Huggingface错误：AttributeError:“ByTelevelbeTokenizer”对象没有属性“pad\u token\u id”

Question

我正在尝试使用WordLevel/BPE标记器对一些数字字符串进行标记，创建一个数据压缩程序，并最终在PyTorch数据加载器中使用它来从头训练一个新模型

然而，我得到了一个错误

AttributeError: 'ByteLevelBPETokenizer' object has no attribute 'pad_token_id'

当运行以下代码时

from transformers import DataCollatorForLanguageModeling
from tokenizers import ByteLevelBPETokenizer
from tokenizers.pre_tokenizers import Whitespace
from torch.utils.data import DataLoader, TensorDataset

data = ['4814 4832 4761 4523 4999 4860 4699 5024 4788 <unk>']

# Tokenizer
tokenizer = ByteLevelBPETokenizer()
tokenizer.pre_tokenizer = Whitespace()
tokenizer.train_from_iterator(data, vocab_size=1000, min_frequency=1, 
    special_tokens=[
        "<s>",
        "</s>",
        "<unk>",
        "<mask>",
    ])

# Data Collator
data_collator = DataCollatorForLanguageModeling(
    tokenizer=tokenizer, mlm=False
)

train_dataset = TensorDataset(torch.tensor(tokenizer(data, ......)))

# DataLoader
train_dataloader = DataLoader(
    train_dataset, 
    collate_fn=data_collator
)

这个错误是因为没有为标记器配置pad_token_id吗？如果是的话，我们怎么做

谢谢

错误跟踪：

AttributeError: Caught AttributeError in DataLoader worker process 0.
Original Traceback (most recent call last):
  File "/opt/anaconda3/envs/x/lib/python3.8/site-packages/torch/utils/data/_utils/worker.py", line 198, in _worker_loop
    data = fetcher.fetch(index)
  File "/opt/anaconda3/envs/x/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 47, in fetch
    return self.collate_fn(data)
  File "/opt/anaconda3/envs/x/lib/python3.8/site-packages/transformers/data/data_collator.py", line 351, in __call__
    if self.tokenizer.pad_token_id is not None:
AttributeError: 'ByteLevelBPETokenizer' object has no attribute 'pad_token_id'

康达套餐

pytorch                   1.7.0           py3.8_cuda10.2.89_cudnn7.6.5_0    pytorch
pytorch-lightning         1.2.5              pyhd8ed1ab_0    conda-forge
tokenizers                0.10.1                   pypi_0    pypi
transformers              4.4.2                    pypi_0    pypi

Answer 1

该错误告诉您标记器需要一个名为pad_token_id的属性。您可以将ByteLevelBPETokenizer包装到具有这样一个属性的类中（…并且遇到了其他丢失的属性），或者使用transformers库中的包装类：

from transformers import PreTrainedTokenizerFast

#your code
tokenizer.save(SOMEWHERE)
tokenizer = PreTrainedTokenizerFast(tokenizer_file=tokenizer_path)