| /external/tensorflow/tensorflow/core/ops/compat/ops_history_v1/ |
| D | AllToAll.pbtxt | 2 name: "AllToAll"
39 name: "AllToAll"
|
| /external/tensorflow/tensorflow/core/api_def/base_api/ |
| D | api_def_AllToAll.pbtxt | 2 graph_op_name: "AllToAll"
|
| /external/tensorflow/tensorflow/compiler/mlir/xla/ir/ |
| D | hlo_ops_base.td | 673 string summary = "AllToAll";
676 AllToAll is a collective operation that sends data from all cores to all
687 be retrieved using ReplicaId op). AllToAll will be applied within
689 `replica_groups` = {{1,2,3}, {4,5,0}} means that an AllToAll will be applied
691 will be concatenated in the same order of 1, 2, 3. Then, another AllToAll
|
| D | hlo_ops.td | 343 // AllToAll). Non-positive channel_id handle is equivalent to no channel id.
|
| /external/tensorflow/tensorflow/compiler/xla/python/ |
| D | xla_client.py | 1026 def AllToAll(self, member in ComputationBuilder
1053 return ops.AllToAll(operand, split_dimension, concat_dimension, split_count,
|
| D | xla.cc | 823 ops.def("AllToAll", &AllToAll); in PYBIND11_MODULE()
|
| D | xla_client_test.py | 694 c.AllToAll(c.Constant(lhs), 0, 0)
|
| /external/tensorflow/tensorflow/compiler/xla/service/ |
| D | hlo_parser_test.cc | 1472 "AllToAll",
1473 R"(HloModule AllToAll
1475 ENTRY AllToAll {
|
| D | hlo_matchers.h | 207 HLO_MATCHER(AllToAll);
|
| D | hlo.proto | 131 // AllToAll). Non-positive channel_id is equivalent to no channel id.
|
| /external/tensorflow/tensorflow/compiler/xla/client/ |
| D | xla_builder_test.cc | 373 TEST_F(XlaBuilderTest, AllToAll) { in TEST_F() argument
376 AllToAll(x, /*split_dimension=*/1, /*concat_dimension=*/0, in TEST_F()
|
| D | xla_builder.h | 520 XlaOp AllToAll(XlaOp operand, int64 split_dimension, int64 concat_dimension,
928 friend XlaOp AllToAll(XlaOp operand, int64 split_dimension,
1681 XlaOp AllToAll(XlaOp operand, int64 split_dimension, int64 concat_dimension,
|
| D | xla_builder.cc | 2189 XlaOp XlaBuilder::AllToAll(XlaOp operand, int64 split_dimension, in AllToAll() function in xla::XlaBuilder
3314 XlaOp AllToAll(const XlaOp operand, int64 split_dimension, in AllToAll() function
3317 return operand.builder()->AllToAll(operand, split_dimension, concat_dimension, in AllToAll()
|
| /external/tensorflow/tensorflow/compiler/xla/g3doc/ |
| D | operation_semantics.md | 72 ## AllToAll section in Operation Semantics
75 [`XlaBuilder::AllToAll`](https://www.tensorflow.org/code/tensorflow/compiler/xla/client/xla_builder…
77 AllToAll is a collective operation that sends data from all cores to all cores.
90 retrieved using [`ReplicaId`](#replicaid)). AllToAll will be applied within
92 {4,5,0}}` means that an AllToAll will be applied within replicas `{1, 2,
94 in the same order of 1, 2, 3. Then, another AllToAll will be applied within
105 <b> `AllToAll(operand, split_dimension, concat_dimension, split_count,
135 AllToAll(x, /*split_dimension=*/1, /*concat_dimension=*/0, /*split_count=*/4);
|
| /external/tensorflow/tensorflow/tools/api/golden/v1/ |
| D | tensorflow.raw_ops.pbtxt | 84 name: "AllToAll"
|
| /external/tensorflow/tensorflow/tools/api/golden/v2/ |
| D | tensorflow.raw_ops.pbtxt | 84 name: "AllToAll"
|
| /external/tensorflow/tensorflow/go/op/ |
| D | wrappers.go | 13985 func AllToAll(scope *Scope, input tf.Output, group_assignment tf.Output, concat_dimension int64, sp… func
|
| /external/tensorflow/tensorflow/core/ops/compat/ |
| D | ops_history.v2.pbtxt | 1550 name: "AllToAll"
1587 name: "AllToAll"
|
| /external/tensorflow/tensorflow/core/ops/ |
| D | ops.pbtxt | 657 name: "AllToAll"
|