Commits

Breadcrumbs

History for

onr0.10

User selector

All users

All time

Commit History

Commits on Aug 1, 2016

• Simplify tf.nn.sufficient_statistics.

  Show description for b1d9ef5
  tensorflower-gardener

  committed
  b1d9ef5

  Copy full SHA for b1d9ef5

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 Jul 22, 2016

• Fixing bug in which a Tensor was being checked as a Python bool, causing an error.

  Show description for 3acdd0a
  tensorflower-gardener

  committed
  3acdd0a

  Copy full SHA for 3acdd0a

Commits on Jul 19, 2016

• Improved support for fp16

  Show description for 9223380
  

  benoitsteiner

  authored and

  tensorflower-gardener

  committed
  9223380

  Copy full SHA for 9223380

Commits on Jun 30, 2016

• Merge changes from github.

  Show description for 533d891
  tensorflower-gardener

  committed
  533d891

  Copy full SHA for 533d891

Commits on Jun 29, 2016

• Updated the batch normalization code to make it work with fp16

  Show description for 3a48e6a
  

  benoitsteiner

  authored and

  tensorflower-gardener

  committed
  3a48e6a

  Copy full SHA for 3a48e6a

Commits on Jun 21, 2016

• Fixing bug in which negative variances were produced due to order of operations.

  Show description for f6c59f0
  tensorflower-gardener

  committed
  f6c59f0

  Copy full SHA for f6c59f0

Commits on Jun 16, 2016

• Move CTC out of contrib and document.

  Show description for 67d3c91
  

  ebrevdo

  authored and

  tensorflower-gardener

  committed
  67d3c91

  Copy full SHA for 67d3c91

Commits on Jun 14, 2016

• Execute TODOs to

  Show description for a558c6e
  

  itsmeolivia

  authored and

  tensorflower-gardener

  committed
  a558c6e

  Copy full SHA for a558c6e

Commits on Jun 8, 2016

• Make RNN api public.

  Show description for bab4fc4
  

  ebrevdo

  authored and

  tensorflower-gardener

  committed
  bab4fc4

  Copy full SHA for bab4fc4

Commits on Jun 3, 2016

• Merge changes from github.

  Show description for 60796d7
  

  martinwicke

  authored and

  tensorflower-gardener

  committed
  60796d7

  Copy full SHA for 60796d7

• Morphological filtering operations: Dilation and erosion.

  Show description for 283782a
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  283782a

  Copy full SHA for 283782a

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 26, 2016

• Allow the output of the separable_conv2d to have the same number of channels as the corresponding input.

  Show description for a9112b4
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  a9112b4

  Copy full SHA for a9112b4

Commits on May 24, 2016

• Documentation fix for tf.nn.sufficient_statistics().

  Show description for ef120cb
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  ef120cb

  Copy full SHA for ef120cb

• Allow callers to provide a shift value in tf.nn.moments().

  Show description for 4ea0733
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  4ea0733

  Copy full SHA for 4ea0733

Commits on May 23, 2016

• Fix a documentation bug in sampled_softmax_loss

  Show description for 9eb50f1
  

  girving

  authored and

  tensorflower-gardener

  committed
  9eb50f1

  Copy full SHA for 9eb50f1

Commits on May 20, 2016

• There is evidence that the 'shift' strategy in computing the sufficient statistics of the moments is actually leading to worse numerical stability for batch normalization.

  Show description for e47dc85
  

  vincentvanhoucke

  authored and

  tensorflower-gardener

  committed
  e47dc85

  Copy full SHA for e47dc85

Commits on May 13, 2016

• Change sigmoid_cross_entropy_with_logits to fix gradients at 0.

  Show description for 9d2e920
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  9d2e920

  Copy full SHA for 9d2e920

Commits on May 4, 2016

• Add a comment on a more stable form of the logistic loss.

  Show description for edb8f90
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  edb8f90

  Copy full SHA for edb8f90

• Add support for 3d convolutions and pooling. CPU kernels use Eigen, GPU kernels use CuDNN.

  Show description for 6a187cc
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  6a187cc

  Copy full SHA for 6a187cc

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 Apr 19, 2016

• Merge changes from github.

  Show description for 5c9bc51
  

  ilblackdragon

  authored and

  tensorflower-gardener

  committed
  5c9bc51

  Copy full SHA for 5c9bc51

Commits on Apr 18, 2016

• Make native depthwise Conv2d the default.

  Show description for 975094c
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  975094c

  Copy full SHA for 975094c

Commits on Apr 2, 2016

• PyLint updates.

  Show description for e01f6bf
  

  josh11b

  authored and

  tensorflower-gardener

  committed
  e01f6bf

  Copy full SHA for e01f6bf

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 23, 2016

• Merge changes from github, some fixes to adhere somewhat

  Show description for 37606a4
  

  Vijay Vasudevan

  authored and

  tensorflower-gardener

  committed
  37606a4

  Copy full SHA for 37606a4

Commits on Mar 11, 2016

• Implements "bool strictness" for `tf.Tensor`.

  Show description for f6f77ac
  

  mrry

  authored and

  tensorflower-gardener

  committed
  f6f77ac

  Copy full SHA for f6f77ac

Commits on Mar 7, 2016

• fix docstring for zero_fraction

  Show description for 972998e
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  972998e

  Copy full SHA for 972998e

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 Mar 2, 2016

• Make the image, nn, and train packages only expose documented symbols.

  Show description for 64017d1
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  64017d1

  Copy full SHA for 64017d1

Commits on Mar 1, 2016

• Added a benchmark for the performance of dyanmic rnn with memory swapping. For the large model, the overhead is roughly 35%. More improvements over this base-line implementation are coming. But if…

  Show description for 384480d
  

  yuanbyu

  authored and

  tensorflower-gardener

  committed
  384480d

  Copy full SHA for 384480d

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