: {}, should be >= 1"
.format(args.gradient_accumulation_steps))
args.train_batch_size = int(args.train_batch_size /
args.gradient_accumulation_steps)
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
if n_gpu > 0:
torch.cuda.manual_seed_all(args.seed)
if not args.do_train and not args.do_predict:
raise ValueError(
"At least one of `do_train` or `do_predict` must be True.")
if args.do_train:
if not args.train_file:
raise ValueError(
"If `do_train` is True, then `train_file` must be specified.")
if args.do_predict:
if not args.predict_file:
raise ValueError(
"If `do_predict` is True, then `predict_file` must be specified."
)
if os.path.exists(args.output_dir) and os.listdir(args.output_dir):
raise ValueError(
"Output directory () already exists and is not empty.")
os.makedirs(args.output_dir, exist_ok=True)
tokenizer = BertTokenizer.from_pretrained(args.bert_model)
train_examples = None
num_train_steps = None
if args.do_train:
train_examples = read_squad_examples(input_file=args.train_file,
is_training=True)
num_train_steps = int(
len(train_examples) / args.train_batch_size /
args.gradient_accumulation_steps * args.num_train_epochs)
# Prepare model
model = BertForQuestionAnswering.from_pretrained(args.bert_model)
if args.fp16:
model.half()
model.to(device)
if args.local_rank != -1:
model = torch.nn.parallel.DistributedDataParallel(
model, device_ids=[args.local_rank], output_device=args.local_rank)
elif n_gpu > 1:
model = torch.nn.DataParallel(model)
# Prepare optimizer
if args.fp16:
param_optimizer = [(n, param.clone().detach().to('cpu').float().requires_grad_()) \
for n, param in model.named_parameters()]
elif args.optimize_on_cpu:
param_optimizer = [(n, param.clone().detach().to('cpu').requires_grad_()) \
for n, param in model.named_parameters()]
else:
param_optimizer = list(model.named_parameters())
no_decay = ['bias', 'gamma', 'beta']
optimizer_grouped_parameters = [{
'params':
[p for n, p in param_optimizer if not any(nd in n for nd in no_decay)],
'weight_decay_rate':
0.01
}, {
'params':
[p for n, p in param_optimizer if any(nd in n for nd in no_decay)],
'weight_decay_rate':
0.0
}]
t_total = num_train_steps
if args.local_rank != -1:
t_total = t_total // torch.distributed.get_world_size()
optimizer = BertAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
warmup=args.warmup_proportion,
t_total=t_total)
global_step = 0
save_path_ls = []
if args.do_train:
cached_train_features_file = args.train_file + '_{0}_{1}_{2}_{3}'.format(
args.bert_model, str(args.max_seq_length), str(args.doc_stride),
str(args.max_query_length))
train_features = None
try:
with open(cached_train_features_file, "rb") as reader:
train_features = pickle.load(reader)
except:
train_features = convert_examples_to_features(
examples=train_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=True)
if args.local_rank == -1 or torch.distributed.get_rank() == 0:
logger.info(" Saving train features into cached file %s",
cached_train_features_file)
with open(cached_train_features_file, "wb") as writer:
pickle.dump(train_features, writer)
logger.info("***** Running training *****")
logger.info(" Num orig examples = %d", len(train_examples))
logger.info(" Num split examples = %d", len(train_features))
logger.info(" Batch size = %d", args.train_batch_size)
logger.info(" Num steps = %d", num_train_steps)
all_input_ids = torch.tensor([f.input_ids for f in train_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in train_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in train_features],
dtype=torch.long)
all_start_positions = torch.tensor(
[f.start_position for f in train_features], dtype=torch.long)
all_end_positions = torch.tensor(
[f.end_position for f in train_features], dtype=torch.long)
train_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_start_positions,
all_end_positions)
if args.local_rank == -1:
train_sampler = RandomSampler(train_data)
else:
train_sampler = DistributedSampler(train_data)
train_dataloader = DataLoader(train_data,
sampler=train_sampler,
batch_size=args.train_batch_size)
model.train()
for _ in trange(int(args.num_train_epochs), desc="Epoch"):
for step, batch in enumerate(
tqdm(train_dataloader, desc="Iteration")):
if n_gpu == 1:
batch = tuple(
t.to(device)
for t in batch) # multi-gpu does scattering it-self
input_ids, input_mask, segment_ids, start_positions, end_positions = batch
loss = model(input_ids, segment_ids, input_mask,
start_positions, end_positions)
if n_gpu > 1:
loss = loss.mean() # mean() to average on multi-gpu.
if args.fp16 and args.loss_scale != 1.0:
# rescale loss for fp16 training
# see https://docs.nvidia.com/deeplearning/sdk/mixed-precision-training/index.html
loss = loss * args.loss_scale
if args.gradient_accumulation_steps > 1:
loss = loss / args.gradient_accumulation_steps
loss.backward()
if (step + 1) % args.gradient_accumulation_steps == 0:
if args.fp16 or args.optimize_on_cpu:
if args.fp16 and args.loss_scale != 1.0:
# scale down gradients for fp16 training
for param in model.parameters():
if param.grad is not None:
param.grad.data = param.grad.data / args.loss_scale
is_nan = set_optimizer_params_grad(
param_optimizer,
model.named_parameters(),
test_nan=True)
if is_nan:
logger.info(
"FP16 TRAINING: Nan in gradients, reducing loss scaling"
)
args.loss_scale = args.loss_scale / 2
model.zero_grad()
continue
optimizer.step()
copy_optimizer_params_to_model(
model.named_parameters(), param_optimizer)
else:
optimizer.step()
model.zero_grad()
global_step += 1
print("Saving model...")
save_path = os.path.join(args.output_dir,
"step_" + str(global_step) + ".pth")
torch.save(model.state_dict(), save_path)
save_path_ls.append(save_path)
if args.do_predict and (args.local_rank == -1
or torch.distributed.get_rank() == 0):
eval_examples = read_squad_examples(input_file=args.predict_file,
is_training=False)
eval_features = convert_examples_to_features(
examples=eval_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=False)
logger.info("***** Running predictions *****")
logger.info(" Num orig examples = %d", len(eval_examples))
logger.info(" Num split examples = %d", len(eval_features))
logger.info(" Batch size = %d", args.predict_batch_size)
all_input_ids = torch.tensor([f.input_ids for f in eval_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in eval_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in eval_features],
dtype=torch.long)
all_example_index = torch.arange(all_input_ids.size(0),
dtype=torch.long)
eval_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_example_index)
# Run prediction for full data
eval_sampler = SequentialSampler(eval_data)
eval_dataloader = DataLoader(eval_data,
sampler=eval_sampler,
batch_size=args.predict_batch_size)
for save_dir in save_path_ls:
model.load_state_dict(torch.load(save_dir))
model_name = save_dir.split('/')[-1]
step_name = model_name[:-4]
output_path = os.path.join(args.output_dir, step_name)
if not os.path.exists(output_path):
os.makedirs(output_path)
model.eval()
all_results = []
logger.info("Start evaluating")
for input_ids, input_mask, segment_ids, example_indices in tqdm(
eval_dataloader, desc="Evaluating"):
if len(all_results) % 1000 == 0:
logger.info("Processing example: %d" % (len(all_results)))
input_ids = input_ids.to(device)
input_mask = input_mask.to(device)
segment_ids = segment_ids.to(device)
with torch.no_grad():
batch_start_logits, batch_end_logits = model(
input_ids, segment_ids, input_mask)
for i, example_index in enumerate(example_indices):
start_logits = batch_start_logits[i].detach().cpu().tolist(
)
end_logits = batch_end_logits[i].detach().cpu().tolist()
eval_feature = eval_features[example_index.item()]
unique_id = int(eval_feature.unique_id)
all_results.append(
RawResult(unique_id=unique_id,
start_logits=start_logits,
end_logits=end_logits))
output_prediction_file = os.path.join(output_path,
"predictions.json")
output_nbest_file = os.path.join(output_path,
"nbest_predictions.json")
write_predictions(eval_examples, eval_features, all_results,
args.n_best_size, args.max_answer_length,
args.do_lower_case, output_prediction_file,
output_nbest_file, args.verbose_logging)
def main():
parser = argparse.ArgumentParser()
## Required parameters
parser.add_argument(
"--bert_model",
default=None,
type=str,
required=True,
help="Bert pre-trained model selected in the list: bert-base-uncased, "
"bert-large-uncased, bert-base-cased, bert-large-cased, bert-base-multilingual-uncased, "
"bert-base-multilingual-cased, bert-base-chinese.")
parser.add_argument(
"--output_dir",
default=None,
type=str,
required=True,
help=
"The output directory where the model checkpoints and predictions will be written."
)
parser.add_argument(
"--preds_dir",
default=None,
type=str,
required=True,
help="The directory where the predictions will be written.")
## Other parameters
parser.add_argument("--train_file",
default=None,
type=str,
help="SQuAD json for training. E.g., train-v1.1.json")
parser.add_argument(
"--predict_file",
default=None,
type=str,
help="SQuAD json for predictions. E.g., dev-v1.1.json or test-v1.1.json"
)
parser.add_argument(
"--max_seq_length",
default=384,
type=int,
help=
"The maximum total input sequence length after WordPiece tokenization. Sequences "
"longer than this will be truncated, and sequences shorter than this will be padded."
)
parser.add_argument(
"--doc_stride",
default=128,
type=int,
help=
"When splitting up a long document into chunks, how much stride to take between chunks."
)
parser.add_argument(
"--max_query_length",
default=64,
type=int,
help=
"The maximum number of tokens for the question. Questions longer than this will "
"be truncated to this length.")
parser.add_argument("--do_train",
action='store_true',
help="Whether to run training.")
parser.add_argument("--do_predict",
action='store_true',
help="Whether to run eval on the dev set.")
parser.add_argument("--do_evaluate",
action='store_true',
help="Whether to eval after training.")
parser.add_argument("--train_batch_size",
default=32,
type=int,
help="Total batch size for training.")
parser.add_argument("--predict_batch_size",
default=8,
type=int,
help="Total batch size for predictions.")
parser.add_argument("--learning_rate",
default=5e-5,
type=float,
help="The initial learning rate for Adam.")
parser.add_argument("--num_train_epochs",
default=3.0,
type=float,
help="Total number of training epochs to perform.")
parser.add_argument(
"--warmup_proportion",
default=0.1,
type=float,
help=
"Proportion of training to perform linear learning rate warmup for. E.g., 0.1 = 10%% "
"of training.")
parser.add_argument(
"--n_best_size",
default=20,
type=int,
help=
"The total number of n-best predictions to generate in the nbest_predictions.json "
"output file.")
parser.add_argument(
"--max_answer_length",
default=30,
type=int,
help=
"The maximum length of an answer that can be generated. This is needed because the start "
"and end predictions are not conditioned on one another.")
parser.add_argument(
"--verbose_logging",
action='store_true',
help=
"If true, all of the warnings related to data processing will be printed. "
"A number of warnings are expected for a normal SQuAD evaluation.")
parser.add_argument("--no_cuda",
action='store_true',
help="Whether not to use CUDA when available")
parser.add_argument('--seed',
type=int,
default=42,
help="random seed for initialization")
parser.add_argument(
'--gradient_accumulation_steps',
type=int,
default=1,
help=
"Number of updates steps to accumulate before performing a backward/update pass."
)
parser.add_argument(
"--do_lower_case",
action='store_true',
help=
"Whether to lower case the input text. True for uncased models, False for cased models."
)
parser.add_argument("--local_rank",
type=int,
default=-1,
help="local_rank for distributed training on gpus")
parser.add_argument(
'--fp16',
action='store_true',
help="Whether to use 16-bit float precision instead of 32-bit")
parser.add_argument('--overwrite_output_dir',
action='store_true',
help="Overwrite the content of the output directory")
parser.add_argument(
'--loss_scale',
type=float,
default=0,
help=
"Loss scaling to improve fp16 numeric stability. Only used when fp16 set to True.\n"
"0 (default value): dynamic loss scaling.\n"
"Positive power of 2: static loss scaling value.\n")
parser.add_argument(
'--version_2_with_negative',
action='store_true',
help=
'If true, the SQuAD examples contain some that do not have an answer.')
parser.add_argument(
'--null_score_diff_threshold',
type=float,
default=0.0,
help=
"If null_score - best_non_null is greater than the threshold predict null."
)
parser.add_argument('--server_ip',
type=str,
default='',
help="Can be used for distant debugging.")
parser.add_argument('--server_port',
type=str,
default='',
help="Can be used for distant debugging.")
parser.add_argument('--eval_period', type=int, default=2000)
parser.add_argument('--wait_step', type=int, default=4)
parser.add_argument('--load_from_cache',
action='store_true',
help="Load train features from cache.")
args = parser.parse_args()
#print(args)
if args.server_ip and args.server_port:
# Distant debugging - see https://code.visualstudio.com/docs/python/debugging#_attach-to-a-local-script
import ptvsd
print("Waiting for debugger attach")
ptvsd.enable_attach(address=(args.server_ip, args.server_port),
redirect_output=True)
ptvsd.wait_for_attach()
if args.local_rank == -1 or args.no_cuda:
device = torch.device("cuda" if torch.cuda.is_available()
and not args.no_cuda else "cpu")
n_gpu = torch.cuda.device_count()
else:
torch.cuda.set_device(args.local_rank)
device = torch.device("cuda", args.local_rank)
n_gpu = 1
# Initializes the distributed backend which will take care of sychronizing nodes/GPUs
torch.distributed.init_process_group(backend='nccl')
logging.basicConfig(
format='%(asctime)s - %(levelname)s - %(name)s - %(message)s',
datefmt='%m/%d/%Y %H:%M:%S',
level=logging.INFO if args.local_rank in [-1, 0] else logging.WARN)
logger.info(
"device: {} n_gpu: {}, distributed training: {}, 16-bits training: {}".
format(device, n_gpu, bool(args.local_rank != -1), args.fp16))
if args.gradient_accumulation_steps < 1:
raise ValueError(
"Invalid gradient_accumulation_steps parameter: {}, should be >= 1"
.format(args.gradient_accumulation_steps))
args.train_batch_size = args.train_batch_size // args.gradient_accumulation_steps
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
if n_gpu > 0:
torch.cuda.manual_seed_all(args.seed)
if not args.do_train and not args.do_predict:
raise ValueError(
"At least one of `do_train` or `do_predict` must be True.")
if args.do_train:
if not args.train_file:
raise ValueError(
"If `do_train` is True, then `train_file` must be specified.")
if args.do_predict:
if not args.predict_file:
raise ValueError(
"If `do_predict` is True, then `predict_file` must be specified."
)
with open(args.predict_file, encoding='utf-8') as pf:
dev_data = json.load(pf)["data"]
if os.path.exists(args.output_dir) and os.listdir(
args.output_dir
) and args.do_train and not args.overwrite_output_dir:
raise ValueError(
"Output directory () already exists and is not empty.")
if not os.path.exists(args.output_dir):
os.makedirs(args.output_dir)
if args.local_rank not in [-1, 0]:
torch.distributed.barrier(
) # Make sure only the first process in distributed training will download model & vocab
if args.do_train:
tokenizer = BertTokenizer.from_pretrained(
args.bert_model, do_lower_case=args.do_lower_case)
model = BertForQuestionAnswering.from_pretrained(args.bert_model)
elif args.do_evaluate:
# Load a trained model and vocabulary that you have fine-tuned
model = BertForQuestionAnswering.from_pretrained(args.output_dir)
tokenizer = BertTokenizer.from_pretrained(
args.output_dir, do_lower_case=args.do_lower_case)
if args.local_rank == 0:
torch.distributed.barrier()
model.to(device)
if n_gpu > 1:
model = torch.nn.DataParallel(model)
if args.do_train:
if args.local_rank in [-1, 0]:
tb_writer = SummaryWriter()
# Prepare data loader
train_examples = read_squad_examples(
input_file=args.train_file,
is_training=True,
version_2_with_negative=args.version_2_with_negative)
cached_train_features_file = args.train_file + '_{0}_{1}_{2}_{3}'.format(
list(filter(None, args.bert_model.split('/'))).pop(),
str(args.max_seq_length), str(args.doc_stride),
str(args.max_query_length))
try:
if args.load_from_cache:
with open(cached_train_features_file, "rb") as reader:
train_features = pickle.load(reader)
else:
raise Exception
except:
train_features = convert_examples_to_features(
examples=train_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=True)
if args.local_rank == -1:
logger.info(" Saving train features into cached file %s",
cached_train_features_file)
with open(cached_train_features_file, "wb") as writer:
pickle.dump(train_features, writer)
all_input_ids = torch.tensor([f.input_ids for f in train_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in train_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in train_features],
dtype=torch.long)
all_start_positions = torch.tensor(
[f.start_position for f in train_features], dtype=torch.long)
all_end_positions = torch.tensor(
[f.end_position for f in train_features], dtype=torch.long)
train_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_start_positions,
all_end_positions)
train_sampler = RandomSampler(train_data)
train_dataloader = DataLoader(train_data,
sampler=train_sampler,
batch_size=args.train_batch_size)
num_train_optimization_steps = len(
train_dataloader
) // args.gradient_accumulation_steps * args.num_train_epochs
# Prepare optimizer
param_optimizer = list(model.named_parameters())
# hack to remove pooler, which is not used
# thus it produce None grad that break apex
param_optimizer = [n for n in param_optimizer if 'pooler' not in n[0]]
no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [{
'params': [
p for n, p in param_optimizer
if not any(nd in n for nd in no_decay)
],
'weight_decay':
0.01
}, {
'params':
[p for n, p in param_optimizer if any(nd in n for nd in no_decay)],
'weight_decay':
0.0
}]
optimizer = BertAdam(optimizer_grouped_parameters,
lr=args.learning_rate,
warmup=args.warmup_proportion,
t_total=num_train_optimization_steps)
logger.info("***** Running training *****")
logger.info(" Num orig examples = %d", len(train_examples))
logger.info(" Num split examples = %d", len(train_features))
logger.info(" Batch size = %d", args.train_batch_size)
logger.info(" Num steps = %d", num_train_optimization_steps)
if args.do_predict and args.local_rank == -1:
eval_examples = read_squad_examples(
input_file=args.predict_file,
is_training=False,
version_2_with_negative=args.version_2_with_negative)
eval_features = convert_examples_to_features(
examples=eval_examples,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
doc_stride=args.doc_stride,
max_query_length=args.max_query_length,
is_training=False)
logger.info("***** Running predictions *****")
logger.info(" Num orig examples = %d", len(eval_examples))
logger.info(" Num split examples = %d", len(eval_features))
logger.info(" Batch size = %d", args.predict_batch_size)
all_input_ids = torch.tensor([f.input_ids for f in eval_features],
dtype=torch.long)
all_input_mask = torch.tensor([f.input_mask for f in eval_features],
dtype=torch.long)
all_segment_ids = torch.tensor([f.segment_ids for f in eval_features],
dtype=torch.long)
all_example_index = torch.arange(all_input_ids.size(0),
dtype=torch.long)
eval_data = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_example_index)
# Run prediction for full data
eval_sampler = SequentialSampler(eval_data)
eval_dataloader = DataLoader(eval_data,
sampler=eval_sampler,
batch_size=args.predict_batch_size)
if args.do_train:
best_f1 = 0
wait_step = 0
global_step = 0
stop_training = False
model.train()
for epoch in trange(int(args.num_train_epochs), desc="Epoch"):
for step, batch in enumerate(
tqdm(train_dataloader,
desc="Iteration",
disable=args.local_rank not in [-1, 0])):
if n_gpu == 1:
batch = tuple(
t.to(device)
for t in batch) # multi-gpu does scattering it-self
input_ids, input_mask, segment_ids, start_positions, end_positions = batch
loss = model(input_ids, segment_ids, input_mask,
start_positions, end_positions)
if n_gpu > 1:
loss = loss.mean() # mean() to average on multi-gpu.
if args.gradient_accumulation_steps > 1:
loss = loss / args.gradient_accumulation_steps
loss.backward()
if (step + 1) % args.gradient_accumulation_steps == 0:
optimizer.step()
optimizer.zero_grad()
global_step += 1
if global_step % args.eval_period == 0:
model.eval()
scores = predict(args, model, eval_examples, eval_features,
eval_dataloader, dev_data, device)
em, f1 = scores['exact_match'], scores['f1']
logger.info("f1: %.3f, em: %.3f on epoch=%d" %
(f1, em, epoch))
print("f1: %.3f, em: %.3f on epoch=%d" % (f1, em, epoch))
if best_f1 < f1:
logger.info("Saving model with best f1: %.3f -> %.3f on epoch=%d" % \
(best_f1, f1, epoch))
save_model(args, model, device, tokenizer)
best_f1 = f1
wait_step = 0
stop_training = False
else:
wait_step += 1
if best_f1 > 10 and wait_step == args.wait_step:
stop_training = True
model.train()
if stop_training:
break
elif args.do_evaluate:
model.eval()
scores = predict(args, model, eval_examples, eval_features,
eval_dataloader, dev_data, device, True)
em, f1 = scores['exact_match'], scores['f1']
print("f1: %.3f, em: %.3f" % (f1, em))
