Commits

Branch selector

 master

User selector

keveman

All time

Commit History

Commits on Feb 15, 2017

• Added documentation for imperative mode (#7534)

  keveman

  committedFeb 15, 2017
  969585b

Commits on Jun 2, 2016

• Fix data race in constant folding.

  

  keveman

  authored and

  tensorflower-gardener

  committedJun 2, 2016
  9538634

Commits on May 24, 2016

• Register GPU device OpKernel for Mod, for int32 type.

  

  keveman

  authored and

  tensorflower-gardener

  committedMay 24, 2016
  249106f

Commits on May 19, 2016

• Support negative values in the reduction_indices argument of reduce_*

  

  keveman

  authored and

  tensorflower-gardener

  committedMay 19, 2016
  cce41d3

Commits on May 18, 2016

• Enforce the rule that custom op libraries don't depend on core:framework or

  

  keveman

  authored and

  tensorflower-gardener

  committedMay 18, 2016
  cf9d71a

Commits on May 17, 2016

• Added framework/register_types.h to core:framework_headers

  

  keveman

  authored and

  mrry

  committedMay 17, 2016
  66cda38

Commits on May 16, 2016

• Added framework/register_types.h to core:framework_headers

  

  keveman

  authored and

  tensorflower-gardener

  committedMay 16, 2016
  78bc7ca

Commits on May 13, 2016

• Enforce the rule that custom op libraries don't depend on core:framework or

  

  keveman

  authored and

  tensorflower-gardener

  committedMay 13, 2016
  febd091

• Added linker script to limit the symbols that are visible in

  

  keveman

  authored and

  tensorflower-gardener

  committedMay 13, 2016
  5405394

Commits on Apr 25, 2016

• Do not replace a constant while constant folding.

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 25, 2016
  eb6fe62

Commits on Apr 16, 2016

• Do not add large constants (> 10M) to the graph while constant folding.

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 16, 2016
  73d026e

Commits on Apr 13, 2016

• Removed unnecessary link options when build shared libraries for custom ops.

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 13, 2016
  0afe5b3

Commits on Apr 8, 2016

• Remove unused mutex.h from file_system.h

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 8, 2016
  f77c9fb

• Move the FileSystemRegistry inside Env instead of it being global and shared

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 8, 2016
  6cdcc59

• Parse the URI from scheme and allow only [a-zA-Z][0-9a-zA-Z.]*

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 8, 2016
  74cba0c

Commits on Apr 7, 2016

• Implement a file factory mechanism to handle network file systems.

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 7, 2016
  fe5c6ae

• Implement a file factory mechanism to handle network file systems.

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 7, 2016
  9cc7cf6

• Implement a file factory mechanism to handle network file systems.

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 7, 2016
  9671b6c

Commits on Apr 5, 2016

• Making sure there is a kernel implementation for the constant on the device

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 5, 2016
  37baf7a

• Use convert_to_tensor to do type/shape check instead of make_tensor_proto,

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 5, 2016
  84476b2

Commits on Apr 4, 2016

• Avoid HostMemory constraint violation when doing constant folding.

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 4, 2016
  6398cb9

Commits on Apr 3, 2016

• Fixed the declaration of ReplaceTensorWithConstant.

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 3, 2016
  320b2e2

• Disable optimizations when running benchmarks.

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 3, 2016
  784cb15

Commits on Apr 2, 2016

• Moved a variable declaration inside GOOGLE_CUDA to fix an unused variable

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 2, 2016
  1bdd2c3

• Enable constant folding in L0 optimization level.

  

  keveman

  authored and

  tensorflower-gardener

  committedApr 2, 2016
  0cd025b

Commits on Mar 31, 2016

• Add type check for concat dim, add tests.

  

  keveman

  authored and

  tensorflower-gardener

  committedMar 31, 2016
  9f0f274

Commits on Mar 26, 2016

• Merge commit for internal changes

  keveman

  committedMar 26, 2016
  84184a3

• Added instructions for installing binary pip package with fast cpp protos

  

  keveman

  authored and

  tensorflower-gardener

  committedMar 26, 2016
  8ac0097

Commits on Mar 23, 2016

• Merge pull request #1608 from keveman/branch_117961829.

  keveman

  committedMar 23, 2016
  1a25c92

• Fixed merge conflicts in skflow

  keveman

  committedMar 23, 2016
  16454ee

• Added standalone kernel_library and gen_op_wrapper_py targets

  

  keveman

  authored and

  tensorflower-gardener

  committedMar 23, 2016
  b7babfd

Commits on Mar 22, 2016

• Merge pull request #1584 from keveman/master.

  keveman

  committedMar 22, 2016
  ebfebf5

• Added a note about compiling user op library with gcc 5

  keveman

  committedMar 22, 2016
  da16fd6

Commits on Mar 21, 2016

• Added -fPIC to the command line in instructions to compile user op.

  

  keveman

  authored and

  tensorflower-gardener

  committedMar 21, 2016
  0c599d7

Commits on Mar 15, 2016

• Fix GetOpList and GetPythonWrappers SWIG wrappers for Python 3.

  

  keveman

  authored and

  tensorflower-gardener

  committedMar 15, 2016
  b27022d

Pagination

PreviousNext