| c2c9e3beef18c5f1ed443d372dec18c8acc5d991 |
|
23-Jan-2018 |
Skye Wanderman-Milne <skyewm@google.com> |
_DefinedFunction.name should be a string.
With the C API enabled and python2, it was being set to a unicode
object.
PiperOrigin-RevId: 182865087
/external/tensorflow/tensorflow/python/framework/function.py
|
| 9b03bcd74d117a2a6ee270af438a3a28e7123111 |
|
19-Jan-2018 |
Alexandre Passos <apassos@google.com> |
Fixes issue with using lookup ops in datasets from eager.
Addresses issue #16160
PiperOrigin-RevId: 182554969
/external/tensorflow/tensorflow/python/framework/function.py
|
| 90e42f3ac8c43474633136af4242dca04b6a1e09 |
|
16-Dec-2017 |
Dandelion Man? <dandelion@google.com> |
Automated g4 rollback of changelist 179260538
PiperOrigin-RevId: 179263865
/external/tensorflow/tensorflow/python/framework/function.py
|
| 9648f8040a559f6cf9bbe0501ba96f2b2c2864b1 |
|
16-Dec-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Automated g4 rollback of changelist 179258973
PiperOrigin-RevId: 179260538
/external/tensorflow/tensorflow/python/framework/function.py
|
| d55f532867a3670d66460c5ee3b774519542adc1 |
|
16-Dec-2017 |
Dandelion Man? <dandelion@google.com> |
Merge changes from github.
PiperOrigin-RevId: 179258973
/external/tensorflow/tensorflow/python/framework/function.py
|
| 1c4810141e71289d71bfd94a74434bd09ee6b20f |
|
30-Nov-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Hoist function input placeholders out of any control flow context.
Prior to this change, functions that closed over external tensors in a
while loop would cause a segfault at runtime. This is because the
external tensors are temporarily represented as placeholders in the
function body before being replaced by input parameters, and the
placeholders would be created directly in the while loop body. This
would eventually lead to using the input tensor in the while loop body
without an enter node. This wasn't caught by the runtime check because
it isn't applied to function bodies.
This change adds tests for capturing tensors in a while loop body and
in a cond context. Note that the cond test passed without this fix.
PiperOrigin-RevId: 177464541
/external/tensorflow/tensorflow/python/framework/function.py
|
| 4640cc88ffcdaeb8724ffc129bc4d900277a5403 |
|
10-Nov-2017 |
Yifei Feng <yifeif@google.com> |
Fix typo in tensorflow/python/framework/function.py
PiperOrigin-RevId: 175219920
/external/tensorflow/tensorflow/python/framework/function.py
|
| 3bafe0a86f67dd54197c6d60bdb5053f510de7d8 |
|
10-Oct-2017 |
Peter Hawkins <phawkins@google.com> |
Add uint32 and uint64 types to TensorFlow.
This change merely creates the types, but does not register kernels that act on uint32/uint64 values. It also does not alter most op registration lists to include uint32/uint64 values. If desirable, that can be done in a subsequent change, although binary size will likely prove problematic if adding more kernels.
The intent of the change is so XLA-compiled code can make use uint32/uint64 types. Since XLA does not use traditional TensorFlow kernels, using uint32/uint64 operators from XLA will require only uint32/uint64 op registrations, but will require few new kernel registrations.
PiperOrigin-RevId: 171681867
/external/tensorflow/tensorflow/python/framework/function.py
|
| 2198b8cfe8acb5af7bb5a1dac54c18ff72c98002 |
|
05-Oct-2017 |
Igor Ganichev <iga@google.com> |
Minimize python code in supporting TF_Function.
After this change when C API is enabled, function support in Python is
done with minimal use of Python code. In particular, we don't create
or store FunctionDef in Python. Small changes include:
- We don't use _hash_str for function comparisons in Python. Instead,
we delegate this logic to TF_GraphCopyFunction in C API.
- We checking for duplication function additions from
_DefinedFunction.add_to_graph(graph) to Graph._add_function in all
cases. This is more logical and make it easier to support both modes.
- We change some error messages to be same in both modes.
- Since we don't store FunctionDef in C API mode in Python but get it
on demand, access to common attributes like name or signature can
become expensive. To mitigate this, we cache the signature (OpDef)
of the function in Python. Signatures are generally much smaller
than whole definitions.
- Add context manager for creating and destroying TF_Buffers.
- Allow zero output tensorflow functions in Python
The C API and C++ runtime support functions without outputs, but Python
APIs explicitly disallowed them before this change. This change allows
zero output functions in Python and cleans some hacks that were added
to side-step regular Python function APIs before.
PiperOrigin-RevId: 171201162
/external/tensorflow/tensorflow/python/framework/function.py
|
| bdab2691068757ee4872167898bc8768a7303ae9 |
|
28-Sep-2017 |
Igor Ganichev <iga@google.com> |
Add append_hash_to_fn_name arg to TF_GraphToFunction
PiperOrigin-RevId: 170379490
/external/tensorflow/tensorflow/python/framework/function.py
|
| 818dc210881aadcce9013b693e071b96a6e1f1f6 |
|
26-Sep-2017 |
Igor Ganichev <iga@google.com> |
Remove function declaration feature
PiperOrigin-RevId: 170043510
/external/tensorflow/tensorflow/python/framework/function.py
|
| 3e4521bd290e4654a8b1e432d16ca893181ab018 |
|
20-Sep-2017 |
Igor Ganichev <iga@google.com> |
Add function description parameter to TF_GraphToFunction
PiperOrigin-RevId: 169421145
/external/tensorflow/tensorflow/python/framework/function.py
|
| 7ad8e25495a2793ea14189359af736d2c662a694 |
|
20-Sep-2017 |
Igor Ganichev <iga@google.com> |
Add attribute setting and getting support to TF_Function
PiperOrigin-RevId: 169337159
/external/tensorflow/tensorflow/python/framework/function.py
|
| 1da763a1cc94fc5e4ad1822788b444b77623538c |
|
19-Sep-2017 |
Igor Ganichev <iga@google.com> |
Add function gradient support to C API
Also, change the internal representation of TF_Function and
rename TF_GraphAddFunction to TF_GraphAddFunctionCopy to make it
clear that a copy of the function is added to the graph. Any
subsequent modifications to the function will not be reflected
in the copy added to the graph.
PiperOrigin-RevId: 169187793
/external/tensorflow/tensorflow/python/framework/function.py
|
| 87f5182a5644993e747c2f42dfe6da75b7431e66 |
|
16-Sep-2017 |
Asim Shankar <ashankar@google.com> |
Fix broken test: tensorflow/contrib/eager/python:datasets_test
PiperOrigin-RevId: 168914742
/external/tensorflow/tensorflow/python/framework/function.py
|
| b57bbdecb24afa18d1716403dd05a08566e3e516 |
|
16-Sep-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Adding support for capture-by-value.
PiperOrigin-RevId: 168903482
/external/tensorflow/tensorflow/python/framework/function.py
|
| ea4e0ad9dbfb49c17be1cbe6112cbb68ea1b77a0 |
|
14-Sep-2017 |
Asim Shankar <ashankar@google.com> |
eager: Support function-valued operation attributes.
This enables use of MapDataset, FilterDataset etc. when
eager execution is enabled.
PiperOrigin-RevId: 168729488
/external/tensorflow/tensorflow/python/framework/function.py
|
| 9624d165f1f2c717eda96464fee8bf7229cc14f5 |
|
31-Aug-2017 |
Igor Ganichev <iga@google.com> |
Add function support to Tensorflow C API
This change adds minimal functionality. Support for FunctionOptions,
attributes, output name rewriting, function name generation, etc is
comming next.
PiperOrigin-RevId: 167091238
/external/tensorflow/tensorflow/python/framework/function.py
|
| 48c48729a1770eb1e59e86345b8768b466722041 |
|
16-Aug-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Allow Defun to inherit enclosing XLA compilation scope.
PiperOrigin-RevId: 165457067
/external/tensorflow/tensorflow/python/framework/function.py
|
| 28ce1d163eeffe618a6972c5245be0e660d94e85 |
|
15-Aug-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Merge changes from github.
END_PUBLIC
---
Commit 9f81374c3 authored by raymondxyang<zihao.yang@microsoft.com>
Committed by Rasmus Munk Larsen<rmlarsen@google.com>:
Add option for build more python tests in Cmake (#11853)
* Ignore Windows built project
* Fix deprecated methods in tf.contrib.python
* Fix regex match for Windows build in contrib.keras
* Fix Regex match for Windows build in session_bundle
* * Fix deprecated methods
* Fix regex match for Windows
* Fix compatibility issue with Python 3.x
* Add missing ops into Windows build for test
* Enabled more testcases for Windows build
* Clean code and fix typo
* Add conditional cmake mode for enabling more unit testcase
* Add Cmake mode for major Contrib packages
* Add supplementary info in RAEDME for new cmake option
* * Update tf_tests after testing with TF 1.3
* Clean code and resolve conflicts
* Fix unsafe regex matches and format code
* Update exclude list after testing with latest master branch
* Fix missing module
---
Commit 98f0e1efe authored by Yong Tang<yong.tang.github@outlook.com>
Committed by Rasmus Munk Larsen<rmlarsen@google.com>:
Dynamic ksize and strides with MaxPool (#11875)
* Dynamic ksize with max_pool
This fix tries to fix the issue raised in 4746 where ksize
is static (attr) with max_pool.
This fix changes ksize to input tensor so that it is dynamic now.
This fix fixes 4746.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Add dynamic ksize to MaxPoolGrad and MaxPoolGradGrad
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Add test cases for max_pool_v2
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Fix GPU Jenkins issue.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Enable MaxPoolV2 in GPU
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Hide MaxPoolV2 and other fixes.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
---
Commit 02d6bc185 authored by Bairen Yi<byronyi@users.noreply.github.com>
Committed by Rasmus Munk Larsen<rmlarsen@google.com>:
remove useless variable (#12212)
---
Commit ed6b0d905 authored by namrata-ibm<bhavenamrata@gmail.com>
Committed by Rasmus Munk Larsen<rmlarsen@google.com>:
Adding support for s390x in calculation of cpu_frequency (#12201)
---
Commit 627dfc9dd authored by Taehoon Lee<taehoonlee@snu.ac.kr>
Committed by Taehoon Lee<taehoonlee@snu.ac.kr>:
Fix typos
---
Commit c0f9b0a91 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
In fast-math mode emit a tanh that has a faster min/max.
PiperOrigin-RevId: 164943597
---
Commit 87605f3d6 authored by Kay Zhu<kayzhu@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[TF:XLA] Use HloEvaluator for ComputeConstant, remove the need of a dedicated
compute constant backend.
PiperOrigin-RevId: 164940970
---
Commit 881de45c2 authored by Taehoon Lee<me@taehoonlee.com>
Committed by Rasmus Munk Larsen<rmlarsen@google.com>:
Add bool type supports for GPU kernels (#11927)
* Add bool type supports for GPU kernels
* Add bool type test codes for GPU kernels
---
Commit eeacdcdb1 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add missing "CPU" suffix in registrations.
PiperOrigin-RevId: 164939527
---
Commit de01be952 authored by namrata-ibm<bhavenamrata@gmail.com>
Committed by Rasmus Munk Larsen<rmlarsen@google.com>:
Adding support for Big Endian in graph_constructor_test and wav_io (#12179)
---
Commit 26719d29f authored by QingYing Chen<pkudysj@126.com>
Committed by Rasmus Munk Larsen<rmlarsen@google.com>:
Implement CRF decode (Viterbi decode) for tensor (#12056)
* Implement CRF decoding for tensors
* add test code for tensor version's CRF decoding
* made modifications according to pylint
* add some comments for crf decode
* remove useless code
* add comments at the top comment of crf module and add more comments in crf_test
* capitalize first char of first word in comments
* replace crf_decode test code with a deterministic example
---
Commit f9a81ca2f authored by Pete Warden<pete@petewarden.com>
Committed by gunan<gunan@google.com>:
Create CI build script for Raspberry Pi (#12190)
* Create CI build script for Raspberry Pi
* Moved location of Pi build script
---
Commit e2a163a90 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Merge code from PR #11940 with internal changes from cl/164796436, and update Python tests to also run on GPU.
PiperOrigin-RevId: 164929133
---
Commit 08bbfa187 authored by Taehoon Lee<me@taehoonlee.com>
Committed by Rasmus Munk Larsen<rmlarsen@google.com>:
Fix typos (#12195)
---
Commit ab96f41fb authored by Luke Iwanski<luke@codeplay.com>
Committed by Rasmus Munk Larsen<rmlarsen@google.com>:
[OpenCL] Extends matmul_benchmark.py to cover SYCL (#11697)
* [OpenCL] Extends matmul_benchmark.py to cover SYCL
* Fixed typo
* /gpu:0 -> /device:GPU:0
* Fixes control_flow_ops_py_test
* /gpu: -> /device:GPU:
* Fixes //tensorflow/python/profiler/internal:run_metadata_test
* gpu: -> GPU:
* Fixes tfprof_node
* [OpenCL] Fixes device path to name with many colons (#123)
The device path is constructed from a device name by replacing all
colons with underscores. Some device names contain more than one colon,
for example 'device:SYCL:0' which gives a path 'device_SYCL_0'. The
previous code would not convert this back to the original device name,
but rather to 'device:SYCL_0'.
An alternative fix would be to convert all underscores to colons in the
device name (i.e. remove the restriction inside `replace("_", ":", 1)`),
however I'm not sure if there are any device names which contain
underscores.
* If no gpu device aviable fake one
* gpu: -> device:GPU
* Fixes profiler test
* /gpu:x -> /device:GPU:x
* Fixes debug_io_utils_test.cc test
* Fixes device_name_utils_test.cc
---
Commit 35e7a3665 authored by Yong Tang<yong.tang.github@outlook.com>
Committed by Rasmus Munk Larsen<rmlarsen@google.com>:
Remove unneeded casting of int64 for reverse_sequence (#12192)
This fix remove unneeded cast of int64 for reverse_sequence:
```
lengths = math_ops.to_int64(lengths)
```
as int32 has already been enabled for reverse_sequence.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
---
Commit 9fba8c185 authored by Anna R<annarev@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add benchmark dashboard link to benchmarks doc. Also, I added a link and
description for Benchmarks page to Community index page.
PiperOrigin-RevId: 164924906
---
Commit bb6f32fa7 authored by Mark Heffernan<meheff@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Make HloAliasAnalysis updatable after changes to the HLO graph.
As part of this change make HloAliasAnalysis a thinner layer which
basically only holds a map from HloValue to HloBuffer and vice versa.
PiperOrigin-RevId: 164923041
---
Commit 9103096c1 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by Thomas K?ppe<tkoeppe@google.com>:
Merged commit includes the following changes:
164923041 by meheff:
Make HloAliasAnalysis updatable after changes to the HLO graph.
As part of this change make HloAliasAnalysis a thinner layer which
basically only holds a map from HloValue to HloBuffer and vice versa.
--
PiperOrigin-RevId: 164923041
---
Commit 822603aed authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Merging sibling fusion instruction using multi_output_fusion
PiperOrigin-RevId: 164920220
---
Commit c035aa2a8 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Go: Update generated wrapper functions for TensorFlow ops.
PiperOrigin-RevId: 164917891
---
Commit e1e81d9ba authored by Luke Iwanski<luke@codeplay.com>
Committed by Rasmus Munk Larsen<rmlarsen@google.com>:
[OpenCL] Fixes double memcpy bug (#151) (#12173)
* [OpenCL] Fixes double memcpy bug (#151)
As the debg CopyOp is called on a Tensor without type, we need to use
the DataType enum to get type information, and use this to pass the type
on to Eigen. This is a workaround Eigen's need to have a type when
calling memcpy. If the Eigen memcpy can be provided without a type
requirement, then the memcpy in sycl_util is unnecessary.
* Acts on feedback from: #12173/files/32cb12a9001b672425867b5a3110fd98e737a20b#r132496277
---
Commit d9ca2d86d authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Internal change
PiperOrigin-RevId: 164916465
---
Commit b8d13d218 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Remove more parts of DCASGD missed in the first pass. (47949b)
PiperOrigin-RevId: 164914552
---
Commit 73b3d52c7 authored by Alexandre Passos<apassos@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
cmake fix
PiperOrigin-RevId: 164911656
---
Commit 2173b5b0a authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Allow TFE_TensorHandleCopyToDevice to have the same device as src and
destination. It will reuse the same underlying buffer in those cases.
PiperOrigin-RevId: 164909906
---
Commit 13eb3b90e authored by Alexandre Passos<apassos@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Experimental C and Python APIs to invoke TensorFlow kernels on concrete values.
PiperOrigin-RevId: 164902588
---
Commit 7dfabcc01 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Initialize ExecutionOptions in ComputeConstant to default values.
PiperOrigin-RevId: 164894867
---
Commit c8897e9bc authored by Benoit Steiner<bsteiner@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Static required time computation
PiperOrigin-RevId: 164894645
---
Commit 076158f9b authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Enable implicit->explicit conversion by default.
PiperOrigin-RevId: 164890915
---
Commit 58c4a4cb1 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Bugfix: number of input channels is not necessarily in the last dimension, after introduction of data_format param.
PiperOrigin-RevId: 164889729
---
Commit 8f9b1af8a authored by Igor Saprykin<isaprykin@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Recover MonitoredSession when the Coordinator is requested to stop with one of the _PREEMPTION_ERRORS.
When SyncReplicasOptimizer is used, a preemption in the Coordinator may result in two cases:
Case 1) the session gets silently marked as complete
Case 2) the session gets stuck
This CL aims to solve and verify solutions for both of these problems. Fix 1 changes the should_stop logic. Fix 2 changes the CoordinatedSession.run() logic.
SyncReplicasOptimizer runs a separate set of threads using a Coordinator instance. Those threads do FIFOQueue.enqueue; the main thread does a blocking FIFOQueue.dequeue.
`sync_token_q` FIFOQueue is on parameter-servers. When one of the PS instances gets preempted, an AbortedError causes the Coordinator to stop via request_stop(ex). That by itself changes the state of MonitoredSession.should_stop() to True (Fix 1).
Results of the blocking Dequeue operation are sent to the chief worker via Recv. What happens next depends on the amount of tokens in `sync_token_q`. If there are enough for the next call to Dequeue to return, then the low-level "tf session run() call" returns. The next iteration of the `while not MonitoredSession.should_stop()` loop decides that the training is complete (Case 1).
If there are not enough tokens in `sync_token_q`, then the blocking Dequeue is going to keep waiting for them. This results in the graph execution getting stuck and the whole session getting garbage collected after 10 minutes (Case 2).
We decided to fix that by re-creating a session after it gets garbage collected (Fix 2). An alternative was to try to cancel the pending Dequeue operation, but it's not clear that it is the right thing to do and it is also not easy.
PiperOrigin-RevId: 164888390
---
Commit 46e4de6e5 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Undo loop fusion changes for now as they seem to be altering a few results.
END_PUBLIC
RELNOTES: n/a
BEGIN_PUBLIC
BEGIN_PUBLIC
Automated g4 rollback of changelist 164825735
PiperOrigin-RevId: 165340331
/external/tensorflow/tensorflow/python/framework/function.py
|
| 0815de21239955e346b562e899640649c8d2b9cb |
|
05-Aug-2017 |
Benoit Steiner <bsteiner@google.com> |
Merge changes from github.
END_PUBLIC
---
Commit cf375f067 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by Benoit Steiner<benoitsteiner@users.noreply.github.com>:
Adds cudnn_rnn_ops_op_lib and cudnn_rnn_kernels to contrib_ops_op_lib and
contrib_kernels respectively.
PiperOrigin-RevId: 164170971
---
Commit 95ec58e27 authored by Asim Shankar<ashankar@google.com>
Committed by Benoit Steiner<benoitsteiner@users.noreply.github.com>:
C API: Make TF_TensorFromTensor return an error instead of just logging it.
PiperOrigin-RevId: 164167582
---
Commit 15175c870 authored by Jonathan Hseu<jhseu@google.com>
Committed by Benoit Steiner<benoitsteiner@users.noreply.github.com>:
Build fixes.
- Allow var_list as a positional argument in CrossShardOptimizer.
- Set the number of shards to 1 when not running on TPU, to allow evaluate() and predict() on CPU/GPU to work.
PiperOrigin-RevId: 164161640
---
Commit bd3e894f7 authored by Yao Zhang<yaozhang@google.com>
Committed by Benoit Steiner<benoitsteiner@users.noreply.github.com>:
Support freeze mode for fused batch norm.
PiperOrigin-RevId: 164149032
---
Commit e6b6b84c0 authored by Asim Shankar<ashankar@google.com>
Committed by Benoit Steiner<benoitsteiner@users.noreply.github.com>:
C API: TF_Tensors will always be in host memory.
This change undoes some experimentation in
commit 22651083406ca01ac9d481e3367a3510d25f88cd
and restores TF_Tensor behavior to what is was prior to that change.
PiperOrigin-RevId: 164146670
---
Commit 8bf3f88f7 authored by Peter Hawkins<phawkins@google.com>
Committed by Benoit Steiner<benoitsteiner@users.noreply.github.com>:
[TF:XLA] Add _XLASend and _XLARecv TF ops that wrap the XLA Send/Recv HLO ops.
PiperOrigin-RevId: 164124764
---
Commit 626d3200f authored by Peter Hawkins<phawkins@google.com>
Committed by Benoit Steiner<benoitsteiner@users.noreply.github.com>:
[XLA] Add test blacklist mechanism for XLA C++ unit tests.
PiperOrigin-RevId: 164124423
---
Commit 359cc5f5e authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by Benoit Steiner<benoitsteiner@users.noreply.github.com>:
Document dict ordering in nest and make it consistent with sonnet.
PiperOrigin-RevId: 164114335
---
Commit 05813b531 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by Benoit Steiner<benoitsteiner@users.noreply.github.com>:
Go: Update generated wrapper functions for TensorFlow ops.
PiperOrigin-RevId: 164089206
---
Commit c451f465d authored by Anna R<annarev@google.com>
Committed by Benoit Steiner<benoitsteiner@users.noreply.github.com>:
BEGIN_PUBLIC
Automated g4 rollback of changelist 164078808
PiperOrigin-RevId: 164318935
/external/tensorflow/tensorflow/python/framework/function.py
|
| b876065afe85ef7b8e7334e506af2f84b3a6add1 |
|
31-Jul-2017 |
Alexandre Passos <apassos@google.com> |
graph_to_function_def gets its own file
PiperOrigin-RevId: 163709410
/external/tensorflow/tensorflow/python/framework/function.py
|
| 8e82134acd68c313f0ecadb168a66916006b17a6 |
|
12-Jul-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Use C API to implement Operation._input_types
This change first converts _input_types into a property and renames the
member to _input_types_val. We keep _input_dtypes as an alias for
_input_types as it was before this change.
Similarly to _output_types, we can't enable normal tests yet. Instead,
we add a simple temporary test for _input_types.
Also, fix two minor typos in doc strings of function.py
PiperOrigin-RevId: 161597185
/external/tensorflow/tensorflow/python/framework/function.py
|
| 98eb5270e2f9b61408f04035c7edde66c21e3fa7 |
|
09-Jun-2017 |
Ali Yahya <alive@google.com> |
Reorder code in function.py for better readability.
PiperOrigin-RevId: 158525581
/external/tensorflow/tensorflow/python/framework/function.py
|
| c048e2938ceced754f95e5a15dff81e37646aaa3 |
|
01-Jun-2017 |
Alexandre Passos <apassos@google.com> |
Adds support to non-placeholder inputs in _graph_to_function_def.
Specifically, supports input ops with more than one output tensor.
PiperOrigin-RevId: 157640908
/external/tensorflow/tensorflow/python/framework/function.py
|
| eae7e833f104b0afc0aa7bba21a59c56612dc3df |
|
26-May-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Change shape inference so that a single resource tensor can carry
shape and type information for multiple tensors.
Apply this to QueueDequeueV2 handled by grappler.
PiperOrigin-RevId: 157163757
/external/tensorflow/tensorflow/python/framework/function.py
|
| 56b2b7ab193b6fedb3dee58e847bad9776f69a16 |
|
05-May-2017 |
Alexandre Passos <apassos@google.com> |
Internal change.
Change: 155236037
/external/tensorflow/tensorflow/python/framework/function.py
|
| 8e5041918f2e709ded94e63fb1779d6bb363becb |
|
21-Apr-2017 |
Charles Nicholson <nicholsonc@google.com> |
Introduce TFDecorator, a base class for Python TensorFlow decorators. Provides basic introspection and "unwrap" services, allowing tooling code to fully 'understand' the wrapped object.
Change: 153854044
/external/tensorflow/tensorflow/python/framework/function.py
|
| e4a8dc831dbf2894c79659d50aea73999c1ff173 |
|
31-Mar-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Allow variable reuse in function.
Change: 151807504
/external/tensorflow/tensorflow/python/framework/function.py
|
| 4d35ecb853f5e9c7d41df3629a31dcbc7b4032ca |
|
30-Mar-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Copy handle dtype and shape to the placeholders created inside Defun context
for captured external tensors.
Also fix a minor bug in ResourceVariable that dtype from deserialization contains integer enum value rather than dtypes.DType instance.
Change: 151723784
/external/tensorflow/tensorflow/python/framework/function.py
|
| a75b6df69e1b6965fcbac5df68e89dc3cbe9931e |
|
13-Mar-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
[TF:XLA] Add separate_compiled_gradients to control gradient scopes.
Change: 149973410
/external/tensorflow/tensorflow/python/framework/function.py
|
| eb281da2d4379cce5bcb16c6910a085ade84b24b |
|
11-Mar-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Fix unsuccessful opdef lookup when function was not created via Defun.
Change: 149826365
/external/tensorflow/tensorflow/python/framework/function.py
|
| 020501c695119f76bba0dd2bb47abfeaa939d669 |
|
07-Mar-2017 |
Alexandre Passos <apassos@google.com> |
Hoist resource variable reads outside functions.
Change: 149360110
/external/tensorflow/tensorflow/python/framework/function.py
|
| 00a294d68e0f36c44fefcf2e07bf40068250a884 |
|
01-Mar-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Make import_graph_def work with functions
This patch adds logic for recreating functions originally defined using @Defun and serialized in a GraphDef.
Change: 148893654
/external/tensorflow/tensorflow/python/framework/function.py
|
| 93a975e114ee1c35f01ed3bdd47170e6f7129014 |
|
18-Feb-2017 |
Vijay Vasudevan <vrv@google.com> |
Merge changes from github.
Change: 147897309
/external/tensorflow/tensorflow/python/framework/function.py
|
| e62619519b788a73393ca64651b8a081ec9c63c1 |
|
14-Feb-2017 |
Jonathan Hseu <jhseu@google.com> |
Doc fixit task 2.
Change: 147404726
/external/tensorflow/tensorflow/python/framework/function.py
|
| 46c8e8bd966286ad3f192948fa64f2e9ee8f0dcb |
|
11-Feb-2017 |
Eugene Brevdo <ebrevdo@google.com> |
Add _XlaScope attribute to jit_scope to avoid fusing separate adjacent fused blocks.
Gradients get their own separate scope based on the scope of the forward op.
Provide proper XlaScope for Defuns as well (each Defun gets its own scope; their gradients get their own scope).
Also move jit scope gradient unit tests out of core gradients to contrib.compiler.
This is just the python side that sets the attribute; the C++ changes will come
in a separate CL.
Change: 147216860
/external/tensorflow/tensorflow/python/framework/function.py
|
| 8343176bb003f3591cde09ea47636fa5a6a9cc23 |
|
10-Feb-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Add shape_func argument to function.Defun, to specify the shape of output Tensors.
Change: 147149282
/external/tensorflow/tensorflow/python/framework/function.py
|
| 8567b761999bd051eaebdae5f2bd6d751eb4ec81 |
|
03-Feb-2017 |
Peter Hawkins <phawkins@google.com> |
Ensure that the function definition has been built when evaluating the captured variables of a function.
Change: 146511347
/external/tensorflow/tensorflow/python/framework/function.py
|
| 6237c439bdc3c3f882c1532ef85708956410e6f6 |
|
02-Feb-2017 |
Alexandre Passos <apassos@google.com> |
Function library tests pass with resource variables.
Also makes resource variables dense-tensor-like.
Change: 146310220
/external/tensorflow/tensorflow/python/framework/function.py
|
| 873473ef0fff3d5f106f87cbd83048222f1dd93a |
|
12-Jan-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Delete the FunctionDef.Node code paths, now that we have switched
to the NodeDef representation.
Change: 144281952
/external/tensorflow/tensorflow/python/framework/function.py
|
| 2eaaadae5a0afc0a92ed81cca550d57bb9b29cc1 |
|
12-Dec-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Hashing function input/output args should be a part of the function
name hash when the name is not given explicitly.
Change: 141790846
/external/tensorflow/tensorflow/python/framework/function.py
|
| 2d00e6f17df644077af331e5bcb47a0e8a0fa1b7 |
|
09-Dec-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Re-enable the NodeDef version of functions, fixing issue where Python
functions create signatures with duplicate output argument names.
Change: 141500446
/external/tensorflow/tensorflow/python/framework/function.py
|
| 165a1442c8eebb32d9edb11a74a612fef7536a78 |
|
07-Dec-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Add tf2xla tests.
Change: 141346498
/external/tensorflow/tensorflow/python/framework/function.py
|
| afb3e3e1b26e33b52fc2e7b6516b0f6e3f3efdd7 |
|
01-Dec-2016 |
Asim Shankar <ashankar@google.com> |
Switch FIXME to TODO
Change: 140733125
/external/tensorflow/tensorflow/python/framework/function.py
|
| eeb2632cd6bb77a88fdb6d2d20bfda23ad3381b8 |
|
23-Nov-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Switch function definition hash to NodeDef, fixing a TODO.
Change: 139963371
/external/tensorflow/tensorflow/python/framework/function.py
|
| 5591ca5f02377b27c3827b34c14b3f2f86451e06 |
|
22-Nov-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Python DefFun now creates functions that include NodeDefs
(in addition to the older FunctionDef::Node format). Add an
optional out_names argument to Defun and names to Declare's
arguments, so that signatures for forward-declared DefFuns
can have signatures in the name (required for this change).
Change: 139919259
/external/tensorflow/tensorflow/python/framework/function.py
|
| 85eeec0d415a1478bbeffc3d4545c795bee64e9f |
|
19-Nov-2016 |
Jonathan Hseu <jhseu@google.com> |
Automated rollback of change 139400135
Change: 139632235
/external/tensorflow/tensorflow/python/framework/function.py
|
| 7e8728662120df0a80720bb7527613f96d58271e |
|
17-Nov-2016 |
Jonathan Hseu <jhseu@google.com> |
Rename `Tensor` to `Output` in all Python docs
Generated by running:
$ find . -name '*.py' | xargs sed -i 's/a `Tensor`/an `Output`/g'
$ find . -name '*.py' | xargs sed -i 's/A `Tensor`/An `Output`/g'
$ find . -name '*.py' | xargs sed -i 's/`Tensor`/`Output`/g'
$ find . -name '*.py' | xargs sed -i 's/`tf.Tensor`/`tf.Output`/g'
$ find . -name '*.py' | xargs sed -i 's/`Tensors`/`Output`s/g'
$ find . -name '*.py' | xargs sed -i 's/#Tensor)/#Output)/g'
$ find . -name '*.py' | xargs sed -i 's/#Tensor\./#Output./g'
Manually fixed up lines that exceeded 80 characters after the change.
Change: 139400135
/external/tensorflow/tensorflow/python/framework/function.py
|
| 54afc49e09b4b24b166cbad8c115f794c149c06f |
|
16-Nov-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
More clearly document that functions are experimental.
Change: 139372609
/external/tensorflow/tensorflow/python/framework/function.py
|
| aac685b7209b03ffd356ea6860366467b335d402 |
|
08-Nov-2016 |
Dan Smilkov <smilkov@google.com> |
Internal change.
Change: 138451572
/external/tensorflow/tensorflow/python/framework/function.py
|
| b4cc63c5e59a922ad8040aa5da06d128733d5516 |
|
11-Oct-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
* Exposes some properties of a defined function.
* Passes along shape information of captured tensors/variables.
* Writes a hash function agnostic to function def's serialization order
because map is not guaranteed to serialized determinsticly across build.
* Passes along trainable bool and collections list to variable scopes
so that captured variables are placed into proper collections in the
outer graph's collections.
Change: 135752415
/external/tensorflow/tensorflow/python/framework/function.py
|
| d680b19889fa7cb6c25a3a32006199e542b3d411 |
|
06-Oct-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Hoist variables defined inside a function.
Change: 135405674
/external/tensorflow/tensorflow/python/framework/function.py
|
| 7a0cde1dd7b9656da3be8aebc7bfe9ec65e0b6b0 |
|
05-Oct-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
* Supports defining overloaded functions. I.e., defers the
instantiation of the function definition until all input types are known.
* for convenience, adds more convert_to_tensor calls.
* adds None checking for better error reporting.
* cleanup function_test a bit so that it uses tf.logging.info consistently.
* changes the convention for coming up a function's name:
* if func_name is explicitly given, stick with it.
* otherwise, come up a name based on the python function info and a hash of the
function definition.
* uses __doc__ to fill in func's op def description in the proto.
Change: 135156837
/external/tensorflow/tensorflow/python/framework/function.py
|
| c12b2121ba76eb9246f8d787094d96c9d1b54b11 |
|
04-Oct-2016 |
Martin Wicke <wicke@google.com> |
Automated rollback of change 135036941
Change: 135125564
/external/tensorflow/tensorflow/python/framework/function.py
|
| 74d0655d3f606a93efecd8f9520d6b2fe2d248aa |
|
04-Oct-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
* Supports defining overloaded functions. I.e., defers the
instantiation of the function definition until all input types are known.
* for convenience, adds more convert_to_tensor calls.
* adds None checking for better error reporting.
* cleanup function_test a bit so that it uses tf.logging.info consistently.
* changes the convention for coming up a function's name:
* if func_name is explicitly given, stick with it.
* otherwise, come up a name based on the python function info and a hash of the
function definition.
* uses __doc__ to fill in func's op def description in the proto.
Change: 135036941
/external/tensorflow/tensorflow/python/framework/function.py
|
| b95814208554ff6dac745c3f58a93929214a5363 |
|
27-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Supports capturing tensors from outside of a Defun.
Change: 134421429
/external/tensorflow/tensorflow/python/framework/function.py
|
| fa8a4a110d64c71339a33f4d2e67f76ef997afc0 |
|
26-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Allows calling a function w/ python literals and tf variables.
Change: 134288941
/external/tensorflow/tensorflow/python/framework/function.py
|
| f61c9d02f769bdf701d94cc09ecc94f21027f2ae |
|
26-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Handles an op within a function without return values.
Change: 134229005
/external/tensorflow/tensorflow/python/framework/function.py
|
| adc15a46c56070ef92b890b65b6de0147abeff80 |
|
22-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Improves the syntax for passing in Defun.__init__'s kwargs.
Change: 133997045
/external/tensorflow/tensorflow/python/framework/function.py
|
| 75b07290907abf759f05d8ea087f4623de316db0 |
|
22-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Defun.__init__ can take some extra kwargs which is passed
to every call site.
Change: 133899430
/external/tensorflow/tensorflow/python/framework/function.py
|
| e5b7e1e846f3e35d90a6bb260284b041d0036059 |
|
16-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Cleanup
Change: 133341191
/external/tensorflow/tensorflow/python/framework/function.py
|
| 8d54103171827b17cd062a71bc1b50fb80b3bb82 |
|
15-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Some cleanup.
Change: 133273363
/external/tensorflow/tensorflow/python/framework/function.py
|
| 74eeb2021564da80158e92744146a0c8bd7c6a70 |
|
15-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Supports forward declaration of functions.
Change: 133197452
/external/tensorflow/tensorflow/python/framework/function.py
|
| ddb917930c7f2519166c879e8f5273afb9c5c59a |
|
14-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Formating.
Change: 133163348
/external/tensorflow/tensorflow/python/framework/function.py
|
| 1d2aa9451d920ad4bc8ad1ac86c06863b590e81f |
|
09-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Removes a few limitation of Defun:
* Defun can appear anyway. Previously, it must be under a with tf.Graph() scope;
Now, it can be used to decorate module level functions.
* Defun constructs the function definition lazily so that if a Defun is not called,
it cost little;
* Defun can be used within another Defun. The inner ones are lifted to the top-level.
Change: 132620258
/external/tensorflow/tensorflow/python/framework/function.py
|
| 0de021876cbfc65c3d55505f3f6e8ffd3aa80239 |
|
26-Aug-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Remove some unnecessary work done during function
construction (and fix a bad comment).
Change: 131340944
/external/tensorflow/tensorflow/python/framework/function.py
|
| e058f2d1382aae39c88b6d3a71f25ec6114ffb31 |
|
23-Aug-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Begin transition to use NodeDef in FunctionDef instead of
FunctionDef.Node.
Change: 131009401
/external/tensorflow/tensorflow/python/framework/function.py
|
| 1df3fb0b4ae5915364f09e233496e98a99a4a886 |
|
09-Aug-2016 |
Lukasz Kaiser <lukaszkaiser@google.com> |
Deprecate op_scope by extending name_scope and move all callers to the new name_scope.
Instead of tf.op_scope(values, name, default_name) please use now
tf.name_scope(name, default_name, values). Only the name parameter is required,
and when other are absent, it behaves exactly as tf.name_scope did before.
Change: 129776424
/external/tensorflow/tensorflow/python/framework/function.py
|
| 01925fe23a5d8ebf17a1ddfcfaa1503f67e32575 |
|
24-Jul-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Adds python_grad_func argument to function.Defun.
This allows the gradient for a function to be specified the same way as tf.RegisterGradient (except at the graph scope, rather than global scope). This allows the gradient to be expanded directly in the graph, which in turn enables efficiencies (in memory and space) that do not depend on the graph optimizer (which requires configuration entirely separate from the model definition, and does not always kick in, e.g. when control flow ops are present).
Change: 128270792
/external/tensorflow/tensorflow/python/framework/function.py
|
| df94f63281a40acdd695a9a28862905af976adfd |
|
13-Jun-2016 |
Zongheng Yang <zongheng@google.com> |
Minor: fix typo in Defun doc.
Change: 124731079
/external/tensorflow/tensorflow/python/framework/function.py
|
| 0cf9ed3a719c0782695154d5a0bca260001cec15 |
|
02-Jun-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Update copyright for 3p/tf/python.
Change: 123900456
/external/tensorflow/tensorflow/python/framework/function.py
|
| 6bf83b7061df648d4751bc443782348fc8ea5c17 |
|
16-May-2016 |
Eugene Brevdo <ebrevdo@gmail.com> |
Ease-of-use updates to function and function.Defun.
- function.Defun now uses function refs for grad_func.
- function.Defun also now accepts a func_name string argument,
so users can have more control over the Function name.
Change: 122459066
/external/tensorflow/tensorflow/python/framework/function.py
|
| 0cc25f0b95af9a99dd228c75d411670831fd6cec |
|
24-Mar-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Allows the caller of a function to specify that a function is not
inlined into the call site regardless of the graph optimizer option
setting.
Change: 117994922
/external/tensorflow/tensorflow/python/framework/function.py
|
| f10637b372b3216400129b42abacd4fe50a6d7ea |
|
23-Mar-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Allows explicitly setting a gradient function for a user defined function.
Change: 117950844
/external/tensorflow/tensorflow/python/framework/function.py
|
| f6f77accbb9c978b5d3562922425bffd01690f79 |
|
11-Mar-2016 |
Derek Murray <mrry@google.com> |
Implements "bool strictness" for `tf.Tensor`.
Using a `tf.Tensor` as a Python boolean value is ambiguous: in library
code it typically means "is this tensor not None?", whereas in user
code it can mean "does this tensor evaluate to `True`" (which does
not work and leads to buggy code).
This change adds a warning when a tensor is used as a bool, and
replaces all known instances in TensorFlow with the appropriate `is
None` or `is not None` checks.
This is the first step towards addressing Issue #1454.
Change: 116984574
/external/tensorflow/tensorflow/python/framework/function.py
|
| 19b7fd80780d02372f76076bc8eb40d55a89a301 |
|
01-Mar-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Improves the syntax to define a tf function w/ a list of arguments.
Previously,
@Defun(x=tf.float32, y=tf.float32)
def Foo(x, y):
return x + y
Now,
@Defun(tf.float32, tf.float32)
def Foo(x, y):
return x + y
Supports varargs. E.g.,
@Defun([tf.float32]*10)
def Foo(*args):
sum = args[0]
for x in args[1:]:
sum = sum + x
return sum
Change: 116011595
/external/tensorflow/tensorflow/python/framework/function.py
|
| 59f1eba5fb94506a205fa2e81145667754739da5 |
|
07-Feb-2016 |
Martin Wicke <wicke@google.com> |
Merge changes from GitHub, clean up linter errors, fix dependencies test.
Change: 114064632
/external/tensorflow/tensorflow/python/framework/function.py
|
| 3f9101e95e3d1359cb81863657b51c5909f89baa |
|
25-Jan-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
More function related.
Change: 112912683
/external/tensorflow/tensorflow/python/framework/function.py
|
| 084075e4e967677fc0be9fe921a0157126e69617 |
|
20-Jan-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
More function support.
Change: 112611228
/external/tensorflow/tensorflow/python/framework/function.py
|
| a84a81a7379507f8fcdd0d6118afc2d5044d159e |
|
18-Jan-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
* Supports nested function calls;
* Supports uses python class method as a tf function prototype.
Derives the generate function name from the class name and the method name.
* Changes one LOG(INFO) to VLOG(1), which is too verbose.
Change: 112384148
/external/tensorflow/tensorflow/python/framework/function.py
|
| f9f26e04929ab20653b2b6ce86fcad71f906b193 |
|
14-Jan-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Fix Scatter*.png figures
Change: 112109184
/external/tensorflow/tensorflow/python/framework/function.py
|
| 0c62942e2bda15fa815fff440c9f5e466bff59b5 |
|
13-Jan-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Change visibility to public in bower.BUILD
A remote repository cannot refer to the main repository that way and a fix in a future Bazel release will break TensorFlow with the current set-up.
Error: http://ci.bazel.io/job/TensorFlow/BAZEL_VERSION=HEAD,PLATFORM_NAME=linux-x86_64/74/console
Change: 112056797
/external/tensorflow/tensorflow/python/framework/function.py
|
| 5515148be85f369484d6179b7c1baab30995b068 |
|
13-Jan-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Minor fixes (related) to gpu_event_mgr_test.cc
This test got out of sync after some recent changes.
This CL allows the test function to explicitly stop/start
the polling loop so we can test some invarients without
non-deterministic timing issues. Also, the change
to TensorReferenceVector is now handled correctly in ~EventMgr,
and reliably tested.
Change: 112016082
/external/tensorflow/tensorflow/python/framework/function.py
|
| e9434718c2c4f7e1de5488a411a2ea89e2724724 |
|
12-Jan-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Fix overly pessimistic shape inference in tf.batch_matmul for matrices with partially specified shapes.
Change: 111953111
/external/tensorflow/tensorflow/python/framework/function.py
|
| 3034eef57a9e6a080eae457a5ee8480fadd377ea |
|
12-Jan-2016 |
Martin Wicke <wicke@google.com> |
Fix error in link to mnist_with_summaries.py
Change: 111920304
/external/tensorflow/tensorflow/python/framework/function.py
|
| f2de334678c251aa767743d08e66345a45113cc4 |
|
12-Jan-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Fix type handling logic in translation of TensorFlow functions.
Inputs and outputs with lists of fixed types (as opposed to polymorphic lists) caused incorrect translation of some nodes in a function graph - adding a spurious attribute with name == "" and no value.
e.g. the 'indices' argument of DynamicStitch has a fixed int32 type, but 'data' is polymorphic. Handling 'indices' added a bogus attribute to the op in the function body.
REGISTER_OP("DynamicStitch")
.Input("indices: N * int32")
.Input("data: N * T")
.Output("merged: T")
.Attr("N : int >= 2")
.Attr("T : type")
Note that this is because the attr map in the ArgDef behaves like a C++ map in that an attempt to access a non-existent field creates key with a default value (i.e. does not fail like a Python dictionary would).
Change: 111895250
/external/tensorflow/tensorflow/python/framework/function.py
|
| 66c21d60d4becef8c72162015c66492ba975495a |
|
08-Jan-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Finish a TODO where we need to support N*T and type list during
graph-to-func conversion.
Change: 111648019
/external/tensorflow/tensorflow/python/framework/function.py
|
| d12b67833196d5508657d488cb8d96127419f2eb |
|
06-Jan-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Change: 111539073
/external/tensorflow/tensorflow/python/framework/function.py
|