m53297601
/
mindspore
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

!1511 [AutoParallel] handle PRelu as an element-wise activation op

Browse Source 
Merge pull request !1511 from Chong/ReID
pull/1511/MERGE
mindspore-ci-bot 5 years ago committed by Gitee
parent ad279e90fd 26d05be808
commit 2060cbe66a
3 changed files with 33 additions and 18 deletions
Show Stats Download Patch File Download Diff File

46
mindspore/ccsrc/parallel/auto_parallel/rec_core/rec_generate_strategy.cc
Unescape View File

@ -78,8 +78,8 @@ std::vector<std::vector<int32_t>> PrepareVirtualDataset(const std::vector<std::s
return strategies;
}
std::vector<std::vector<int32_t>> PrepareBiasAdd(const std::vector<std::shared_ptr<OperatorInfo>> &ops,
const size_t iter_ops, std::vector<int32_t> s) {
std::vector<std::vector<int32_t>> PrepareScalarInputOperator(const std::vector<std::shared_ptr<OperatorInfo>> &ops,
const size_t iter_ops, std::vector<int32_t> s) {
std::vector<std::vector<int32_t>> strategies;
auto dev_num = g_device_manager->DeviceNum();
@ -190,12 +190,16 @@ std::vector<std::vector<int32_t>> MakeDataParallelStrategy(const std::vector<std
std::vector<int32_t> s;
size_t input_size = origin_strategy->GetInputDim()[iter_op_inputs].size();
for (size_t dim = 0; dim < input_size; dim++) {
if (dim == 0 && input_size == 4) {
size_t max_device_num = g_device_manager->DeviceNum();
size_t target_tensor_batch = ops[iter_ops]->outputs_tensor_info()[0].shape()[0];
s.push_back(std::min(max_device_num, target_tensor_batch));
if (input_size == 1 || input_size == 2 || input_size == 4) {
if (dim == 0) {
size_t max_device_num = g_device_manager->DeviceNum();
size_t target_tensor_batch = ops[iter_ops]->outputs_tensor_info()[0].shape()[0];
s.push_back(std::min(max_device_num, target_tensor_batch));
} else {
s.push_back(1);
}
} else {
s.push_back(1);
MS_LOG(ERROR) << "Tensor's shape is unknown.";
}
}
@ -239,6 +243,8 @@ void GeneratePartitionedOperatorStrategy(const std::shared_ptr<Graph> graph,
std::vector<std::vector<int32_t>> strategies;
size_t iter_graph = index_list->at(iter_ops);
if (iter_graph == SIZE_MAX) {
StrategyPtr sp = std::make_shared<Strategy>(0, strategies);
ops[iter_ops]->SetSelectedStrategyAndCost(sp, ops[iter_ops]->selected_cost());
continue;
}
strategies = PrepareStrategy(graph, ops, iter_graph, iter_ops);
@ -389,7 +395,7 @@ std::vector<int32_t> ModifyStrategyIfReduceIncoming(const std::vector<std::share
std::vector<int32_t> s_Reduce;
std::vector<int32_t> axis_list;
for (size_t i = 0; i < s.size(); i++) {
axis_list.push_back(i + 1);
axis_list.push_back(i);
}
auto dim_list = GetDimList(ops, incoming_op_index);
for (auto axis : dim_list) {
@ -400,7 +406,7 @@ std::vector<int32_t> ModifyStrategyIfReduceIncoming(const std::vector<std::share
axis_list.erase(it);
}
for (size_t i = 0; i < (size_t)axis_list.size(); i++) {
s_Reduce.push_back(s[axis_list[i] - 1]);
s_Reduce.push_back(s[axis_list[i]]);
}
return s_Reduce;
}
@ -418,8 +424,6 @@ std::vector<int32_t> CopyIncomingOperatorInputStrategy(const std::vector<std::sh
ops[incoming_op_index]->type() == REDUCE_MIN || ops[incoming_op_index]->type() == REDUCE_MEAN) {
s = ModifyStrategyIfReduceIncoming(ops, incoming_op_index, s);
}
} else {
no_stra_op_list->push_back(iter_ops);
}
return s;
}
@ -428,12 +432,18 @@ std::vector<std::vector<int32_t>> GenerateStrategiesFromStrategy(const std::vect
const size_t iter_ops, std::vector<int32_t> s) {
std::vector<int32_t> s_empty = {};
std::vector<std::vector<int32_t>> stra;
if (s.size() == 0) {
for (size_t iter_op_inputs = 0; iter_op_inputs < (size_t)ops[iter_ops]->inputs_tensor_info().size();
iter_op_inputs++) {
stra.push_back(s);
}
return stra;
}
MS_EXCEPTION_IF_NULL(ops[iter_ops]);
if (ops[iter_ops]->type() == BIAS_ADD) {
return PrepareBiasAdd(ops, iter_ops, s);
if (ops[iter_ops]->type() == BIAS_ADD || ops[iter_ops]->type() == PRELU) {
return PrepareScalarInputOperator(ops, iter_ops, s);
}
if (ops[iter_ops]->type() == ONEHOT) {
return PrepareOneHot(s);
@ -504,10 +514,14 @@ void GenerateEliminatedOperatorStrategyForward(const std::shared_ptr<Graph> grap
} else {
s = CopyIncomingOperatorInputStrategy(ops, incoming_op_index, iter_ops, no_stra_op_list);
}
} else {
}
if (s.size() == 0) {
no_stra_op_list->push_back(iter_ops);
} else {
stra = GenerateStrategiesFromStrategy(ops, iter_ops, s);
}
stra = GenerateStrategiesFromStrategy(ops, iter_ops, s);
StrategyPtr sp = std::make_shared<Strategy>(0, stra);
ops[iter_ops]->SetSelectedStrategyAndCost(sp, ops[iter_ops]->selected_cost());
}
@ -541,7 +555,7 @@ std::vector<int32_t> ModifyStrategyIfReduceOutgoing(const std::vector<std::share
size_t s_index = 0;
size_t dim_list_index = 0;
for (size_t i = 0; i < (size_t)(s.size() + dim_list.size()); i++) {
if ((i + 1) == (size_t)dim_list[dim_list_index]) {
if (i == (size_t)dim_list[dim_list_index]) {
s_Reduce.push_back(1);
dim_list_index++;
} else {

4
mindspore/ccsrc/parallel/auto_parallel/rec_core/rec_generate_strategy.h
Unescape View File

@ -36,8 +36,8 @@ std::vector<std::vector<int32_t>> PrepareMatMul(const std::shared_ptr<Graph> &gr
const size_t iter_graph, const size_t iter_ops);
std::vector<std::vector<int32_t>> PrepareVirtualDataset(const std::vector<std::shared_ptr<OperatorInfo>> &ops,
const size_t iter_ops);
std::vector<std::vector<int32_t>> PrepareBiasAdd(const std::vector<std::shared_ptr<OperatorInfo>> &ops,
const size_t iter_ops, std::vector<int32_t> s);
std::vector<std::vector<int32_t>> PrepareScalarInputOperator(const std::vector<std::shared_ptr<OperatorInfo>> &ops,
const size_t iter_ops, std::vector<int32_t> s);
std::vector<std::vector<int32_t>> PrepareOneHot(std::vector<int32_t> s);
std::vector<std::vector<int32_t>> MakeRecSearchStrategy(const std::shared_ptr<Graph> &graph,
const std::vector<std::shared_ptr<OperatorInfo>> &ops,

1
mindspore/ccsrc/parallel/auto_parallel/rec_core/rec_parse_graph.h
Unescape View File

@ -55,6 +55,7 @@ const std::map<std::string, OperatorType> DictOpType{
{"HSigmoid", OperatorType::kRecReLU},
{GELU, OperatorType::kRecReLU},
{TANH, OperatorType::kRecReLU},
{PRELU, OperatorType::kRecReLU},
{TENSOR_ADD, OperatorType::kRecElmWiseOp},
{SUB, OperatorType::kRecElmWiseOp},

Write Preview
Loading…
Cancel
Save