| a4f1478134cdbf73f0ad7eda3e73407135a54566 |
|
19-Feb-2018 |
Michael Zhou <michaelchouqj@gmail.com> |
make "smart_cond" api public and reusable (#13954)
* make smart_cond api public and reusable
* adjust function comment space
* change `smart_cond` naming schema consistent with `cond`
* Move `smart_cond` and `smart_constant_value` to tf.contrib.framework
* resolve loop import
* remove redundant constant_value
* fix function name typo
* Fix function parameter name usage
* Fix function parameter name usage(2)
* To support 0/1 interger pred
* check integer value after type check
* Lint fixes
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| b37913a6d1a2327b3aebcef857638e4ad2f465c3 |
|
09-Feb-2018 |
Jacques Pienaar <jpienaar@google.com> |
Fix lint warnings. NFC
PiperOrigin-RevId: 185167035
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 23851760b7b099214bdd4f1b88156d7ac2bdd2a2 |
|
05-Feb-2018 |
Skye Wanderman-Milne <skyewm@google.com> |
Get control_flow_ops.py ready to support de/serializing nested control flow.
With this change, ControlFlowContexts keep track of their nested
contexts (the reverse lookup as ControlFlowContext.outer_context).
This is to enable de/serializing the nested contexts of each "root"
context, and only adding the root contexts to collections. This allows
for simple deserialization of each root context by recursively
deserializing its nested contexts.
The de/serialization logic is disabled and the corresponding
control_flow.proto changes are omitted for now for forwards
compatability (i.e. three-week-old binaries must be ready to accept
the new proto format once its commited). After this is committed for
three weeks, I'll commit a follow-up change enabling the new behavior.
Design note: I chose to serialize the nested contexts, rather than the
outer contexts, because it makes it easy to deserialize the contexts
in topological order and to assign the right outer context. If we
serialized the outer contexts, there'd need to be some mechanism for
either sorting all the serialized contexts first, or deserializing all
of them and then doing another pass to assign the outer contexts.
PiperOrigin-RevId: 184533406
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 216878ea3dafdc5fbe6a15d389edb003ad2fd4b4 |
|
13-Dec-2017 |
Alexander Gorban <gorban@google.com> |
Simplify tf.case implementation.
PiperOrigin-RevId: 178853258
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 9810da1b87aae689cac42bae754e7e4cb5a99d57 |
|
12-Dec-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Adds XLA support for tf.nn.dynamic_rnn
Changes tf.nn.dynamic_rnn to specify `maximum_iterations` argument for the while_loop.
When `maximum_iterations` argument is supplied to tf.while_loop, use this to provide an upper bound on the size of Stacks used for gradient computation.
By specifying the stack limit we can generate gradient code for while loops that uses fixed shape TensorArrays and hence can be compiled with XLA.
PiperOrigin-RevId: 178802710
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 28807c5666c9f574ef415fed7b18b99ebed41ecc |
|
09-Dec-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Add Operation._remove_all_control_inputs and use in ControlFlowContext.
This allows while loop gradients to work with the C API. This change
also enables the C API for control flow tests.
PiperOrigin-RevId: 178438424
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| bcdcb78854e8dfa1b2eda813b9e2910df522abb4 |
|
07-Dec-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Fix tests in control_flow_ops_test.py to not access Tensor._shape
Accessing _shape directly doesn't work with the C API enabled. Rewrite
these tests to use public APIs instead.
PiperOrigin-RevId: 178162875
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 01365dbc2c257ff2ab409a2a5122a06739272737 |
|
24-Oct-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Allow lists to be passed to tf.group().
PiperOrigin-RevId: 173308794
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 48591e00fe917bfeecc31e501ef133447b81e161 |
|
24-Oct-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Better error message if you pass a list to tf.group().
PiperOrigin-RevId: 173260210
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 728e238d26669a358ff296364b83325ce0e14c34 |
|
28-Sep-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Enable _USE_C_API for CondTest in control_flow_ops_test.py
The only change required to make CondTest pass is changing the various
Operation methods to check self._c_op to determine if the C API is
enabled, instead of self._graph._c_graph or _USE_C_API. This is
because CondContext.AddOp() is called before creating self._c_op in
Operation.__init__(), and AddOp() uses the Operation methods that call
the C API. We need to use the original Python-only code before
self._c_op has been created. I added a comment in ops.py explaining an
alternative to this solution that we may wish to implement later.
PiperOrigin-RevId: 170350199
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 05c9966edafea9db76a3e275d09f078070680d6d |
|
07-Sep-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Adding test for tf.cond with a mutable boolean pred.
PiperOrigin-RevId: 167844952
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 6f898c6b2cbbc257d0966ee313a3670e88919463 |
|
05-Aug-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Make "default" argument for tf.case optional.
PiperOrigin-RevId: 164368207
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 8a90e8487d39cedc8fdd2cbab5a6c237c2f3a1cb |
|
18-Jul-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Fixes in control_flow_ops.
PiperOrigin-RevId: 162325768
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 068781de74e4623297943fd31856d1b918c69a4e |
|
23-Jun-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Modify ControlFlowContext to also respect import_scope in 'values_' and keys of 'external_values_'
PiperOrigin-RevId: 159985290
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 7c561e09c05100fe68f00d66a2d27d1b490ee74e |
|
02-May-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Explain when callables passed to tf.cond & tf.while_loop are run.
Rename the parameters to tf.cond.
Change: 154774725
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 2f9421cfc449d2cad8bdbbce786bcd07d462de66 |
|
27-Apr-2017 |
Tim Harley <tharley@google.com> |
Make switches outside of a cond context differentiable.
Change: 154441734
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| d3aa99075fdd9b5ce0d65537e12af4e3a448d18a |
|
11-Apr-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Fixes tf.cond with nested return values inside while loops.
Change: 152830112
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 3aaf220fef04c1ab71301c4fc76739c6a4df605e |
|
08-Apr-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
- Adds support for nested types in tf.case and tf.cond.
- Adds a "strict" mode which disables silent unpacking of singleton lists.
- Adds shape inference to tf.case.
- Adds a lot of unit tests.
Change: 152581097
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| a1d2a4ab90bd9df7312408f3971a2236810a1074 |
|
13-Feb-2017 |
Alexandre Passos <apassos@google.com> |
Backprop through resource handles.
Change: 147371282
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| fbf32f8cccdb9879e460ecf13293a1b846e9cecd |
|
02-Feb-2017 |
Alexandre Passos <apassos@google.com> |
Handle resource variable reads inside loops.
Change: 146314715
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| a9e161988996416f37b3cfce3d56c1e8d9dcd5a0 |
|
20-Dec-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Change `with_dependencies` to handle any iterable for `dependencies` arg.
Change: 142591245
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 58201a058853de647b37ddb0ccf63d89b2357f03 |
|
17-Dec-2016 |
Justine Tunney <jart@google.com> |
Remove hourglass imports from even more tests
Change: 142318245
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| b183b6d0c8ff5ddf0cc48f21fc03341bf56c75f7 |
|
16-Dec-2016 |
Andrew Selle <aselle@google.com> |
Change remaining tf.mul -> tf.multiply, tf.neg -> tf.negative, and tf.sub -> tf.subtract
tf.negative, tf.multiply, tf.subtract are the new names
Change: 142257628
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 35a58c8141a13bf78fd88a4bc13a6181f815a2e3 |
|
10-Dec-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Change references to TensorArray.pack, TensorArray.unpack to TensorArray.stack
and TensorArray.unstack since pack and unpack are getting deprecated. Also, I switched a few references to tf.pack/tf.unpack to tf.stack and tf.unstack.
Change: 141619037
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| cbd3cacfb73bbea912b9d01c2540187684f751a7 |
|
03-Nov-2016 |
Illia Polosukhin <ipolosukhin@google.com> |
Replace usages initialize_all_variables -> global_variables_initializer
Change: 138128703
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| e91ea774cfce323ef5812274f92fa2f13d2f3637 |
|
13-Sep-2016 |
Sherry Moore <sherrym@google.com> |
Added to_proto and from_proto for CondContext and WhileContext.
Change: 133013495
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| e98f46a30b58e086bf2dacaf371acf7e5a5b469c |
|
06-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Fix AddBackPropIndexedSlicesAccumulator for partially defined shapes and non-float32 dtypes.
Change: 132333359
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| c46abae176717ef6c6649ba0ca1099c35b90194e |
|
25-Aug-2016 |
Yuan Yu <yuanbyu@google.com> |
Strictly enforce shape invariance for while loop. TF Python frontend has increasingly been taking advantage of shape information in graph construction. This could lead to incorrect graph construction for while loop if the shapes of the loop variables are not loop invariants. This CL implements the following solution.
We add a new optional argument `shape_invariants` to tf.while_loop. If `shape_invariants` is None, we consider the initial shapes of `loop_vars` as the shape invariants for the loop. We statically check and raise an error if the shapes of `loop_vars` are not shape invariants. In that case, users are required to provide shape invariants, which we also check statically. We expect that in most cases the initial shapes of loop_vars are loop invariants and therefore there is no need to provide shape_invariants.
NOTE. This can be a *breaking* change. There are unfortunately cases that we have to provide the shape invariants.
NOTE. The deprecated `While` is removed. Use `tf.while_loop`.
Change: 131324247
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 92527942d7d55d5e20f16f94d58483178f377157 |
|
22-Aug-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Split NodeDef out of graph.proto into node_def.proto. Needed so we
can use NodeDef in FunctionDef.
Change: 130982373
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| ec225aea6b899f187e68251be194e70f6ccb2440 |
|
08-Jul-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Fix IndexedSlices gradients within a tf.while_loop
Change: 126924005
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 0cf9ed3a719c0782695154d5a0bca260001cec15 |
|
02-Jun-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Update copyright for 3p/tf/python.
Change: 123900456
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 3402f51ecd11a26d0c071b1d06b4edab1b0ef351 |
|
18-Apr-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Fix for IndexedSlices gradient accumulation in while loop.
Change: 120136109
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 9394661924a37fe17e5f11f43833e9bbffbf23ff |
|
23-Mar-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Fix dense_shape check.
This rectifies the following error:
TypeError: Using a `tf.Tensor` as a Python `bool` is not allowed. Use `if t is
not None:` instead of `if t:` to test if a tensor is defined, and use the
logical TensorFlow ops to test the value of a tensor.
when the conditional branch contains a tf.IndexedSlices object with
dense_shape=tf.constant(...).
Change: 117937593
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 7760ce56fc3ab4ab8cdc408e29d8ad8b539c417e |
|
11-Feb-2016 |
Josh Levenberg <josh11b@tensorflow.org> |
Get rid of some import cruft.
Change: 114374558
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| 9c3043ff3bf31a6a81810b4ce9e87ef936f1f529 |
|
20-Nov-2015 |
Manjunath Kudlur <keveman@gmail.com> |
TensorFlow: Improve performance of Alexnet
Changes:
* error message that refers to removed `DefaultSession` method.
* -Wnull-conversion warnings
* the "_start_time" attr for recvs when the flag "--brain_enable_scheduling_for_recvs" is set.
* typo in tutorial data download progress message.
* a typo ("however their installing"=>"however installing").
* typo, rename "TensorFlow Mechanics" to "How To" to be consistent with the website.
* a typo ("subtact"=>"subtract").
* protobuf examples in comments in tensorflow::Example.proto.
* formula formatting in MNIST beginner tutorial
* negative fraction-of-queue-full stats
* protobuf inclusion path so that Android demo will build under Blaze.
* small typo (moderatly > moderately)
* Session.run() to check that tensor arguments come from the session's graph.
* another six import
* seq2seq typo in bazel command
Base CL: 108349164
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| f2102f4e2c1c87f1d1bf9ab856a2849c54478760 |
|
12-Nov-2015 |
Vijay Vasudevan <vrv@google.com> |
TensorFlow: upstream changes from the afternoon.
Changes:
- futurize --stage2 changes for Python 3 compatibility by @girving.
- Small updates to documentation by @vrv, schuster and others
- Account for failure of std::thread::hardware_concurrency by @ebrevdo.
- More changes for backwards-compatibility tests by Josh
- Updates to python op doc generation by Josh
- Added support for using the best-fit allocator via ConfigProto by @vrv.
- Rename LocalSession to DirectSession, since local was a bad name for
it.
- Enable tf.nn.moments() to work with tensors of unknown shape by @mrry.
GITHUB_ISSUE: 139
- Changes for Android build by Andrew.
Base CL: 107645181
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|
| f41959ccb2d9d4c722fe8fc3351401d53bcf4900 |
|
07-Nov-2015 |
Manjunath Kudlur <keveman@gmail.com> |
TensorFlow: Initial commit of TensorFlow library.
TensorFlow is an open source software library for numerical computation
using data flow graphs.
Base CL: 107276108
/external/tensorflow/tensorflow/python/ops/control_flow_ops_test.py
|