| .. |
|
parallel_device
|
Parallel device: Minor rename+slight generalization to make it easier to create literal/constant parallel tensors
|
2021-02-08 17:12:35 -08:00 |
|
abstract_context.h
|
A very basic start on some op handler infrastructure
|
2020-11-02 09:45:24 -08:00 |
|
abstract_function.h
|
Implement MLIR and graph tracing using abstract interfaces shared with immediate execution mode.
|
2020-06-24 21:40:46 -07:00 |
|
abstract_operation.h
|
A very basic start on some op handler infrastructure
|
2020-11-02 09:45:24 -08:00 |
|
abstract_tensor_handle.cc
|
Add a debug string to AbstractTensorHandle
|
2020-12-10 15:24:54 -08:00 |
|
abstract_tensor_handle.h
|
Add a debug string to AbstractTensorHandle
|
2020-12-10 15:24:54 -08:00 |
|
BUILD
|
Move custom device placement from eager/execute.cc to c_api.cc. Then it can be reused by TFRT.
|
2021-02-08 17:30:34 -08:00 |
|
c_api_cluster_test.cc
|
Fix c_api_remote_test tsan flakiness.
|
2020-05-29 13:08:17 -07:00 |
|
c_api_debug_test.cc
|
Add Tensor & TensorHandle C APIs taking a Context
|
2020-04-06 15:10:09 -07:00 |
|
c_api_debug.cc
|
[TF2XLA] Remove unused and broken call to enable XLA, remove unneeded build macro
|
2020-08-27 11:25:21 -07:00 |
|
c_api_distributed_test.cc
|
Update function runtime to not run function optimization passes for component functions.
|
2020-10-08 11:17:43 -07:00 |
|
c_api_experimental_test.cc
|
Add functions that return device type and ID for eager.
|
2020-10-17 18:39:48 -07:00 |
|
c_api_experimental.cc
|
Switch cancellation manager to wrap/unwrap instead of pimpl
|
2021-02-02 10:56:02 -08:00 |
|
c_api_experimental.h
|
Allow components of parallel device tensors to have different shapes
|
2021-01-14 17:22:46 -08:00 |
|
c_api_internal.h
|
Remove the internal flag of lazy_remote_inputs_copy.
|
2021-01-07 10:42:36 -08:00 |
|
c_api_remote_function_test.cc
|
Fix an issue of out of order execution. For a multi-device function, don't send a packed input to the function device until all underlying remote handles are ready on remote devices. Otherwise, on a remote worker, a remote component function execution request could be enqueued before a request for producing a function input.
|
2020-12-11 14:56:03 -08:00 |
|
c_api_remote_test_util.cc
|
Fix an issue of out of order execution. For a multi-device function, don't send a packed input to the function device until all underlying remote handles are ready on remote devices. Otherwise, on a remote worker, a remote component function execution request could be enqueued before a request for producing a function input.
|
2020-12-11 14:56:03 -08:00 |
|
c_api_remote_test_util.h
|
Fix an issue of out of order execution. For a multi-device function, don't send a packed input to the function device until all underlying remote handles are ready on remote devices. Otherwise, on a remote worker, a remote component function execution request could be enqueued before a request for producing a function input.
|
2020-12-11 14:56:03 -08:00 |
|
c_api_remote_test.cc
|
Separate function tests from c_api_remote_test
|
2020-08-11 17:14:22 -07:00 |
|
c_api_test_util.cc
|
Support savedmodels that reference assets.
|
2020-09-18 14:29:39 -07:00 |
|
c_api_test_util.h
|
Support savedmodels that reference assets.
|
2020-09-18 14:29:39 -07:00 |
|
c_api_test.cc
|
Disable failing tests in c_api_test.cc
|
2021-01-20 09:49:36 -08:00 |
|
c_api_unified_experimental_eager.cc
|
Implement MLIR and graph tracing using abstract interfaces shared with immediate execution mode.
|
2020-06-24 21:40:46 -07:00 |
|
c_api_unified_experimental_graph.cc
|
Implements AbstractTensor::Shape(PartialTensorShape*).
|
2020-11-13 14:20:40 -08:00 |
|
c_api_unified_experimental_internal.h
|
Implements AbstractTensor::Shape(PartialTensorShape*).
|
2020-11-13 14:20:40 -08:00 |
|
c_api_unified_experimental_test.cc
|
Implements AbstractTensor::Shape(PartialTensorShape*).
|
2020-11-13 14:20:40 -08:00 |
|
c_api_unified_experimental.cc
|
Implements AbstractTensor::Shape(PartialTensorShape*).
|
2020-11-13 14:20:40 -08:00 |
|
c_api_unified_experimental.h
|
Implements AbstractTensor::Shape(PartialTensorShape*).
|
2020-11-13 14:20:40 -08:00 |
|
c_api.cc
|
Move custom device placement from eager/execute.cc to c_api.cc. Then it can be reused by TFRT.
|
2021-02-08 17:30:34 -08:00 |
|
c_api.h
|
Update lint check for ContextDevicePlacementPolicy since it has been moved.
|
2020-09-28 11:53:45 -07:00 |
|
custom_device_test.cc
|
Give custom devices the option to do type-based dispatch for ops with no explicit placement
|
2020-08-27 16:39:24 -07:00 |
|
custom_device_testutil.cc
|
Give custom devices the option to do type-based dispatch for ops with no explicit placement
|
2020-08-27 16:39:24 -07:00 |
|
custom_device_testutil.h
|
Give custom devices the option to do type-based dispatch for ops with no explicit placement
|
2020-08-27 16:39:24 -07:00 |
|
dlpack.cc
|
Fix multiple vulnerabilities in tf.experimental.dlpack.to_dlpack.
|
2020-09-18 19:19:43 -07:00 |
|
dlpack.h
|
fix
|
2020-03-28 08:41:24 +00:00 |
|
gradient_checker_test.cc
|
Disable TF32 in a few tests.
|
2021-01-08 15:50:46 -06:00 |
|
gradient_checker.cc
|
use outputs directly
|
2020-12-15 13:18:38 +07:00 |
|
gradient_checker.h
|
refactor gradient_check to use unified_api_testutil
|
2020-12-11 09:34:32 +07:00 |
|
gradients_internal.h
|
First pass at implementing a C++ API for gradients built on top of abstract interfaces.
|
2020-07-10 15:03:08 -07:00 |
|
gradients_test.cc
|
remove getValue in gradients_test
|
2021-01-15 03:05:17 +07:00 |
|
gradients_util.cc
|
Implements AbstractTensor::Shape(PartialTensorShape*).
|
2020-11-13 14:20:40 -08:00 |
|
gradients_util.h
|
Rename SparseSoftmaxCrossEntropyLoss to SparseSoftmaxCrossEntropyWithLogits to match up with the op name.
|
2020-09-17 17:53:59 -07:00 |
|
gradients.cc
|
Simplify C++ tape APIs to match the proposal in https://github.com/tensorflow/community/pull/335
|
2020-12-03 14:28:48 -08:00 |
|
gradients.h
|
Simplify C++ tape APIs to match the proposal in https://github.com/tensorflow/community/pull/335
|
2020-12-03 14:28:48 -08:00 |
|
immediate_execution_context.h
|
Move custom device placement from eager/execute.cc to c_api.cc. Then it can be reused by TFRT.
|
2021-02-08 17:30:34 -08:00 |
|
immediate_execution_distributed_manager.h
|
Integrate multi-client TFRT initialization with TFE_EnableCollectiveOps C API.
|
2021-02-02 11:36:26 -08:00 |
|
immediate_execution_operation.h
|
Move custom device placement from eager/execute.cc to c_api.cc. Then it can be reused by TFRT.
|
2021-02-08 17:30:34 -08:00 |
|
immediate_execution_tensor_handle.h
|
Add functions that return device type and ID for eager.
|
2020-10-17 18:39:48 -07:00 |
|
mnist_gradients_test.cc
|
move TestReluGrad to nn_grad_test
|
2020-12-18 00:25:24 +07:00 |
|
mnist_gradients_testutil.cc
|
move TestReluGrad to nn_grad_test
|
2020-12-18 00:25:24 +07:00 |
|
mnist_gradients_testutil.h
|
move TestReluGrad to nn_grad_test
|
2020-12-18 00:25:24 +07:00 |
|
tape.h
|
Simplify C++ tape APIs to match the proposal in https://github.com/tensorflow/community/pull/335
|
2020-12-03 14:28:48 -08:00 |
|
tfe_cancellation_manager_internal.h
|
Switch cancellation manager to wrap/unwrap instead of pimpl
|
2021-02-02 10:56:02 -08:00 |
|
tfe_context_internal.h
|
Split Abstract interfaces into Abstract and ImmediateExecution interfaces.
|
2020-06-17 12:39:13 -07:00 |
|
tfe_executor_internal.h
|
Attempting to split c_api_internal.h into separate files.
|
2020-04-21 11:04:58 -07:00 |
|
tfe_monitoring_internal.h
|
Attempting to split c_api_internal.h into separate files.
|
2020-04-21 11:04:58 -07:00 |
|
tfe_op_attrs_internal.h
|
Less pointer indirection for TFE_OpAttrs, add TFE_OpGetAttrs
|
2020-05-04 09:24:03 -07:00 |
|
tfe_op_internal.h
|
Split Abstract interfaces into Abstract and ImmediateExecution interfaces.
|
2020-06-17 12:39:13 -07:00 |
|
tfe_tensor_debug_info_internal.h
|
Attempting to split c_api_internal.h into separate files.
|
2020-04-21 11:04:58 -07:00 |
|
tfe_tensorhandle_internal.h
|
Split Abstract interfaces into Abstract and ImmediateExecution interfaces.
|
2020-06-17 12:39:13 -07:00 |
|
tracing_utils.cc
|
- Integrate C++ tape with op building APIs via TapeContext and TapeOperation which delegate calls to a parent execution context and record operations on the tape. Please see gradients_test.cc for usage.
|
2020-09-23 16:12:39 -07:00 |
|
tracing_utils.h
|
[Cleanup] Add a tracing::MaybeSetOpName utility to reduce the verbosity of C++ ops.
|
2020-09-23 09:18:26 -07:00 |
|
unified_api_test.cc
|
Fix unified_api_test
|
2020-12-11 09:34:32 +07:00 |
|
unified_api_testutil.cc
|
add GetValue and TestTensorHandleWithDimsInt to unified_api_testutil
|
2020-12-11 09:33:41 +07:00 |
|
unified_api_testutil.h
|
add GetValue and TestTensorHandleWithDimsInt to unified_api_testutil
|
2020-12-11 09:33:41 +07:00 |
