Commits

Breadcrumbs

History for

onquantize-graph-docs

User selector

All users

All time

Commit History

Commits on Jan 13, 2017

• Make tfdbg tests pass in Windows Bazel build (#6821)

  

  caisq

  authored and

  drpngx

  committed
  018583b

  Copy full SHA for 018583b

• Merge changes from github.

  Show description for 769496b
  

  Patrick Nguyen

  authored and

  tensorflower-gardener

  committed
  769496b

  Copy full SHA for 769496b

• More new ops for speeding up input pipline: stage and unstage.

  Show description for 85acf52
  tensorflower-gardener

  committed
  85acf52

  Copy full SHA for 85acf52

Commits on Jan 12, 2017

• Kernels and ops for all optimizers when using resource variables.

  Show description for dd9684f
  tensorflower-gardener

  committed
  dd9684f

  Copy full SHA for dd9684f

• Add a placeholder for hexagon_controller

  Show description for 6ee46a3
  tensorflower-gardener

  committed
  6ee46a3

  Copy full SHA for 6ee46a3

• Switch open-source to use jemalloc for CPU Tensor memory allocation, gRPC, and other places where we call malloc/free.

  Show description for 83c6e0c
  

  Jonathan Hseu

  authored and

  tensorflower-gardener

  committed
  83c6e0c

  Copy full SHA for 83c6e0c

Commits on Jan 11, 2017

• Add an option to run inference on hvx into build script

  Show description for 54e2be2
  tensorflower-gardener

  committed
  54e2be2

  Copy full SHA for 54e2be2

• Add optimized reverse kernel for reversing rows of a 3D tensor. This is used in

  Show description for 51e6e84
  tensorflower-gardener

  committed
  51e6e84

  Copy full SHA for 51e6e84

• Optimize eight-bit graphs by removing RequantizationRanges

  Show description for b42ba8a
  

  petewarden

  authored and

  tensorflower-gardener

  committed
  b42ba8a

  Copy full SHA for b42ba8a

Commits on Jan 10, 2017

• Register Slice op for quantized types.

  Show description for a81cff2
  tensorflower-gardener

  committed
  a81cff2

  Copy full SHA for a81cff2

Commits on Jan 9, 2017

• Change reduce_join to treat reduction_indices in the same way as

  Show description for a910b52
  tensorflower-gardener

  committed
  a910b52

  Copy full SHA for a910b52

• Fused version of adjust_saturation.

  Show description for 3e265db
  tensorflower-gardener

  committed
  3e265db

  Copy full SHA for 3e265db

Commits on Jan 7, 2017

• Define variables derived from int64s as int64.

  Show description for 004fc82
  tensorflower-gardener

  committed
  004fc82

  Copy full SHA for 004fc82

Commits on Jan 6, 2017

• Additional Ops files were added to the android build rules.

  Show description for 33ccce4
  

  yoni

  authored and

  tensorflower-gardener

  committed
  33ccce4

  Copy full SHA for 33ccce4

• Add support for gradient calculation for concat_v2 with negative axis to fix github

  Show description for f3d5990
  tensorflower-gardener

  committed
  f3d5990

  Copy full SHA for f3d5990

• Allow empty tf.Features in the FeatureLists for sparse sequence examples.

  Show description for a245f73
  tensorflower-gardener

  committed
  a245f73

  Copy full SHA for a245f73

• Fix input image handling in hvx execution test

  Show description for 6745072
  tensorflower-gardener

  committed
  6745072

  Copy full SHA for 6745072

• Improve speed of scatter_update on the cpu for ASSIGN operations.

  Show description for 9d612e7
  tensorflower-gardener

  committed
  9d612e7

  Copy full SHA for 9d612e7

Commits on Jan 5, 2017

• Mark gemmlowp result as initialized.

  Show description for b380aae
  

  Patrick Nguyen

  authored and

  tensorflower-gardener

  committed
  b380aae

  Copy full SHA for b380aae

• Add a way to set image bmp, dump result and change log level of SOC graph transferer

  Show description for 93ee6cc
  tensorflower-gardener

  committed
  93ee6cc

  Copy full SHA for 93ee6cc

• Make EmptyOp compatible with GPU.

  Show description for 9563060
  

  zheng-xq

  authored and

  tensorflower-gardener

  committed
  9563060

  Copy full SHA for 9563060

Commits on Jan 4, 2017

• Adds V2 versions of Queue and Reader ops using ResourceHandles.

  Show description for 2482564
  tensorflower-gardener

  committed
  2482564

  Copy full SHA for 2482564

• Mark gemmlowp result as initialized.

  Show description for 1118de0
  

  Patrick Nguyen

  authored and

  tensorflower-gardener

  committed
  1118de0

  Copy full SHA for 1118de0

• Optimize im2col section of quantized convolution

  Show description for d4109cb
  

  petewarden

  authored and

  tensorflower-gardener

  committed
  d4109cb

  Copy full SHA for d4109cb

• Adds the following new ops:

  Show description for 6ec984e
  tensorflower-gardener

  committed
  6ec984e

  Copy full SHA for 6ec984e

Commits on Jan 3, 2017

• Internal only change.

  Show description for e39df7d
  

  dm-jrae

  authored and

  tensorflower-gardener

  committed
  e39df7d

  Copy full SHA for e39df7d

• Corrected typo.

  Show description for eb020c9
  tensorflower-gardener

  committed
  eb020c9

  Copy full SHA for eb020c9

• Merge changes from github.

  Show description for 46d2c28
  tensorflower-gardener

  committed
  46d2c28

  Copy full SHA for 46d2c28

Commits on Dec 29, 2016

• Adds V3 version of TensorArray ops. All use resource handles and TensorArrayV3 has a flow output.

  Show description for 0295391
  tensorflower-gardener

  committed
  0295391

  Copy full SHA for 0295391

Commits on Dec 28, 2016

• Minor fix: allowing min_object_overlap to be 0.

  Show description for 4dfc5d6
  tensorflower-gardener

  committed
  4dfc5d6

  Copy full SHA for 4dfc5d6

Commits on Dec 23, 2016

• Fix for quantized convolution optimization

  Show description for e27bc77
  

  petewarden

  authored and

  tensorflower-gardener

  committed
  e27bc77

  Copy full SHA for e27bc77

Commits on Dec 22, 2016

• Merge changes from github.

  Show description for bed8383
  

  Jonathan Hseu

  authored and

  tensorflower-gardener

  committed
  bed8383

  Copy full SHA for bed8383

• Define GEMMLOWP_ALLOW_SLOW_SCALAR_FALLBACK to allow building with

  Show description for c4b09b5
  tensorflower-gardener

  committed
  c4b09b5

  Copy full SHA for c4b09b5

• Create Graph Transform Tool for rewriting model files.

  Show description for 0f0e29e
  

  petewarden

  authored and

  tensorflower-gardener

  committed
  0f0e29e

  Copy full SHA for 0f0e29e

• Only register Reshape CPU kernels with type constraint of int32 on

  Show description for 21d8bb6
  

  Vijay Vasudevan

  authored and

  tensorflower-gardener

  committed
  21d8bb6

  Copy full SHA for 21d8bb6