Commits

Breadcrumbs

History for

onmaster

User selector

All users

All time

Commit History

Commits on Jun 21, 2017

• Make changes to prepare for the fused option of batch norm to be set to None (None means using fused batch norm if possible).

  Show description for b3f33ad
  

  Yao Zhang

  authored and

  tensorflower-gardener

  committed
  b3f33ad

  Copy full SHA for b3f33ad

Commits on Jun 14, 2017

• Updating a comment in Conv2D to correctly specify the kernel dimension ordering (height, width).

  Show description for 4fc83c7
  tensorflower-gardener

  committed
  4fc83c7

  Copy full SHA for 4fc83c7

Commits on Jun 7, 2017

• tf.layers.conv2d use_bias=True to use nn.bias_add

  Show description for b94540e
  

  tfboyd

  authored and

  tensorflower-gardener

  committed
  b94540e

  Copy full SHA for b94540e

Commits on Jun 6, 2017

• Add fp16 support to convolutional layers that support it.

  Show description for cc1a02d
  

  fchollet

  authored and

  tensorflower-gardener

  committed
  cc1a02d

  Copy full SHA for cc1a02d

Commits on Jun 5, 2017

• Fixed triggering create device multiple times

  Show description for a5909d6
  

  tfboyd

  authored and

  tensorflower-gardener

  committed
  a5909d6

  Copy full SHA for a5909d6

Commits on Jun 3, 2017

• Add fused batch norm to tf.layers.

  Show description for 675d36b
  

  Yao Zhang

  authored and

  tensorflower-gardener

  committed
  675d36b

  Copy full SHA for 675d36b

Commits on May 31, 2017

• Refactoring of layer name autogeneration, to remove a graph serialization warning.

  Show description for e405b0f
  

  fchollet

  authored and

  tensorflower-gardener

  committed
  e405b0f

  Copy full SHA for e405b0f

Commits on May 24, 2017

• Support tensors as dropout rates again, by removing the min(max(..)..) clipping

  Show description for c98dab0
  tensorflower-gardener

  committed
  c98dab0

  Copy full SHA for c98dab0

Commits on May 19, 2017

• Fixed up the documentation for tf.contrib.layers.batch_norm and tf.layers.batch_normalization explaining how to include the update ops in the train_op. Also simplified the existing tf.layers.batch_…

  Show description for 675f0f9
  

  nealwu

  authored and

  tensorflower-gardener

  committed
  675f0f9

  Copy full SHA for 675f0f9

Commits on May 17, 2017

• Refactor Keras layers to rely on core TF layers, specifically:

  Show description for 35253fa
  

  fchollet

  authored and

  tensorflower-gardener

  committed
  35253fa

  Copy full SHA for 35253fa

Commits on May 15, 2017

• Fixed the documentation for tf.layers.conv2d to indicate that `use_bias` takes precedence over `bias_initializer`

  Show description for 652252e
  

  nealwu

  authored and

  tensorflower-gardener

  committed
  652252e

  Copy full SHA for 652252e

Commits on May 11, 2017

• Fixes markdown formatting.

  Show description for a9c50ea
  tensorflower-gardener

  committed
  a9c50ea

  Copy full SHA for a9c50ea

• Merge changes from github.

  Show description for ee112cf
  

  benoitsteiner

  authored and

  tensorflower-gardener

  committed
  ee112cf

  Copy full SHA for ee112cf

Commits on May 5, 2017

• Merge changes from github.

  Show description for 692fad2
  

  ringw

  authored and

  tensorflower-gardener

  committed
  692fad2

  Copy full SHA for 692fad2

Commits on May 4, 2017

• TensorFlow documentation fixes

  Show description for 4a4a1eb
  

  nealwu

  authored and

  tensorflower-gardener

  committed
  4a4a1eb

  Copy full SHA for 4a4a1eb

Commits on May 3, 2017

• [tf layers] Delay marking a layer as built until the end of its first apply().

  Show description for 485a24e
  

  ebrevdo

  authored and

  tensorflower-gardener

  committed
  485a24e

  Copy full SHA for 485a24e

Commits on Apr 27, 2017

• Refactor Keras layers to rely on core TF layers.

  Show description for 1e3e5d4
  

  fchollet

  authored and

  tensorflower-gardener

  committed
  1e3e5d4

  Copy full SHA for 1e3e5d4

Commits on Apr 22, 2017

• Merge changes from github.

  Show description for 3269423
  

  caisq

  authored and

  tensorflower-gardener

  committed
  3269423

  Copy full SHA for 3269423

• RNNCell is now a subclass of tf.layers._Layer.

  Show description for e8482ab
  

  ebrevdo

  authored and

  tensorflower-gardener

  committed
  e8482ab

  Copy full SHA for e8482ab

Commits on Apr 21, 2017

• Introduce TFDecorator, a base class for Python TensorFlow decorators. Provides basic introspection and "unwrap" services, allowing tooling code to fully 'understand' the wrapped object.

  Show description for 8e50419
  

  Charles Nicholson

  authored and

  tensorflower-gardener

  committed
  8e50419

  Copy full SHA for 8e50419

Commits on Apr 18, 2017

• Merge changes from github.

  Show description for 69a4cf8
  

  Patrick Nguyen

  authored and

  tensorflower-gardener

  committed
  69a4cf8

  Copy full SHA for 69a4cf8

Commits on Apr 12, 2017

• Ensure that moving_mean and moving_variance are grabbed before being updated. Note that this should not be necessary since they are only updated during training and used during testing (unlike reno…

  Show description for 348b3b1
  tensorflower-gardener

  committed
  348b3b1

  Copy full SHA for 348b3b1

Commits on Apr 9, 2017

• Prevent the renorm_weight from being updated too early.

  Show description for a61cff6
  tensorflower-gardener

  committed
  a61cff6

  Copy full SHA for a61cff6

Commits on Apr 7, 2017

• Adds batch renormalization.

  Show description for 55c809b
  tensorflower-gardener

  committed
  55c809b

  Copy full SHA for 55c809b

Commits on Apr 5, 2017

• Merge changes from github.

  Show description for ccbc899
  tensorflower-gardener

  committed
  ccbc899

  Copy full SHA for ccbc899

Commits on Mar 29, 2017

• Ensure tf.layers._Layer objects are not used across multiple graphs.

  Show description for 6dcfef1
  

  ebrevdo

  authored and

  tensorflower-gardener

  committed
  6dcfef1

  Copy full SHA for 6dcfef1

• Fix shape handling errir in convolutional layer.

  Show description for 58bfd6e
  tensorflower-gardener

  committed
  58bfd6e

  Copy full SHA for 58bfd6e

Commits on Mar 23, 2017

• Merge changes from github.

  Show description for bc456e3
  

  martinwicke

  authored and

  tensorflower-gardener

  committed
  bc456e3

  Copy full SHA for bc456e3

• Fixed badly formed error message.

  Show description for 61de656
  tensorflower-gardener

  committed
  61de656

  Copy full SHA for 61de656

• Undo a breaking change to the public API; while retaining the bugfix of #8504.

  Show description for 81761d7
  

  ebrevdo

  authored and

  tensorflower-gardener

  committed
  81761d7

  Copy full SHA for 81761d7

Commits on Mar 21, 2017

• Add support for the NCHW data_format for 3d operations (convolution, pooling).

  Show description for 61f3022
  tensorflower-gardener

  committed
  61f3022

  Copy full SHA for 61f3022

• Change behavior of tf.layers._Layer: scope (if not provided) is set lazily.

  Show description for cb9d67d
  

  ebrevdo

  authored and

  tensorflower-gardener

  committed
  cb9d67d

  Copy full SHA for cb9d67d

Commits on Mar 16, 2017

• Fix separable convolution bias check

  Show description for a0ca4bc
  tensorflower-gardener

  committed
  a0ca4bc

  Copy full SHA for a0ca4bc

Commits on Mar 10, 2017

• Merge changes from github.

  Show description for 0386a01
  

  Dandelion Mané

  authored and

  tensorflower-gardener

  committed
  0386a01

  Copy full SHA for 0386a01

Commits on Mar 2, 2017

• Fix typos and inconsistencies in layers docstrings.

  Show description for 37b4005
  

  fchollet

  authored and

  tensorflower-gardener

  committed
  37b4005

  Copy full SHA for 37b4005