Merge pull request !3726 from zhaozhenlong/lite/op/prior_box

Merge pull request !3688 from ghzl/add-deconv-parser

Merge pull request !3844 from lixian/master

Merge pull request !3713 from ghzl/deconv-adapter

Merge pull request !3835 from sunsuodong/int8_relu_hswish

Merge pull request !3843 from fuzhiye/mindspore

Merge pull request !3810 from pengyongrong/master

modify CMakeLists.txt

remove old MINDSPORE_HCCL_CONFIG_PATH in model zoo

decoupling core and debug

add lite/test

add gpu python

pad fp32 support less than 4d input

repair distribute training for vgg in D

Merge pull request !3837 from liucunwei/add_test_for_mindrecord3

Merge pull request !3833 from fuzhiye/mindspore

Merge pull request !3814 from liucunwei/enhance_map_file_for_imagenet

Merge pull request !3808 from 张学同/to_merge

Merge pull request !3836 from zhanyuan/master

Merge pull request !3776 from lixian/master

Merge pull request !3815 from xutianchun/quant_0731

2. fix anf_exporter bug: pool, concat op may not set into metagraph
3. fix weight trans pass will return error when post quantization
4. make anf_exporter reentrant: do not set PrimitiveT * to nullptr

Merge pull request !3827 from hangq/master

Merge pull request !3610 from TinaMengtingZhang/cpp-api-concat

Merge pull request !3829 from cathwong/ckw_c_api_transform_hdr_alpha

Merge pull request !3622 from EricZ/distributed_sampler_fix

Merge pull request !3760 from 34bunny/GPU-ResizeNearestNeighbor-fix

Merge pull request !3619 from danishnxt/GPU_One

Added Distributed sampler option

Fix style

Merge pull request !3693 from dessyang/master

Merge pull request !3549 from anthonyaje/device_augment_mapop_refactor1

Merge pull request !3553 from tony_liu2/staging

Merge pull request !3801 from liubuyu/master

Merge pull request !3780 from baihuawei/losscuda