We create a reservoir that organizes health pill values by node name. Also added a
method onto the event multiplexor to fetch the health pills per run.
Change: 146272468

Change: 146272519

…py insertion HLO passes. Switches implementation to use simple expression tree matching, and reduces the number of patterns to match to those that we care about in production (i.e. post fusion and copy insertion).

// ExecuteThunks runtime FRNN forward pass (sequence_length=30)  WhileThunk: 112ms

// ExecuteThunks runtime FRNN forward pass (sequence_length=30)  ForThunk:     6ms

// ExecuteThunks runtime FRNN forward pass (sequence_length=100) WhileThunk: 585ms

// ExecuteThunks runtime FRNN forward pass (sequence_length=100) ForThunk:   191ms
Change: 146273048

Helps with tensorflow#6267
Change: 146277178

…leak.

Change: 146278252

Resource variables are quite new, so this change should not affect any preexisting models. Follows up on the fix for Optimizers dealing with sparse gradients which have repeated indices.
Change: 146279378

…ch is not numerically accurate for x near 0).

Make some cleanups to unary_ops_test.py.
Change: 146282294

…ll actions: training, evaluation, inference, serving.

Change: 146285386

- Create the directory if it doesn't exist
- Output formatted code
Change: 146286990

Change: 146288841

Change: 146289701

…delete unused methods in contrib metrics.

Change: 146294596

Change: 146298002

- Register attribute and op renames
Change: 146298778

- Marked control flow ops as hidden
Change: 146300232

Change: 146300684

Change: 146301808

Change: 146305718

Change: 146308381

Comment out all the overrides except what was in the
first iteration of the change. Enable them after we've
ensured there are no build breakages.
Change: 146309050

Change: 146309688

Also makes resource variables dense-tensor-like.
Change: 146310220

The better fix is to add Windows support to ParseURI, but that requires more significant changes, so leaving it as a TODO for now.
Change: 146314223

…'s currently exposed in Python

Change: 146314254

Change: 146314715

Change: 146316019

Change: 146316196

Change: 146318441

Change: 146318735

test_log_output_filename. This lets us create test_log_output_filename
automatically if it is not passed in.
Change: 146320874

Change: 146325322

Change: 146325723

Change: 146327228

Fixes the problem of XLA being unable to find libdevice files if not executed
through bazel. For example,
http://stackoverflow.com/questions/41729019/notfounderror-running-tensorflow-xla-example-libdevice-compute-35-10-bc/41800414#41800414
Change: 146329331

commented out plus some new overrides.
Change: 146330232

Resolved conflicts in:
tensorflow/contrib/slim/python/slim/learning_test.py