Paddle

Author	SHA1	Message	Date
Dong Daxiang	8d2896f1fe	【paddle.fleet】Fleet run graph in Executor and add two more strategies (#25844 ) * split meta optimizer files * add graph execution in execution, update two properties in DistributedStrategy, unit tests for these features	6 years ago
tangwei12	caa90a6510	Integrated Trainer of Parameter Server (API add `fluid.contrib.layers.sparse_embedding` only) (#22957 ) * Integrated Trainer of Parameter Server	6 years ago
Dong Daxiang	a96d54ac19	Generate final strategy (#25782 ) * refine strategy compiler and meta optimizers make async as a_sync	6 years ago
Dong Daxiang	920d998f1e	add more settings for distributed strategy (#25685 ) * add more settings for distributed strategy Basically, DistributedStrategy has several parts of configurations: - BuildStrategy: the same as paddle.fluid.BuildStrategy, but the distributed arguments are moved out of BuildStrategy - ExecutionStrategy: the same as paddle.fluid.ExecutionStrategy - collective communication configs: nccl_comm_num, hierarchical allreduce and so on - distributed algorithms: async_update(mainly used in PS), lars, lamb and so on	6 years ago
Dong Daxiang	28064c2dc9	fix gen nccl id bug (#25669 ) * fix gen nccl id bug	6 years ago
Dong Daxiang	e657d7062d	fleet base initial implementation and the API (#25442 ) refactor fleet api under paddle.fleet update DistributedStrategy	6 years ago