distributed_strategy.proto

// Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

syntax = "proto2";
package paddle.fleet;

enum Mode {
  COLLECTIVE = 1;
  PS = 2;
  PIPELINE = 3;
  HETER = 4; // support XPU and GPU computing server
}

message DistributedStrategy {
  optional Mode mode = 1 [ default = COLLECTIVE ]; // just for serialization
  // collective training strategy
  optional bool amp = 2 [ default = false ];
  optional int32 amp_loss_scaling = 3 [ default = 32768 ];
  optional bool recompute = 4 [ default = false ];
  repeated string recompute_checkpoints = 5;
  optional bool localsgd = 6 [ default = false ];
  optional int32 localsgd_k_step = 7 [ default = 4 ];
  optional bool dgc = 8 [ default = false ];
  optional bool hierachical_allreduce = 9 [ default = false ];
  optional int32 nccl_comm_num = 10 [ default = 1 ];
  optional bool gradient_merge = 11 [ default = false ];
  optional int32 gradient_merge_k_step = 12 [ default = 1 ];
  optional bool sequential_execution = 13 [ default = false ];
  optional bool enable_backward_optimizer_op_deps = 14 [ default = true ];
  optional bool lars = 15 [ default = false ];
  optional bool lamb = 16 [ default = false ];
  optional bool fuse_elewise_add_act_ops = 17 [ default = false ];
  optional bool fuse_bn_act_ops = 18 [ default = false ];
  optional bool enable_auto_fusion = 19 [ default = false ];
  optional bool fuse_relu_depthwise_conv = 20 [ default = false ];
  optional bool enable_inplace = 21 [ default = false ];
  optional bool fuse_all_reduce_ops = 22 [ default = false ];
  optional int32 num_iteration_per_drop_scope = 23 [ default = 1 ];
  optional bool sync_batch_norm = 24 [ default = false ];
  optional bool fuse_all_optimizer_ops = 25 [ default = false ];

  // pipeline training
  optional bool pipeline = 101 [ default = false ];
  optional int32 pipeline_micro_batch = 102;

  // parameter server training
  optional bool sync = 201 [ default = false ];
  optional bool async = 202 [ default = true ];
  optional int32 async_k_step = 203 [ default = -1 ];
  optional int32 max_merge_var_num = 204 [ default = 1 ];
  optional int32 send_queue_size = 205 [ default = 16 ];
  optional bool independent_recv_thread = 206 [ default = false ];
  optional int32 min_send_grad_num_before_recv = 207 [ default = 1 ];
  optional int32 thread_pool_size = 208 [ default = 1 ];
  optional int32 send_wait_times = 209 [ default = 1 ];
  optional bool runtime_split_send_recv = 210 [ default = false ];
  optional bool use_thread_barrier = 211 [ default = false ];

  // elastic deep learning strategies
  optional bool elastic = 301 [ default = false ];

  // auto parallel
  optional bool auto = 401 [ default = false ];
}

message DistributedJobInfo {
  optional int32 worker_num = 1;
  optional int32 server_num = 2;
  repeated string worker_ips = 3;
  repeated string server_endpoints = 4;
  optional string origin_startup = 5;
  optional string origin_main = 6; // without backpropagation and optimization
  optional string distributed_main = 7; // with backpropagation and optimization
  optional string optimizer_name = 8;   // optimizer name
  optional DistributedStrategy strategy = 101;
}