* fix selected_gpus bug for local_cluster

* fix bugs about single machine multicards training for local_cluster

* fix python -m config.yaml -b backend.yaml

* re add selected_gpus_num config for single train mode

* fix bug for submit for paddlecloud k8s
Co-authored-by: Nwuzhihua <35824027+fuyinno4@users.noreply.github.com>

* add cluster support

* fix split files at PY3

* fix linux at PY3

* fix desc error

* fix collective cards and worknum
Co-authored-by: Ntangwei <tangwei12@baidu.com>

* fix setup

* fix bug for dssm reader

* fix net bug at PY3 for afm


* fix multi cards with files

* fix ctr

* add validation

* add validation

* add validation

* fix compile

* fix ci

* fix user define runner

* fix gnn reader at PY3

* fix fast yaml config at PY3
Co-authored-by: Ntangwei <tangwei12@baidu.com>

* add cluster running

* add k8s/paddle cloud

* fix local cluster without gpu

* define engine to train/infer/local_cluster_train/cluster_train

* fix gpu cluster wait

* fix bug for 2ps 2tr
Co-authored-by: Ntangwei <tangwei12@baidu.com>

* fix gpu default card & doc

* fix yaml import

* add value checker

* add development.md & update yaml.md

* fix tdm

* update general trainer

* update design

* rename model & reader

* update startup

* add gpu support

* after self-test

* code clean

* fix

* change dnn yaml

* add mpi support

* fix pslib

* fix pslib

* fix pslib network

* fix pslib program

* fix pslib runner

* fix fleet mode

* fix is_distribute

* fix deepfm

* change pslib code

* fix din

* update & fix

* fix reader base

* add ci & fix

* fix fnn

* fix ci sudo
Co-authored-by: Ntangwei <tangwei12@baidu.com>