Tensorflow bert标记化未知词

from tokenizers import BertWordPieceTokenizer

tokenizer = BertWordPieceTokenizer('bert-base-uncased-vocab.txt')
tokens = tokenizer.encode("Hello, world!")

import tensorflow as tf
from transformers import BertTokenizer, TFBertForMaskedLM

tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
model = TFBertForMaskedLM.from_pretrained('bert-base-uncased')

text = "Hello, [MASK]!"
inputs = tokenizer(text, return_tensors="tf")
outputs = model(inputs)
predictions = tf.argmax(outputs.logits, axis=-1)
predicted_token = tokenizer.decode(predictions[0][0])

import tensorflow as tf
from tensorflow.keras.layers import Embedding, LSTM, Dense

vocab_size = 1000
embedding_dim = 16
max_length = 100

inputs = tf.keras.Input(shape=(max_length,))
embedding_layer = Embedding(vocab_size, embedding_dim, input_length=max_length)(inputs)
lstm_layer = LSTM(64)(embedding_layer)
outputs = Dense(vocab_size, activation='softmax')(lstm_layer)

model = tf.keras.Model(inputs=inputs, outputs=outputs)