Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Branch 192771889 #18497

Merged
merged 110 commits into from Apr 13, 2018
Merged

Branch 192771889 #18497

merged 110 commits into from Apr 13, 2018

Conversation

qlzh727
Copy link
Member

@qlzh727 qlzh727 commented Apr 13, 2018

Third attempt to push to public.

Manually merged

tensorflow/contrib/data/python/ops/BUILD
tensorflow/python/framework/dtypes_test.py

tensorflower-gardener and others added 30 commits April 11, 2018 09:46
@tensorflower-gardener
Import FunctionDef as GrapplerFunctionItem
a9a3b98 
Explicitly track function input arg expansion into Placeholders,
and keep metadata to map between FunctionDef and GraphDef connectivity
formats.

PiperOrigin-RevId: 192462592
@MarkDaoust @tensorflower-gardener
Add a clear error message for when a doc does not have a title.
08a12ca 
PiperOrigin-RevId: 192463583
@jpienaar @tensorflower-gardener
Add gradient in cond test to match CallGradInLoop.
8f75385 
PiperOrigin-RevId: 192463997
@tensorflower-gardener
Use tf.train.get_or_create_global_step() instead of deprecated variab…
5757d09 
…les.get_or_create_global_step().

PiperOrigin-RevId: 192476077
@tensorflower-gardener
Fix uninitialized value.
48b2bdc 
PiperOrigin-RevId: 192479630
@tensorflower-gardener
Disable prelu tests for real now.
2ea5c1e 
PiperOrigin-RevId: 192480452
@tensorflower-gardener
Script to create custom_ops inside a TensorFlow graphdef.
8b17a17 
PiperOrigin-RevId: 192481690
@adria-p @tensorflower-gardener
Update docs for softmax_cross_entropy_with_logits.
abc26c1 
PiperOrigin-RevId: 192482242
@mrry @tensorflower-gardener
Increase size of tensorflow/contrib/data/python/kernel_tests:batch_da…
5eccb5a 
…taset_op_test to "medium".

PiperOrigin-RevId: 192484895
@angerson @tensorflower-gardener
Internal TF Lite test changes
cc15251 
PiperOrigin-RevId: 192491201
@rohan100jain @tensorflower-gardener
Adding hp5y back.
d983832 
PiperOrigin-RevId: 192491335
@annarev @tensorflower-gardener
Internal change.
c5d59c6 
PiperOrigin-RevId: 192498471
@tensorflower-gardener
DepthwiseConv Optimizations
371d513 
PiperOrigin-RevId: 192499401
@fchollet @tensorflower-gardener
When not necessary, avoid the creation of a `placeholder_with_default…
744a5cc 
…` in BN (not yet supported by TPU compilation).

PiperOrigin-RevId: 192502020
@tensorflower-gardener
Factor out the syntactic function scope tracking into the transformer…
3fa224a 
…. Choosing not to do this at static analysis because it exposes the scope to any node, making it easier to use by any specialization of a transformer.

PiperOrigin-RevId: 192502309
@tensorflower-gardener
Replace examples/image_retraining by a pointer to TensorFlow Hub.
1a36eb1 
https://github.com/tensorflow/hub/tree/master/examples/image_retraining
has the same tool, upgraded to use TensorFlow Hub instead of raw graph defs.

PiperOrigin-RevId: 192502469
@tensorflower-gardener
Support for removing unfused quantized activation functions and min/max.
73aef57 
PiperOrigin-RevId: 192503204
@jsimsa @tensorflower-gardener
Porting tests for the decode_proto and encode_proto to OS.
1e283d6 
PiperOrigin-RevId: 192504411
@tensorflower-gardener
BREAKING_CHANGE: Remove event_ndims in Bijector, and require `log_det…
eed6828 
…_jacobian` methods to take event_ndims.

The class level event_ndims parameter is being deprecated in favor of passing it in
to the `log_det_jacobian` methods.

Specific changes:

  - `log_det_jacobian` signatures are now `log_det_jacobian(input, event_ndims)`

  - Constructors no long have event_ndims passed in (e.g. Affine() vs. Affine(event_ndims=0)).

  - All bijectors must specify a subset of [forward_min_event_ndims, inverse_min_event_ndims]. This is the minimal dimensionality the bijector operates on, with it being "broadcasted" to any passed in event_ndims (e.g. Exp has forward_min_event_ndims = 0. That means it operates on scalars. However, we can use the bijector on any event_ndims > 0 (i.e. we've broadcasted the transformation to work on any amount of event_ndims > 0), and jacobian reduction will work in those cases.

As a result of this change, all bijectors should "broadcast" (e.g. Sigmoid now works on any number of event_ndims).

Other changes (internal and documentation):
  - Added clarifications on Jacobian Determinant vs. Jacobian Matrix.
  - Added clarifications on min_event_ndims, and what the jacobian reduction is over.
  - Changed caching of ildj to be keyed on event_ndims.
  - Several bug fixes to bugs unearthed while writing this code (e.g. transformed distribution shape computation being incorrect)

PiperOrigin-RevId: 192504919
@tensorflower-gardener
Adds a nodedef_fn parameter to copy_op_handler, allowing customizatio…
e520167 
…n by mutating

NodeDef before creating the copied operation.

PiperOrigin-RevId: 192505209
@tensorflower-gardener
Adding support for batch_to_space_nd op with crops.
21fb4ee 
PiperOrigin-RevId: 192511036
@tensorflower-gardener
Extend support to remove transpose/reverse on dimensions of size 1.
d2690cf 
PiperOrigin-RevId: 192516190
@tensorflower-gardener
GCS Filesystem should not cache checkpoint file as we need to read th…
079d63d 
…e updated checkpoints from the contents.

PiperOrigin-RevId: 192517819
@alextp @tensorflower-gardener
Fixes issue where name scope collisions could lead to an invalid vari…
4b08b66 
…able in the metagraph.

PiperOrigin-RevId: 192518307
@mrry @tensorflower-gardener
Increase size of //tensorflow/python/kernel_tests:sets_test to "medium".
f029631 
PiperOrigin-RevId: 192519639
@tensorflower-gardener
Revealing the range of node ids in the latest layer via resource' state
9ce7791 
PiperOrigin-RevId: 192520351
@tensorflower-gardener
Adding support of core feature columns and losses to gradient boosted…
acad702 
… trees estimators.

PiperOrigin-RevId: 192521398
@tensorflower-gardener
support profiling multiple tpu through one grpc and one session.
d6e2513 
data are saved with host prefix.

PiperOrigin-RevId: 192523668
@tensorflower-gardener
Speed up computing mean confidence intervals by avoiding tf.while_loop.
e7cfede 
Implement a vectorized way to compute the same thing instead.

PiperOrigin-RevId: 192524667
@mrry @tensorflower-gardener
Move callback into bound function to avoid copying.
2b94b44 
PiperOrigin-RevId: 192530231
fchollet and others added 19 commits April 12, 2018 19:04
@fchollet @tensorflower-gardener
Enable efficient feeding of symbolic tensors to placeholders in the K…
5a6d5a1 
…eras backend.

PiperOrigin-RevId: 192707345
@jsimsa @tensorflower-gardener
Automated g4 rollback of changelist 192691078
4f615ad 
PiperOrigin-RevId: 192708480
@tensorflower-gardener
Add boolean type to tflite in favor of comparison implementations.
3c98705 
PiperOrigin-RevId: 192711203
@jsimsa @tensorflower-gardener
Automated g4 rollback of changelist 192504411
3438c3f 
PiperOrigin-RevId: 192711501
@jsimsa @tensorflower-gardener
Automated g4 rollback of changelist 192698931
1c88fac 
PiperOrigin-RevId: 192718697
@tensorflower-gardener
[XLA] Rename Interpreter{Executor,Platform} -> XlaInterpreter{Executo…
68f0f1a 
…r,Platform}.

These types live inside StreamExecutor's namespace, but they are
specific to XLA.  Therefore they either shouldn't live in SE's namespace
or should have "XLA" in the name.

Moving them out of SE's namespace is ugly, because almost every type
used inside of these headers then needs to be qualified.  So name-change
it is.

This patch was generated by a mechanical find/replace.

PiperOrigin-RevId: 192724238
@tensorflower-gardener
-- Add a new histogram/cdf computation method compatible with the TPU.
73cc1d5 
-- Refactor utility functions into pruning_utils.py and add tests

PiperOrigin-RevId: 192727737
@tensorflower-gardener
Implementation of Less
1b0c277 
PiperOrigin-RevId: 192728635
@tensorflower-gardener
Fix bug in converted_call, and add tests for it.
bb80410 
PiperOrigin-RevId: 192751211
@tensorflower-gardener
Update for upstream LLVM *.def -> *.inc rename
b520022 
PiperOrigin-RevId: 192752798
@tensorflower-gardener
- Fixed small bug in example script
345414c 
PiperOrigin-RevId: 192756152
@tensorflower-gardener
Keep function doc string at the top of the function.
bb8fcd5 
PiperOrigin-RevId: 192761604
@tensorflower-gardener
Demo: RNN colorbot with Estimators.
8c47ec3 
PiperOrigin-RevId: 192765203
@tensorflower-gardener
Split byte_order.h off cpu_info.h
3eb4e4f 
PiperOrigin-RevId: 192768744
@tensorflower-gardener
Automated g4 rollback of changelist 192768744
f9de043 
PiperOrigin-RevId: 192770717
@tensorflower-gardener
Add support to TFLite for dilated convolution.
91c3199 
PiperOrigin-RevId: 192770919
@tensorflower-gardener
Refactor to remove the duplicate calls to obtain a function's namespa…
17aa70e 
…ce. This removes the need to explicitly import internal components (barring the tf module which cannot be imported directly).

PiperOrigin-RevId: 192771440
@tensorflower-gardener
Experiment with pre-shuffled fully-connected weights
554c587 
PiperOrigin-RevId: 192771889
@qlzh727
Merge commit for internal changes
a194572
@qlzh727 qlzh727 added the kokoro:force-run Tests on submitted change label Apr 13, 2018
@kokoro-team kokoro-team removed the kokoro:force-run Tests on submitted change label Apr 13, 2018
@qlzh727 qlzh727 added the kokoro:force-run Tests on submitted change label Apr 13, 2018
@kokoro-team kokoro-team removed the kokoro:force-run Tests on submitted change label Apr 13, 2018
@qlzh727
Copy link
Member Author

qlzh727 commented Apr 13, 2018

XLA test is broken by internal env change, which has nothing to do with this PR. I think this is good to go now.

@qlzh727 qlzh727 merged commit cda36b8 into tensorflow:master Apr 13, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@yifeif yifeif yifeif approved these changes

@jhseu jhseu Awaiting requested review from jhseu

@annarev annarev Awaiting requested review from annarev

Assignees
No one assigned
Labels
cla: yes
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

None yet