From: @ling_qiao_min
Reviewed-by: 
Signed-off-by:

1. description of num_shard in datasets.py
2. add param validation for sampler
3. add description for add_child in sampler.py

fix c++ ut compilation failed of Asan bug

From: @lixiachen
Reviewed-by: 
Signed-off-by:

From: @mengyuanli
Reviewed-by: @hangangqiang,@zhanghaibo5
Signed-off-by: @hangangqiang

From: @yuchaojie
Reviewed-by: @c_34,@liangchenghui
Signed-off-by: @c_34

From: @mengyuanli
Reviewed-by: @zhanghaibo5,@hangangqiang
Signed-off-by: @hangangqiang

From: @jonwe
Reviewed-by: @robingrosman,@tom__chen
Signed-off-by: @tom__chen

From: @wangzhe128
Reviewed-by: 
Signed-off-by:

From: @wuxuejian
Reviewed-by: @kisnwang,@guoqi1024,@c_34
Signed-off-by: @c_34

From: @YeFeng_24
Reviewed-by: 
Signed-off-by:

From: @ch-l
Reviewed-by: @zhoufeng54,@kisnwang
Signed-off-by: @kisnwang

From: @zhangyunshu
Reviewed-by: 
Signed-off-by:

From: @ouwenchang
Reviewed-by: 
Signed-off-by:

From: @qujianwei
Reviewed-by: @linqingke,@liangchenghui
Signed-off-by: @liangchenghui

From: @fuzhiye
Reviewed-by: @zhanghaibo5,@zhang_xue_tong
Signed-off-by: @zhang_xue_tong

From: @sunsuodong
Reviewed-by: @zhang_xue_tong
Signed-off-by:

From: @zhouyaqiang0
Reviewed-by: @c_34,@linqingke
Signed-off-by: @linqingke

From: @pandoublefeng
Reviewed-by: @c_34,@linqingke
Signed-off-by: @c_34

From: @zhengjun10
Reviewed-by: 
Signed-off-by:

From: @cjh9368
Reviewed-by: @zhanghaibo5
Signed-off-by:

增加测试用例

修改测试用例

修改测试用例

修改算子实现，alpha 写死为 1.0

修改算子实现，alpha 写死为 1
修改算子注册

修改算子实现，dy 和 y 搞反了

修改测试代码

修改算子实现

修改算子实现，float16 单独走一个分支

修改算子实现，前端的 float 转换为double

修改算子实现，改回来了

修改算子实现，float16 转换为 Eigen 库中的 half

修改算子实现，float16 转换为 Eigen 库中的 half

修改算子实现，float16 转换为 Eigen 库中的 half

修改算子实现 在单独的文件中实现 EluGrad

修改算子实现 在单独的文件中实现 EluGrad

修改算子实现 在单独的文件中实现 EluGrad

修改算子实现 在单独的文件中实现 EluGrad

格式化代码

修改 pylint 错误

更新注释信息

删除未使用的私有变量

删除未使用的私有变量

From: @changzherui
Reviewed-by: @zh_qh,@zhoufeng54
Signed-off-by: @zh_qh

Renamed illegal variables and functions.

reconverted the nae of function in Lite java files.

modified

From: @yeyunpeng2020
Reviewed-by: @zhanghaibo5,@zhang_xue_tong
Signed-off-by:

From: @ling_qiao_min
Reviewed-by: @zhang_xue_tong,@hangangqiang
Signed-off-by: @zhang_xue_tong

From: @pengyongrong
Reviewed-by: 
Signed-off-by:

Cast the float16-input to float32 before ReduceSum, and cast back to float16 after ReduceSum.
If the op after this ReduceSum is a casting from float16 to float32, then it can be eliminated.