| Commit message (Collapse) | Author | Age |
|
|
|
| |
PiperOrigin-RevId: 208260479
|
|
|
|
|
|
| |
loops.
PiperOrigin-RevId: 208260082
|
|
|
|
| |
PiperOrigin-RevId: 208258892
|
|
|
|
| |
PiperOrigin-RevId: 208257223
|
|\
| |
| |
| | |
PiperOrigin-RevId: 208257112
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Rollback breaking change.
END_PUBLIC
Automated rollback of commit 20622daa9db9d403a66111f5682367566e2bb8db. Revert #21038.
PiperOrigin-RevId: 208256026
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Fixes a bug in which the arg and retval node indices weren't rewritten when
a function was repartitioned by a StatefulPartitionedCallOp. Repartitioning can
occur when the kernel is visited with a previously unseen
FunctionLibraryRuntime.
PiperOrigin-RevId: 208255736
|
| |
| |
| |
| | |
PiperOrigin-RevId: 208255199
|
|\ \
| | |
| | |
| | | |
PiperOrigin-RevId: 208255132
|
| | |
| | |
| | |
| | | |
PiperOrigin-RevId: 208255119
|
| | |
| | |
| | |
| | | |
PiperOrigin-RevId: 208254124
|
|\ \ \
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 208253747
|
|\ \ \ \
| | | | |
| | | | |
| | | | | |
PiperOrigin-RevId: 208251834
|
|\ \ \ \ \
| | | | | |
| | | | | |
| | | | | | |
PiperOrigin-RevId: 208250121
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Allow a different output shape from the input in tf.contrib.image.transform (#17011).
END_PUBLIC
RELNOTES: Allow a different output shape from the input in tf.contrib.image.transform.
Thanks qyu@ for making the original change and fixing a few other prior issues!
Automated rollback of commit 07fdb697d33478d7a72d09fc2371fa834e870b83
PiperOrigin-RevId: 208248183
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Fix by holding reference to passed in model_content and relying on
immutability of PyString (pointer doesn't change).
PiperOrigin-RevId: 208244957
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
built manually.
PiperOrigin-RevId: 208244600
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | |
| | | | | | | |
PiperOrigin-RevId: 208244350
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
tensorflow::string is sometimes ::string and sometimes std::string, which
makes code that uses both subtly dangerous. For example,
FactoryKeyCreator::AddAsKey() has an overload for tensorflow::string but had
many callsites passing a std::string, causing incorrect behavior on
the google platform.
PiperOrigin-RevId: 208244169
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
PiperOrigin-RevId: 208243676
|
|\ \ \ \ \ \ \
| | | | | | | |
| | | | | | | |
| | | | | | | | |
PiperOrigin-RevId: 208238767
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Fixes #19062
PiperOrigin-RevId: 208235214
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
PiperOrigin-RevId: 208232511
|
|\ \ \ \ \ \ \ \
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
PiperOrigin-RevId: 208232199
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
PiperOrigin-RevId: 208229302
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
PiperOrigin-RevId: 208223617
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
PiperOrigin-RevId: 208220542
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
Modify ShapeVerifier::HandleRng to detect illegal Rng instructions.
Add test cases to test the handling of Rng instructions in the HLO verifier.
Modify the document for XLA random operations to reflect the supported data
types.
PiperOrigin-RevId: 208220062
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
PiperOrigin-RevId: 208219138
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
PiperOrigin-RevId: 208218148
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
There are few issues we will need to address in future.
1. TF RNN cell currently accept callable only as activation function, while keras allows string and convert to activation function. This cause a weird backward dependency from tf.nn.op to tf.keras. It will be nice to have a unified place for activation functions within TF.
2. There is a name conflict for few cells between Keras and TF, namely LSTMCell and GRUCell. The serial/de-serialization need special care since the tf.nn.rnn_cell is not visible to Keras.
Update TF LSTM Cell to support proper serialization.
PiperOrigin-RevId: 208217949
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
PiperOrigin-RevId: 208200028
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
PiperOrigin-RevId: 208178874
|
| |_|_|_|/ / / /
|/| | | | | | |
| | | | | | | |
| | | | | | | | |
`find_vc_path(repository_ctx)` returns `None` if VC path is not found.
@gunan
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
PiperOrigin-RevId: 208172537
|
|\ \ \ \ \ \ \ \
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
PiperOrigin-RevId: 208160689
|
|\ \ \ \ \ \ \ \ \
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
yongtang:21435-sparse_softmax_cross_entropy_with_logits-doc
PiperOrigin-RevId: 208154937
|
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
PiperOrigin-RevId: 208153363
|
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
PiperOrigin-RevId: 208153311
|
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
types of py_func-wrapping functions.
PiperOrigin-RevId: 208152501
|
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
PiperOrigin-RevId: 208146858
|
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
PiperOrigin-RevId: 208146417
|
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
PiperOrigin-RevId: 208141063
|
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
PiperOrigin-RevId: 208140215
|
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
managers
Allows custom checkpoint numbering by providing a way to save which doesn't append a save_counter.
A followup will add tf.contrib.checkpoint.CheckpointManager for cleaning up old checkpoints. Neither save() nor write() delete checkpoints after this CL.
RELNOTES: tf.train.Checkpoint does not delete old checkpoints by default.
PiperOrigin-RevId: 208138796
|
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
of keep_dims in a backwards compatible manner.
PiperOrigin-RevId: 208137773
|
|\ \ \ \ \ \ \ \ \ \
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
PiperOrigin-RevId: 208136518
|
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
PiperOrigin-RevId: 208135589
|
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
#20451
#19607
PiperOrigin-RevId: 208135233
|
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
PiperOrigin-RevId: 208133502
|