Change: 144623381

(can be used to have larger batch sizes for smaller input sequences.)
Change: 144621786

Change: 144609556

Change: 144588235

Change: 144584357

Change: 144575750

update callers in tensorflow.

Deprecated tf.initialize_all_tables.
Change: 144575599

Change: 144556462

- Rename gradients in the summary s/:/_
- tf.summary.histogram_summary -> tf.summary.histogram.
Change: 144544329

To avoid another GraphDef -> Graph -> GraphDef conversion, change Device::MaybeRewriteGraph to take a Graph instead of a GraphDef.

Use std::unique_ptr<> in more places to avoid some awkward .release() magic.
Change: 144532446

Change: 144510668

Bug fix in SavedModel shim's named signature up-conversion for classification signatures, where the value was not actually being set.
Change: 144505027

Change: 144502564

Change: 144497247

tensorflow::ops. Leave aliases in the tensorflow::ops for
compatibility until uses have been migrated to the new location.
Change: 144490912

Change: 144489466

Change: 144485165

Change: 144484604

This CL includes wrapper classes for the protocol buffer messages
`tensorflow::RunStepResponse` and `tensorflow::RunGraphResponse` (to
complement the corresponding request message wrappers that were added recently).

This change makes the backend code deal with abstract
`tensorflow::MutableRunStepResponseWrapper` and
`tensorflow::MutableRunGraphResponseWrapper` interfaces and adds three
concrete implementations of each interface:

* A mutable in-memory wrapper, which maintains the tensor data in
  `tensorflow::Tensor` objects, and provides the most efficient
  implementation when the client and master (or master and worker) or
  in the same address space.

* A mutable, owned protobuf wrapper, which has a similar implementation
  to today's client code.

* A mutable, non-owned protobuf wrapper, which has a similar
  implementation to today's server code (where the protobuf message is
  owned by the RPC subsystem).

This is another improvement for issue #6256.
Change: 144481118

Change: 144480955

Models the matrix cI, where c is a [batch] scalar, and I is the identity
matrix.
Change: 144479769

Change: 144479106

Change: 144478664

Change: 144478254

Change: 144477723

Enable record filtering while still allowing for downstream queuing to yield fixed sized filtered batches.
Change: 144477390

transformation of iid Student's t-distributions and should not be confused with
the Multivate Student's T-distribution (which is an example of an
https://en.wikipedia.org/wiki/Elliptical_distribution).
Change: 144474882

The intent is to produce a libtensorflow.tar.gz as part of the release
process to facilitate a leaner (no Python dependencies), smoother installation
process for TensorFlow usage in other languages.
Change: 144470928

Enable customization of architecture dependent compiler optimizations via configure, defaulting to -march=native.
Change: 144470919

Change: 144470271

Change: 144464718

Change: 144464664

Change: 144463826

Change: 144463291

Change: 144457649

Check for and fail immediately if the requested device does not exist.

Add a comment on DeviceMgr's constructor clarifying that it takes ownership of devices.
Change: 144457595

This is safe because this input alternative is currently unused; it was there only for completeness and for potential future uses.
Change: 144454504

Make weight broadcasting more strict (in particular, catch the case where a 1-dim in values shape could be broadcast across a non-1 dimension in weights).
Move broadcast ops to a separate file so they can be shared between `metrics` and `losses`.
Change: 144410889

Change: 144410076

Change: 144409845