STT-tensorflow/tensorflow/core/grappler
Bruce Fontaine c27b834b49 Wrap save/restore logic in tf.function when in eager mode. This allows parallel saving and restoring when using multiple devices.
PiperOrigin-RevId: 317719780
Change-Id: Ifb7e34f708da4121b49fb38d8dad046d45fedc42
2020-06-22 13:23:14 -07:00
..
clusters [BUILD] Clean up unused aliases left over the "common_runtime/BUILD" split. 2020-04-03 11:07:56 -07:00
costs Merge pull request #40496 from tg-at-google:sign-compare-warning-fixes-batch-2 2020-06-17 11:22:40 -07:00
graph_analyzer Cleanup of transitive-fanin node computation to make details of invalid input available to its callers. 2020-04-20 16:18:52 -07:00
inputs Qualify uses of std::string 2020-05-13 06:55:22 -07:00
optimizers Wrap save/restore logic in tf.function when in eager mode. This allows parallel saving and restoring when using multiple devices. 2020-06-22 13:23:14 -07:00
utils Merge pull request #40496 from tg-at-google:sign-compare-warning-fixes-batch-2 2020-06-17 11:22:40 -07:00
verifiers Export grappler code from C++ to Python with pybind11 instead of swig. Adding pywrap_required_hdrs to downstream grappler build files which will go away with bazel first class shared library support. 2020-02-03 13:13:08 -08:00
BUILD [BUILD] Clean up unused aliases left over the "common_runtime/BUILD" split. 2020-04-03 11:07:56 -07:00
devices.cc Merge pull request #36639 from ROCmSoftwarePlatform:google-upstream-grappler 2020-03-11 23:59:13 -07:00
devices.h
graph_topology_view_test.cc
graph_topology_view.cc
graph_topology_view.h
graph_view_test.cc [Grappler] Allow feeding non-zero index tensors to Grappler optimizer 2020-04-27 11:45:41 -07:00
graph_view.cc
graph_view.h
grappler_item_builder_test.cc Remove using _output_shapes for shape annotation. 2019-10-22 09:56:32 -07:00
grappler_item_builder.cc Remove all uses of TENSORFLOW_LITE_PROTOS. 2020-05-06 02:43:41 -07:00
grappler_item_builder.h
grappler_item_test.cc
grappler_item.cc Cleanup of transitive-fanin node computation to make details of invalid input available to its callers. 2020-04-20 16:18:52 -07:00
grappler_item.h Cleanup of transitive-fanin node computation to make details of invalid input available to its callers. 2020-04-20 16:18:52 -07:00
mutable_graph_view_test.cc minor spelling tweaks 2020-02-27 15:42:16 +09:00
mutable_graph_view.cc Merge pull request #36052 from gaurav1086:optimize_expression 2020-01-24 17:02:31 -08:00
mutable_graph_view.h minor spelling tweaks 2020-02-27 15:42:16 +09:00
op_types.cc [Grappler] Add arithmetic optimizer stage for optimizing tf.nn.embedding_lookup_sparse(). 2020-04-24 15:40:51 -07:00
op_types.h [Grappler] Add arithmetic optimizer stage for optimizing tf.nn.embedding_lookup_sparse(). 2020-04-24 15:40:51 -07:00
utils_test.cc Correct #if 2020-03-11 14:33:23 -07:00
utils.cc Merge pull request #40496 from tg-at-google:sign-compare-warning-fixes-batch-2 2020-06-17 11:22:40 -07:00
utils.h [TensorFlow PE] Fix static analysis warnings in grappler: std::set<NodeDef*> does not make sense, since ordering of pointers is not deterministic. Switch to using absl::flat_hash_set, which is also faster. Add a method GetOutputsOrderedByNodeName for uses where deterministic ordering is required. 2020-04-02 18:11:47 -07:00