* refine transformer cpu

* delete run_pretrain.sh

* add transformer for rc0

* predict no_grad

* support pad seq and bsz multi

* alter vocab size

* alter md5sum

* root for reader

* add static predict

* rm position_encoding_init

* fix static predict

* add pad factor to pad vocab

* shuffle batch

* update reader

* update benchmark

* fix hung

* add shuffle batch

* update

* reader_seed to shuffle_seed

* seed for shuffle batch

* update __init__

* fix readme version info

* update __init__

* update seq2seq example, using bleu

* remove target file arg

* update readme

* update __init__

* update couplet api usage, delete useless import

* update seq2seq readme

* update seq2seq readme path

* update readme

* update test=develop

* fix readme

* fix translation dataset and simplify transformer reader

* update seq2seq, using paddlenlp

* Using new paddlenlp API

* update seq2seqREADME

* wrap dev ds

* delete useless comments

* update predict.py

* using paddlenlp.bleu

* remove shard

* update README, using bleu perl

* delete cand

* Remove tokens that make sentences longer than max_len

* remove pdb

* remove useless code.

* update url and dataset name of vae dataset(ptb and yahoo)

* update seq2seq and vae, data and README

* Add TokenEmbedding

* download corpus embedding data
* load embedding data by specifying corpus name
* extend the vocab of tokenizer from corpus embedding data

* add unk token setting

* modify tokenizer

* add extend voacb

* move jieba tokenizer and rename corpus_name->embedding_name

* use bos url instead of localhost

* add log when loading data

* add token dot computation; add __repr__ of TokenEmbedding

* add color logging

* use paddlenlp.utils.log

* adjust repr

* update pretrained embedding table

* fix padding idx