* 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 save model for step

* add PS/Collective support , add definition in yaml.md

* add check for QueueDataset while saving by step, modify dnn config yaml as an example

* add logging about model save operation

* fix bug for adding check for QueueDataset, modify dnn config yaml

* fix bug for dnn model config yaml

* fix bug for checking QueueDataset

* fix UT fail

* fix UT fail
Co-authored-by: Ntangwei12 <tangwei12@baidu.com>

* fix demo

* fix

* fix

* fix code style

* add support for file_list shuffle each epoch, test=develop

* fix float learning rate bug

* optimized code for shuffle files
Co-authored-by: Ntangwei12 <tangwei12@baidu.com>

* fix save inference

* add fine tuning trainer

* add doc

* fix
Co-authored-by: NseiriosPlus <tangwei12@baidu.com>
Co-authored-by: Nwuzhihua <35824027+fuyinno4@users.noreply.github.com>

Revert "fix read hide files bug for core/trainers/framework/dataset.py core/utils/dataloader_instance.py"

This reverts commit 348042af.

* fix

* test

* remove setup change

* fix

* fix
Co-authored-by: Ntangwei12 <tangwei12@baidu.com>