* 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

* Update paddlenlp.metrics README

* Add info for Mcc

* delete DuReader metric

* update task name, add API col

* add custom op gelu

* refine run_pretrain

* add electra pretrain and modify style of electra modeling

* add electra pretrain, modify style of electra modeling and fix problems of review

* delete predict_classifer

* modify accu to acc

* add paddlenlp.metrics.glue

use paddlenlp downloader for the tokenizer

* add model api

* update codes

* update codes

* update codes

* add amp support for model SE_ResNet50_vd

* optimize code, add model ResNet50 and SE_ResNet50_vd to list

* merge develop

* add the support of amp for the fleet train
* remove the attribute for the amp

* update lrscheduler

* minor fix

* add pre-commit

* minor fix

* Add __len__ to squad dataset

* minor fix

* update lrscheduler

* minor fix

* add pre-commit

Co-authored-by: Nfangzeyang <fangzeyang@baidu.com>

* update api 1.8

* fix paddlerec readme

* fix

* fix

* fix 2.0

* fix 2.0

* fix 2.0

* fix 2.0

* fix 2.0

* fix

* add pure fp16 training.

* update the usage of cast_parameters_to_fp16.

* use fp32 to initilize parameter values and then cast to fp16.

* enable multi_precision when training with fp16.

* add static loss scaling.

* add the rescale_grad function in the pure fp16 training.

* use cast to complete the pure fp16 training.

* use sum loss for fp16 trainging and mean loss for fp32 training.

* refine codes according reviewers' suggestions

* Calculate and print the average time for video models.

* Calculate and print the average time for dygraph of slowfast.

* Print ips for nextvlad, slowfast, tsm and tsn.

add mso significant commit, because of creep API change, no more steps toward to be a time killer and all of it is crap (#4953)

average reader cost; test=develop

* add reader cost, and change speed to ips; test=develop

* fix inf bug; test=develop

* Create inception_v3.py

* Update __init__.py

* Create InceptionV3.sh

* fix load in metric learning

* fix load in metric learning

* fix typo

* fix typo

* fix video bug and slowfast benchmark bug

* refine bmn doc

* update api 1.8

* fix paddlerec readme

* fix readme

* update 2.0 api

* fix, test=develop