Commits

Breadcrumbs

History for

onskleinfeld-patch-1

User selector

All users

All time

Commit History

Commits on Sep 9, 2016

• Merge commit for internal changes

  

  Daniel Mané

  committed
  16c35dc

  Copy full SHA for 16c35dc

• Allow seq2seq attention decoder to receive dynamic length attention states.

  Show description for 6a5e421
  tensorflower-gardener

  committed
  6a5e421

  Copy full SHA for 6a5e421

Commits on Sep 1, 2016

• seq2seq.py: Fixed the documentation to be consistent with the code (#3802)

  Show description for a8be586
  

  sahiliitm

  authored and

  Vijay Vasudevan

  committed
  a8be586

  Copy full SHA for a8be586

Commits on Aug 16, 2016

• Recent code change added some lines that inadvertently passed scopes in seq2seq,

  Show description for a0fb26c
  

  Lukasz Kaiser

  authored and

  tensorflower-gardener

  committed
  a0fb26c

  Copy full SHA for a0fb26c

Commits on Aug 10, 2016

• Added support for fp16 to the seq2seq primitives

  Show description for 6011c57
  

  benoitsteiner

  authored and

  tensorflower-gardener

  committed
  6011c57

  Copy full SHA for 6011c57

Commits on Aug 9, 2016

• Deprecate op_scope by extending name_scope and move all callers to the new name_scope.

  Show description for 1df3fb0
  

  Lukasz Kaiser

  authored and

  tensorflower-gardener

  committed
  1df3fb0

  Copy full SHA for 1df3fb0

Commits on Aug 4, 2016

• Make attention_decoder in seq2seq work with state_is_tuple=True. (Fix #3056.)

  Show description for 6215671
  

  Lukasz Kaiser

  authored and

  tensorflower-gardener

  committed
  6215671

  Copy full SHA for 6215671

Commits on Jul 2, 2016

• Make the seq2seq module work well with RNNCells with tuple-like states. (#3056)

  Show description for d1f35b4
  

  Lukasz Kaiser

  authored and

  tensorflower-gardener

  committed
  d1f35b4

  Copy full SHA for d1f35b4

Commits on Jun 4, 2016

• Change embedding_tied_rnn_seq2seq function to provide an option of only predicting from specified decoding symbols (as opposed to predicting from all symbols) during the decoding phase.

  Show description for cc9c311
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  cc9c311

  Copy full SHA for cc9c311

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

• Merge changes from github.

  Show description for 8cc567b
  

  Vijay Vasudevan

  authored and

  tensorflower-gardener

  committed
  8cc567b

  Copy full SHA for 8cc567b

Commits on May 11, 2016

• rnn_cell.linear and models.rnn are going away.

  Show description for 396b586
  

  ebrevdo

  authored and

  tensorflower-gardener

  committed
  396b586

  Copy full SHA for 396b586

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 13, 2016

• Remove RNNCell's input_size property.

  Show description for 871ee01
  

  ebrevdo

  authored and

  tensorflower-gardener

  committed
  871ee01

  Copy full SHA for 871ee01

Commits on Mar 30, 2016

• Make embedding_size an explicit argument of embedding_* functions in seq2seq.

  Show description for a796131
  

  lukaszkaiser

  authored and

  tensorflower-gardener

  committed
  a796131

  Copy full SHA for a796131

Commits on Mar 18, 2016

• Make EmbeddingWrapper accept an explicit embedding_size argument so it does not rely on cell.input_size.

  Show description for 75d60a2
  

  lukaszkaiser

  authored and

  tensorflower-gardener

  committed
  75d60a2

  Copy full SHA for 75d60a2

Commits on Mar 11, 2016

• Fix dependencies bugs

  Show description for 56f1d64
  

  ebrevdo

  authored and

  tensorflower-gardener

  committed
  56f1d64

  Copy full SHA for 56f1d64

Commits on Mar 10, 2016

• Make sparse_softmax_cross_entropy_with_logits handle int32 labels

  Show description for fd839dc
  

  girving

  authored and

  tensorflower-gardener

  committed
  fd839dc

  Copy full SHA for fd839dc

Commits on Feb 17, 2016

• Merge changes from github.

  Show description for fe056f0
  

  Vijay Vasudevan

  authored and

  tensorflower-gardener

  committed
  fe056f0

  Copy full SHA for fe056f0

Commits on Feb 11, 2016

• Add an optional argument to model_with_buckets to get per-example loss and

  Show description for 18cbcdf
  

  lukaszkaiser

  authored and

  tensorflower-gardener

  committed
  18cbcdf

  Copy full SHA for 18cbcdf

Commits on Feb 10, 2016

• Add one-to-many multi-task sequence-to-sequence model.

  Show description for bc918af
  

  A. Unique TensorFlower

  authored and

  tensorflower-gardener

  committed
  bc918af

  Copy full SHA for bc918af

Commits on Feb 5, 2016

• Use sparse_softmax_cross_entropy_with_logits in seq2seq in sequence_losses.

  Show description for 63b7a98
  

  lukaszkaiser

  authored and

  keveman

  committed
  63b7a98

  Copy full SHA for 63b7a98

Commits on Feb 4, 2016

• Improving seq2seq code style and documentation.

  Show description for 6607d30
  

  lukaszkaiser

  authored and

  Vijay Vasudevan

  committed
  6607d30

  Copy full SHA for 6607d30

Commits on Jan 28, 2016

• Breaking change in TF RNN python api: Return the final state instead of the

  Show description for fea55e1
  

  ebrevdo

  authored and

  Vijay Vasudevan

  committed
  fea55e1

  Copy full SHA for fea55e1

Commits on Jan 16, 2016

• Added an option to the attention decoder to initialize the attention

  Show description for a1a3c03
  

  A. Unique TensorFlower

  authored and

  keveman

  committed
  a1a3c03

  Copy full SHA for a1a3c03

Commits on Jan 7, 2016

• some linting fixes to changes brought in from the public.

  Show description for 6cc392e
  

  Vijay Vasudevan

  committed
  6cc392e

  Copy full SHA for 6cc392e

Commits on Dec 18, 2015

• Prevent seq2seq.model_with_buckets from leaking variable scope reuse.

  Show description for 69b38d3
  

  lukaszkaiser

  authored and

  Vijay Vasudevan

  committed
  69b38d3

  Copy full SHA for 69b38d3

Commits on Dec 2, 2015

• TensorFlow: Upstream changes to git.

  Show description for bf6b536
  

  Vijay Vasudevan

  committed
  bf6b536

  Copy full SHA for bf6b536

Commits on Dec 1, 2015

• TensorFlow: upstream changes to git

  Show description for 795f35d
  

  Vijay Vasudevan

  committed
  795f35d

  Copy full SHA for 795f35d

Loading