Commits

Breadcrumbs

History for

onv

User selector

All users

All time

Commit History

Commits on Jan 10, 2017

• Make moments numerically stable by default. Added tests for moments.

  Show description for fdbd02c
  

  sguada

  authored and

  tensorflower-gardener

  committed
  fdbd02c

  Copy full SHA for fdbd02c

Commits on Dec 17, 2016

• Remove hourglass imports from even more tests

  Show description for 58201a0
  

  jart

  authored and

  tensorflower-gardener

  committed
  58201a0

  Copy full SHA for 58201a0

Commits on Dec 13, 2016

• Swap the order of inputs and labels of the sampled_softmax_loss.

  Show description for a467759
  

  Jianwei Xie

  authored and

  tensorflower-gardener

  committed
  a467759

  Copy full SHA for a467759

Commits on Dec 7, 2016

• Swap the input and label arguments in nce_loss

  Show description for 761b12e
  

  Jianwei Xie

  authored and

  tensorflower-gardener

  committed
  761b12e

  Copy full SHA for 761b12e

Commits on Dec 2, 2016

• Swaps the predictions/labels arguments in log_poisson_loss.

  Show description for ee36760
  tensorflower-gardener

  committed
  ee36760

  Copy full SHA for ee36760

Commits on Nov 23, 2016

• Seal tf.nn's interface.

  Show description for af6c3f0
  

  Patrick Nguyen

  authored and

  tensorflower-gardener

  committed
  af6c3f0

  Copy full SHA for af6c3f0

Commits on Aug 24, 2016

• Implement the Concatenated ReLU (CReLu) non-linearity. This non-linearity makes it possible to reduce the depth of lower layers in a convolution tower by a factor of two at no cost in accuracy. See…

  Show description for 8945cec
  

  vincentvanhoucke

  authored and

  tensorflower-gardener

  committed
  8945cec

  Copy full SHA for 8945cec

Commits on Aug 22, 2016

• Fix L2Normalize when passing a list of dims. Fixes #3932.

  Show description for 3b7153d
  

  Vijay Vasudevan

  authored and

  tensorflower-gardener

  committed
  3b7153d

  Copy full SHA for 3b7153d

• Make dropout avoid work if keep_prob == 1

  Show description for 542de31
  

  girving

  authored and

  tensorflower-gardener

  committed
  542de31

  Copy full SHA for 542de31

Commits on Jul 27, 2016

• Adding log poisson loss in order to support Poisson Regression.

  Show description for 6ab6800
  tensorflower-gardener

  committed
  6ab6800

  Copy full SHA for 6ab6800

Commits on Jun 6, 2016

• Merge changes from github.

  Show description for b1b2dc8
  

  martinwicke

  authored and

  tensorflower-gardener

  committed
  b1b2dc8

  Copy full SHA for b1b2dc8

Commits on Jun 2, 2016

• Update copyright for 3p/tf/python.

  Show description for 0cf9ed3
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  0cf9ed3

  Copy full SHA for 0cf9ed3

Commits on May 13, 2016

• Add gradient for log_softmax and test.

  Show description for 1d094c9
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  1d094c9

  Copy full SHA for 1d094c9

Commits on May 12, 2016

• Split nn_test into more manageable pieces.

  Show description for d107fec
  

  martinwicke

  authored and

  tensorflower-gardener

  committed
  d107fec

  Copy full SHA for d107fec

Commits on May 10, 2016

• This change causes Ops with multiple outputs to return them as a NamedTuple

  Show description for adcee0e
  

  davidsoergel

  authored and

  tensorflower-gardener

  committed
  adcee0e

  Copy full SHA for adcee0e

Commits on May 4, 2016

• Fix nn_test.py

  Show description for a1bc10f
  

  gunan

  authored and

  tensorflower-gardener

  committed
  a1bc10f

  Copy full SHA for a1bc10f

• Added support for fp16 to the l2loss kernel

  Show description for 7ae67a1
  

  benoitsteiner

  authored and

  tensorflower-gardener

  committed
  7ae67a1

  Copy full SHA for 7ae67a1

Commits on Apr 21, 2016

• Add atrous convolution (atrous_conv2d op). Fixes #1815.

  Show description for c1aa642
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  c1aa642

  Copy full SHA for c1aa642

Commits on Mar 30, 2016

• Adds weighted_cross_entropy that allows one to trade off recall vs. precision penalty.

  Show description for edb0ad8
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  edb0ad8

  Copy full SHA for edb0ad8

Commits on Mar 25, 2016

• fixing linting errors

  Show description for 8016e72
  

  itsmeolivia

  authored and

  tensorflower-gardener

  committed
  8016e72

  Copy full SHA for 8016e72

Commits on Mar 14, 2016

• Another 1 in 1000 numerical failure.

  Show description for 0bcaf27
  

  vincentvanhoucke

  authored and

  tensorflower-gardener

  committed
  0bcaf27

  Copy full SHA for 0bcaf27

• De-flakify testBatchNormArbitraryShapes

  Show description for bc27f89
  

  vincentvanhoucke

  authored and

  tensorflower-gardener

  committed
  bc27f89

  Copy full SHA for bc27f89

Commits on Mar 3, 2016

• Add shape verification for `tf.nn.sigmoid_cross_entropy_with_logits()`.

  Show description for 80ddfbf
  

  mrry

  authored and

  tensorflower-gardener

  committed
  80ddfbf

  Copy full SHA for 80ddfbf

Commits on Feb 26, 2016

• Initial version of the open-source distributed TensorFlow runtime.

  Show description for 00986d4
  

  mrry

  authored and

  tensorflower-gardener

  committed
  00986d4

  Copy full SHA for 00986d4

Commits on Feb 25, 2016

• Switch nn.moments() to using a one-pass stable algorithm.

  Show description for bce6216
  

  vincentvanhoucke

  authored and

  tensorflower-gardener

  committed
  bce6216

  Copy full SHA for bce6216

Commits on Feb 24, 2016

• Support leaving the offset (beta) parameter out in batch_normalization, in which case no offset will be added after normalization.

  Show description for 4afef14
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  4afef14

  Copy full SHA for 4afef14

Commits on Feb 19, 2016

• Loosen test tolerances which seem to fail on Mac.

  Show description for 369ca60
  

  vincentvanhoucke

  authored and

  tensorflower-gardener

  committed
  369ca60

  Copy full SHA for 369ca60

Commits on Feb 18, 2016

• Add utility function for creating partitioned variables.

  Show description for 03869e6
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  03869e6

  Copy full SHA for 03869e6

Commits on Feb 11, 2016

• Get rid of some import cruft.

  Show description for 7760ce5
  

  josh11b

  authored and

  tensorflower-gardener

  committed
  7760ce5

  Copy full SHA for 7760ce5

Commits on Feb 7, 2016

• https://github.com/tensorflow/tensorflow/issues/999

  Show description for d53f06b
  

  vincentvanhoucke

  authored and

  martinwicke

  committed
  d53f06b

  Copy full SHA for d53f06b

Commits on Jan 6, 2016

• Rename deconv2d to conv2d_transpose and expose publicly

  Show description for 9edd13a
  

  A. Unique TensorFlower

  authored and

  Vijay Vasudevan

  committed
  9edd13a

  Copy full SHA for 9edd13a

Commits on Dec 1, 2015

• TensorFlow: upstream changes to git

  Show description for 795f35d
  

  Vijay Vasudevan

  committed
  795f35d

  Copy full SHA for 795f35d

Commits on Nov 25, 2015

• TensorFlow: Upstream changes to git

  Show description for 854f49b
  keveman

  committed
  854f49b

  Copy full SHA for 854f49b

Commits on Nov 20, 2015

• TensorFlow: Improve performance of Alexnet

  Show description for 9c3043f
  keveman

  committed
  9c3043f

  Copy full SHA for 9c3043f

Commits on Nov 13, 2015

• TensorFlow: Upstream changes from afternoon.

  Show description for d50565b
  

  Vijay Vasudevan

  committed
  d50565b

  Copy full SHA for d50565b