STT-tensorflow

benchmarks/resnet50

Use the new V2 optimizer in resnet50 benchmark.

2020-05-28 17:04:45 -07:00

memory_tests

Move keras related eager memory test to keras

2020-04-08 10:30:35 -07:00

backprop_test.py

Undo some of the composite tensor changes in gradient tape code

2020-06-22 13:02:52 -07:00

backprop_util.py

Mark bfloat16 as trainable.

2020-03-30 14:01:01 -07:00

backprop.py

Special case wrapping of ndarrays in the gradient tape code.

2020-06-22 17:38:38 -07:00

benchmarks_test_base.py

Fix benchmark naming in reporting

2020-05-11 15:30:15 -07:00

benchmarks_test.py

Reduce TensorShape.__init__ overhead by 50%.

2020-06-17 11:08:15 -07:00

BUILD

Special case wrapping of ndarrays in the gradient tape code.

2020-06-22 17:38:38 -07:00

cancellation_test.py

Add a SWIG wrapper for the tensorflow::CancellationManager class.

2019-07-03 08:28:39 -07:00

cancellation.py

Export the Eager classes and functions from C++ to Python with pybind11 instead of swig. This is part of a larger effort to deprecate swig and eventually with modularization break pywrap_tensorflow into smaller components. It will also make exporting C++ ops to Python significantly easier. XLA is using the pybind11 macros already. Please refer to https://github.com/tensorflow/community/blob/master/rfcs/20190208-pybind11.md for more information.

2019-12-17 19:40:40 -08:00

context_test.py

Add a functiondef getter to the context

2020-02-19 10:48:38 -08:00

context.py

Merge pull request #40473 from guillaumekln:fix-executing-eagerly-format

2020-06-16 11:41:10 -07:00

core_test.py

Allow dynamically configuring device placement

2020-05-12 23:17:39 -07:00

core.py

Export the Eager classes and functions from C++ to Python with pybind11 instead of swig. This is part of a larger effort to deprecate swig and eventually with modularization break pywrap_tensorflow into smaller components. It will also make exporting C++ ops to Python significantly easier. XLA is using the pybind11 macros already. Please refer to https://github.com/tensorflow/community/blob/master/rfcs/20190208-pybind11.md for more information.

2019-12-17 19:40:40 -08:00

custom_device_test.py

Add a way to register custom devices with the Python TFE_Context

2020-04-09 17:50:19 -07:00

custom_device_testutil.cc

throwTypeError -> ThrowTypeError

2020-04-16 10:46:01 -07:00

def_function_test_cpu_only.py

Explicitly fail when experimental_compile=True is used, but XLA is not enabled

2020-01-28 15:27:49 -08:00

def_function_test.py

Retrace tf.function when using different variables

2020-06-22 11:01:35 -07:00

def_function_xla_jit_test.py

Rollback of rollback: [TF/XLA] Only force retracing for non-unique XLA context ID for TPUReplicatedContext

2020-06-18 11:07:20 -07:00

def_function_xla_test.py

…

def_function.py

Use trace.Trace instead of traceme.TraceMe in python

2020-06-12 19:48:10 -07:00

device_placement_test.py

Fix device placement logic in ConvertToEagerTensor.

2020-02-25 12:23:56 -08:00

execute.py

Export the Eager classes and functions from C++ to Python with pybind11 instead of swig. This is part of a larger effort to deprecate swig and eventually with modularization break pywrap_tensorflow into smaller components. It will also make exporting C++ ops to Python significantly easier. XLA is using the pybind11 macros already. Please refer to https://github.com/tensorflow/community/blob/master/rfcs/20190208-pybind11.md for more information.

2019-12-17 19:40:40 -08:00

executor.py

Export the Eager classes and functions from C++ to Python with pybind11 instead of swig. This is part of a larger effort to deprecate swig and eventually with modularization break pywrap_tensorflow into smaller components. It will also make exporting C++ ops to Python significantly easier. XLA is using the pybind11 macros already. Please refer to https://github.com/tensorflow/community/blob/master/rfcs/20190208-pybind11.md for more information.

2019-12-17 19:40:40 -08:00

forwardprop_test.py

Merge pull request #39711 from abhichou4:tests/tangent-batch

2020-05-29 14:01:11 -07:00

forwardprop_util.py

minor spelling tweaks

2020-02-11 15:09:21 +09:00

forwardprop.py

Forwardprop: opt the forwardprop utility function out of run_functions_eagerly

2020-05-05 12:05:03 -07:00

function_argument_naming_test.py

…

function_defun_collection_test.py

…

function_gradients_test.py

Make device listing APIs non-experimental

2019-11-05 02:14:04 -08:00

function_test.py

Support packed variable for tf data captured function.

2020-06-22 16:24:37 -07:00

function.py

Special case tfnp ndarrays in some places, and minor improvements to tfnp array

2020-06-22 15:10:53 -07:00

gradient_input_output_exclusions_test.py

Make sure the gradient_input_output_exclusions_test is resilient to windows line endings.

2020-02-13 09:13:58 -08:00

gradient_input_output_exclusions.py

Fix the string representation of string literal subscripts to be x['a'] instead of x[a].

2020-05-25 17:12:18 -07:00

graph_only_ops_test.py

Delete unused graph_zeros_like

2019-10-05 02:11:53 -07:00

graph_only_ops.py

[tfdbg2] Ensure that op_callbacks capture Placeholders for tf.functions

2020-01-20 17:20:00 -08:00

imperative_grad.py

Export the Eager classes and functions from C++ to Python with pybind11 instead of swig. This is part of a larger effort to deprecate swig and eventually with modularization break pywrap_tensorflow into smaller components. It will also make exporting C++ ops to Python significantly easier. XLA is using the pybind11 macros already. Please refer to https://github.com/tensorflow/community/blob/master/rfcs/20190208-pybind11.md for more information.

2019-12-17 19:40:40 -08:00

lift_to_graph_test.py

Fix a bug in lift_to_graph.py.

2020-06-11 02:32:40 -07:00

lift_to_graph.py

Fix a bug in lift_to_graph.py.

2020-06-11 02:32:40 -07:00

monitoring_test.py

Make eager/monitoring_test.py less flaky.

2020-06-08 14:44:00 -07:00

monitoring.py

add a context manager and function decorator for monitoring time.

2020-05-26 10:18:03 -07:00

ops_test.py

Make experimental_ref no longer experimental

2020-02-10 10:56:24 -08:00

profiler_client_test.py

…

profiler_client.py

allow profiler_client.py v2 (our rpc client) to provide ProfilerOptions,

2020-04-13 14:24:44 -07:00

profiler_test.py

Use trace.Trace instead of traceme.TraceMe in python

2020-06-12 19:48:10 -07:00

profiler.py

profiler session allow more options to be specified at python level. such as host trace level and python tracer enable.

2020-04-01 14:39:29 -07:00

pywrap_gradient_exclusions.cc

Add StatelessParameterizedTruncatedNormal sampler.

2020-06-18 10:28:38 -07:00

pywrap_gradient_exclusions.h

Simplify gradient exclusions data to speed up compilation w/ clang on windows.

2020-02-13 14:19:06 -08:00

pywrap_tensor_conversion.cc

Avoid pointer indirection in handle, context & op

2020-04-28 18:58:17 -07:00

pywrap_tensor_conversion.h

Add layer of indirection for Tensor & TensorHandle

2020-01-09 08:43:03 -08:00

pywrap_tensor.cc

Raise error type corresponding to status code instead of generic RuntimeError.

2020-06-19 13:27:01 -07:00

pywrap_tensor.h

Inline status in EagerTensor

2019-09-04 19:55:01 -07:00

pywrap_tfe_src.cc

Split Abstract interfaces into Abstract and ImmediateExecution interfaces.

2020-06-17 12:39:13 -07:00

pywrap_tfe_test.py

Remove Keras dependency from tfe test.

2020-04-07 20:22:09 -07:00

pywrap_tfe.h

Support running a tf.function with packed variable inputs both locally and remotely.

2020-05-18 15:21:13 -07:00

remote_benchmarks_test.py

Fix remote benchmarks with explicit execution_mode.

2020-01-28 12:12:55 -08:00

remote_cloud_tpu_test.py

Add test for connecting to v2-32 Cloud TPU.

2020-02-25 15:34:17 -08:00

remote_cluster_test.py

Use coordinator.join instead of joining threads to make sure thread exceptions are reliably thrown.

2020-06-04 18:21:02 -07:00

remote_execution_test.py

Flip the ordering of test context managers so that parameterized testcases can be correctly processed.

2020-03-06 14:40:15 -08:00

remote_test.py

Support running a tf.function with packed variable inputs both locally and remotely.

2020-05-18 15:21:13 -07:00

remote.py

Fix documentation for connect_to_remote_host

2020-02-04 19:24:27 -08:00

tape_test.py

Do not create an extra GradientTape in custom_gradient/recompute_grad. The extra tape leads the tf.function gradient code to believe that the user intends to compute higher order derivatives. This requires generating a forward function with all possible side outputs which is expensive. This also doesn't work well with control flow and causes the added test to fail.

2020-04-17 19:58:53 -07:00

tape.py

Do not create an extra GradientTape in custom_gradient/recompute_grad. The extra tape leads the tf.function gradient code to believe that the user intends to compute higher order derivatives. This requires generating a forward function with all possible side outputs which is expensive. This also doesn't work well with control flow and causes the added test to fail.

2020-04-17 19:58:53 -07:00

tensor_test.py

Allow floats and ints to be converted to complex numbers.

2020-06-02 10:45:38 -07:00

test.py

…

wrap_function_test.py

Replace == with is in test comparing variable identity.

2019-08-15 13:37:05 -07:00

wrap_function.py

Internal clean-up: move signature handling code that's specific to WrapFunction into WrapFunction subclass.

2020-04-11 09:43:35 -07:00