Change: 125604684

Change: 125601282

- Earlier we were locally copying the map which could be potentially very big.
Change: 125595701

Factors out the interface for control flow ops Enter, Exit, Switch, Merge and NextIteration. Registers the Gather op for quantized parameters.
Change: 125593594

in a subdirectory under the workspace.
Change: 125590691

Change: 125589121

Change: 125584342

Change: 125584181

Change: 125582830

When the range can't be satisfied, the GCS server returns the 416 HTTP code and
some additional information in the body. The curl library expects the exact amount of
bytes in the body as requested in the range, so curl_easy_peform fails with a
message like "Failed writing body (12 != 171)". With this fix, we don't check the
return value of curl_easy_peform if the HTTP code was 416.
Change: 125579779

Change: 125578894

Change: 125575345

and tensorflow/core/distributed_runtime:remote_device_test* from opensource build.
Change: 125573873

local_init_op are run by each worker, while the init_op is only run by the chief.
Fix __all__

Added initial README.md
Change: 125573668

Change: 125572152

Change: 125571683

Fixing bug in which the default value of the init_op was checked against None rather than _USE_DEFAULT. This wasn't caught as the value of _USE_DEFAULT matched supervisor._USE_DEFAULT. A unit test has been added which fails for the old code but succeeds with the new code.
Change: 125564252

Use a function to compute dimensions for backpropagation, rather than using a macro. Return dimensions using a struct rather than individual ints.
Change: 125556701

Fixes #2974
Change: 125530598

Change: 125511710

Change: 125510452

Change: 125508441

Change: 125496972

Add partitioner that partitions a variable such that no partition gets less than the given minimum size of chunk.

Swith from 'layers.legacy_fully_connected' to 'layers.fully_connected' in _DNNLinearCombinedBaseEstimator. Also, partition the weights of the fully connected layers using the added partitioner.
Change: 125492475

* optimize_loss expects 0d Tensor.
* Estimator.get_eval_ops requires to have `targets` not None.
* Adding asserts for shapes in rnn_cells.
* Added error check in rnn.rnn.
Change: 125491759

Change: 125491544

Added tf-storage component for storing TensorBoard URI state and modified tf-tensorboard.html to use it. Added tf-globals component to hold TensorBoard global variables.
Change: 125489418

Change: 125487857

improves by 56%: from 148 to 231 in the forward-backward pass.

Support both the fastest algorithm, and fall back to the fastest algorithm
without using any scratch memory, if the first algorithm fails scratch memory
allocation.

Soumith's conv-benchmarks stay the same before and after this change. But now
it can run with bigger batch size.
Change: 125484122

Change: 125476357

Also, if the TAG is not found, log an info rather than warning.
Change: 125476271

Change: 125475913

Change: 125472018

Also update RELEASE.md to forward port from 0.9 branch, and add
this new 'breaking change to the API.'
Change: 125465285

Change: 125463524

Change: 125460867

Change: 125458571

Change: 125456104

Change: 125453662

Change: 125452820