wehub-resource-sync
74 lines
3.1 KiB
Python
74 lines
3.1 KiB
Python
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
|
|
# Copyright 2021 The HuggingFace Inc. team. All rights reserved.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
import paddle
|
|
|
|
from paddlenlp.data import Dict, Pad
|
|
|
|
|
|
class DataCollatorMLM:
|
|
def __init__(self, tokenizer, batch_pad=None):
|
|
self.batch_pad = batch_pad
|
|
self.mask_token_id = tokenizer.mask_token_id
|
|
self.pad_token_id = tokenizer.pad_token_id
|
|
self.token_len = tokenizer.vocab_size
|
|
if batch_pad is None:
|
|
self.batch_pad = lambda samples, fn=Dict(
|
|
{
|
|
"input_ids": Pad(axis=0, pad_val=self.pad_token_id, dtype="int64"), # input
|
|
# 'token_type_ids': Pad(axis=0, pad_val=0, dtype='int64'), # segment
|
|
"special_tokens_mask": Pad(axis=0, pad_val=True, dtype="int64"), # segment
|
|
}
|
|
): fn(samples)
|
|
else:
|
|
self.batch_pad = batch_pad
|
|
|
|
def __call__(self, examples):
|
|
examples = self.batch_pad(examples)
|
|
examples = [paddle.to_tensor(e) for e in examples]
|
|
examples[0], labels = self._mask_tokens(
|
|
examples[0], paddle.cast(examples[1], dtype=bool), self.mask_token_id, self.token_len
|
|
)
|
|
examples.append(labels)
|
|
return examples
|
|
|
|
def _mask_tokens(self, inputs, special_tokens_mask, mask_token_id, token_len, mlm_prob=0.15, ignore_label=-100):
|
|
"""
|
|
Prepare masked tokens inputs/labels for masked language modeling: 80% MASK, 10% random, 10% original.
|
|
"""
|
|
labels = inputs.clone()
|
|
probability_matrix = paddle.full(labels.shape, mlm_prob)
|
|
probability_matrix[special_tokens_mask] = 0
|
|
|
|
masked_indices = paddle.cast(paddle.bernoulli(probability_matrix), dtype=bool)
|
|
labels[~masked_indices] = ignore_label # We only compute loss on masked tokens
|
|
|
|
# 80% of the time, we replace masked input tokens with tokenizer.mask_token ([MASK])
|
|
indices_replaced = paddle.cast(paddle.bernoulli(paddle.full(labels.shape, 0.8)), dtype=bool) & masked_indices
|
|
inputs[indices_replaced] = mask_token_id
|
|
|
|
# 10% of the time, we replace masked input tokens with random word
|
|
|
|
indices_random = (
|
|
paddle.cast(paddle.bernoulli(paddle.full(labels.shape, 0.5)), dtype=bool)
|
|
& masked_indices
|
|
& ~indices_replaced
|
|
)
|
|
random_words = paddle.randint(low=0, high=token_len, shape=labels.shape)
|
|
inputs[indices_random] = random_words[indices_random]
|
|
|
|
# The rest of the time (10% of the time) we keep the masked input tokens unchanged
|
|
return inputs, labels
|