GloVe = Word2vec + 共现矩阵
ELMo = GloVe +LSTM + pre-train+fine tuning
Transformer = (Seq2Seq) + Attention
GPT = Transformer(self-attention) + ELMo
BERT = GPT + 双向的Transformer + [MASK]token
GloVe = Word2vec + 共现矩阵
ELMo = GloVe +LSTM + pre-train+fine tuning
Transformer = (Seq2Seq) + Attention
GPT = Transformer(self-attention) + ELMo
BERT = GPT + 双向的Transformer + [MASK]token