| Commit message (Collapse) | Author | Age |
|
|
|
|
| |
tensorflow/core/ files and build targets.
Change: 113075110
|
|
|
|
|
| |
tensorflow/core/ files and build targets.
Change: 113075039
|
|
|
|
|
| |
tensorflow/core/ files and build targets.
Change: 113074952
|
|
|
|
|
| |
tensorflow/core/ files and build targets.
Change: 113074643
|
|
|
|
|
| |
tensorflow/core/ files and build targets.
Change: 113073106
|
|
|
|
|
| |
tensorflow/core/ files and build targets.
Change: 113073090
|
|
|
|
|
| |
tensorflow/core/ files and build targets.
Change: 113073087
|
|
|
|
|
|
|
| |
problems."
Doesn't work in OSS
Change: 113069612
|
|
|
|
| |
Change: 113064195
|
|
|
|
| |
Change: 113024053
|
|
|
|
|
| |
case for wide logistic models.
Change: 113020949
|
|
|
|
| |
Change: 113003537
|
|
|
|
| |
Change: 112999739
|
|
|
|
| |
Change: 112993949
|
|
|
|
| |
Change: 112993601
|
|
|
|
|
|
| |
docs from source code, both c++ and python. The generated docs overwrite the docs in the g3doc/api_docs folder.
Change: 112989505
|
|
|
|
| |
Change: 112985792
|
|
|
|
| |
Change: 112984578
|
|
|
|
|
|
| |
- Select linkopts based on a config_setting so we don't pass
invalid options to the crosstool on Mac.
Change: 112984497
|
|
|
|
| |
Change: 112980390
|
|
|
|
|
|
|
| |
- Graph* -> Graph* pass
- Creates a local executor and executes a copy of the constant "slice" of the
original graph, and replaces nodes in original graph with constant nodes.
Change: 112971745
|
|
|
|
| |
Change: 112970147
|
|
|
|
| |
Change: 112968173
|
|
|
|
|
|
| |
The intention is for EqualGraphDef to be the only way downstream tests compare
GraphDefs in golden file tests.
Change: 112967255
|
|
|
|
|
| |
ConstantValue is now constant_value.
Change: 112966774
|
|
|
|
| |
Change: 112965179
|
|
|
|
|
| |
Previously, if activation_fn was set to its default (None) and create_summaries was false, fully_connected would incorrectly give an error. I have fixed this as well as changing the default to tf.identity.
Change: 112963042
|
|
|
|
| |
Change: 112962727
|
|
|
|
|
|
|
|
|
|
| |
Also remove the deprecated TensorShape.as_dimension_list function, and remove
the to_proto argument to as_list. Both features were used only to make protos,
and that is better done directly with the as_proto method to avoid leaking
abstractions.
This brings us one step closer to hiding tf.tensor_util.
Change: 112960305
|
|
|
|
|
| |
The two functions already have the same behavior.
Change: 112959229
|
|
|
|
| |
Change: 112920860
|
|
|
|
| |
Change: 112917524
|
|
|
|
| |
Change: 112917269
|
|
|
|
| |
Change: 112915818
|
|
|
|
| |
Change: 112912683
|
|
|
|
|
|
| |
allocated once in the OpKernelContext::Params struct, then re-used every time a new OpKernelContext uses the Params. Thus in the executor, as long as there is more work to do the PerOpGpuDevice is not freed.
Change: 112909215
|
|
|
|
|
| |
no outputs. Fixes #856
Change: 112903800
|
|
|
|
| |
Change: 112903292
|
|
|
|
|
|
|
|
| |
We still haven't advanced to the scalar strict GraphDef version, but this
change will prevent some (not all) new Python scripts from violating scalar
strictness. That is, it will significantly reduce the amount of bit rot
between now and when I finally get full scalar strictness submitted.
Change: 112866199
|
|
|
|
| |
Change: 112846663
|
|
|
|
| |
Change: 112846648
|
|
|
|
|
|
| |
one from the set when we find one we want is cheaper. Slight performance
improvement (~0.3% on ptb_word_lm model on my desktop).
Change: 112832451
|
|
|
|
| |
Change: 112830289
|
|
|
|
|
|
| |
step. Fixes #837
Change: 112829709
|
|
|
|
|
|
| |
new TensorFlow backend.
Change: 112826468
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* The "core_cpu_internal" build target no longer includes files from the
common_runtime/gpu/ directory.
* tensorflow/core internal targets instead can get access to those headers via
the "gpu_runtime" target.
* The class "CopyTensor" is introduced. It lives in common_runtime/
but supports registration of copy functions so the "gpu_runtime"
target can add a GPU->GPU copy ability if it is linked in.
This registration should make it easier to add more device types
in the future.
* The "core_cpu" and "core_cpu_internal" build targets no longer
reference GPUUtil::CopyViaDMA; rendezvous_mgr uses CopyTensor
instead.
Also the "copy_tensor" build target was not needed.
Change: 112821119
|
|
|
|
|
|
| |
* Move some checks out of inner loops
* Split the mapper in 2: a base mapper, and a sub-mapper. This reduces the number of variables that are contained in the base mapper and helps reduce register spills
Change: 112809881
|
|
|
|
|
|
| |
size of a CNN.
Change: 112809773
|
|
|
|
| |
Change: 112808739
|
|
|
|
|
|
|
|
|
|
| |
If multiple steps are blocked on the same queue, there is a high
chance that their cancellation tokens will collide, since these are
dense and start at 0. This can lead to the wrong step being cancelled.
This change additionally uses the `CancellationManager*` to uniquely
identify an attempt for the purposes of cancellation.
Change: 112801810
|