Commits

Branch selector

 master

User selector

langmore

Datepicker

All time

Commit History

Commits on Sep 26, 2019

• Ensure convert to tensor is only done once for input args to LinearOperators.

  Show description for d9f9ea7
  

  langmore

  authored and

  tensorflower-gardener

  committed
  d9f9ea7

  Copy full SHA for d9f9ea7

Commits on Sep 10, 2019

• Verify linear_operator_circulant and linear_operator_toeplitz are tape safe.

  Show description for 42fcc3b
  

  langmore

  authored and

  tensorflower-gardener

  committed
  42fcc3b

  Copy full SHA for 42fcc3b

Commits on Aug 29, 2019

• Meta linear operators tested for "tape safe"

  Show description for 709d160
  

  langmore

  authored and

  tensorflower-gardener

  committed
  709d160

  Copy full SHA for 709d160

Commits on Aug 28, 2019

• Deprecate `graph_parents` init arg and property in LinearOperator. These were only used for `values` kwarg in name_scope, and therefore not needed in TF2.

  Show description for 65fa1f8
  

  langmore

  authored and

  tensorflower-gardener

  committed
  65fa1f8

  Copy full SHA for 65fa1f8

Commits on Aug 26, 2019

• Remove LinearOperator.graph_parents calls from contrib/distributions/... This will allow

  Show description for 705f4ab
  

  langmore

  authored and

  tensorflower-gardener

  committed
  705f4ab

  Copy full SHA for 705f4ab

Commits on Aug 21, 2019

• LinearOperatorLowRankUpdate made "tape safe" for TF2 compliance.

  Show description for a92016d
  

  langmore

  authored and

  tensorflower-gardener

  committed
  a92016d

  Copy full SHA for a92016d

• s/get_shape()/shape in linalg. This is the TF2 way.

  Show description for c087e10
  

  langmore

  authored and

  tensorflower-gardener

  committed
  c087e10

  Copy full SHA for c087e10

• LinearOperator (base class) now uses `is_tensor or is_ref` to check graph

  Show description for 1f9be63
  

  langmore

  authored and

  tensorflower-gardener

  committed
  1f9be63

  Copy full SHA for 1f9be63

Commits on Jul 31, 2019

• Increase tolerance in LinearOperatorCirculant3D test. Was failing on python 3

  Show description for adc63b1
  

  langmore

  authored and

  tensorflower-gardener

  committed
  adc63b1

  Copy full SHA for adc63b1

Commits on Jul 29, 2019

• Quiet warning about default `to_dense` call in base class LinearOperator.

  Show description for 9d4cb17
  

  langmore

  authored and

  tensorflower-gardener

  committed
  9d4cb17

  Copy full SHA for 9d4cb17

Commits on Jul 24, 2019

• TESTFIX: `tests_to_skip` staticmethod renamed to `skip_these_tests`. This

  Show description for 8cbe52a
  

  langmore

  authored and

  tensorflower-gardener

  committed
  8cbe52a

  Copy full SHA for 8cbe52a

Commits on Jul 23, 2019

• Make non-meta linear operators (other than Circulant/Toeplitz) tape safe.

  Show description for 95bcd43
  

  langmore

  authored and

  tensorflower-gardener

  committed
  95bcd43

  Copy full SHA for 95bcd43

Commits on Jul 17, 2019

• Make LinearOperatorDiag tape safe. Also add a standard "check_tape_safe" method in the base test class.

  Show description for 9eabc5c
  

  langmore

  authored and

  tensorflower-gardener

  committed
  9eabc5c

  Copy full SHA for 9eabc5c

Commits on Jul 16, 2019

• BUGFIX: LinearOperatorAdjoint.matvec was passing an `adjoint_arg` kwarg, but

  Show description for 5819055
  

  langmore

  authored and

  tensorflower-gardener

  committed
  5819055

  Copy full SHA for 5819055

Commits on Jul 15, 2019

• LinearOperator (base class) and LinearOperatorFullMatrix TF2 updates:

  Show description for 258cfc7
  

  langmore

  authored and

  tensorflower-gardener

  committed
  258cfc7

  Copy full SHA for 258cfc7

Commits on Jul 11, 2019

• Change linear_operator_util `convert_immutable_to_tensor` to

  Show description for c07e9e0
  

  langmore

  authored and

  tensorflower-gardener

  committed
  c07e9e0

  Copy full SHA for c07e9e0

Commits on Jul 4, 2019

• Automated rollback of commit 52435ea5a89b9c129894c31d10a3f42e4ed31d85

  Show description for ba730a4
  

  langmore

  authored and

  tensorflower-gardener

  committed
  ba730a4

  Copy full SHA for ba730a4

• LinearOperator (base class) and LinearOperatorFullMatrix TF2 updates:

  Show description for 52435ea
  

  langmore

  authored and

  tensorflower-gardener

  committed
  52435ea

  Copy full SHA for 52435ea

Commits on Jun 27, 2019

• BUGFIX: LinearOperatorAdjoint and LinearOperatorInversion were not using

  Show description for 2f8c378
  

  langmore

  authored and

  tensorflower-gardener

  committed
  2f8c378

  Copy full SHA for 2f8c378

Commits on Jun 20, 2019

• tfp/.../tensor_util functions copy-pasted into linear_operator_util

  Show description for c6a2201
  

  langmore

  authored and

  tensorflower-gardener

  committed
  c6a2201

  Copy full SHA for c6a2201

Commits on Jan 12, 2019

• Auto-set `is_self_adjoint` hint for LinearOperatorScaledIdentity.

  Show description for ff2cd83
  

  langmore

  authored and

  tensorflower-gardener

  committed
  ff2cd83

  Copy full SHA for ff2cd83

Commits on Nov 16, 2018

• LinearOperatorInversion added as a meta-class. Computes the naive inverse of

  Show description for 2226a4d
  

  langmore

  authored and

  tensorflower-gardener

  committed
  2226a4d

  Copy full SHA for 2226a4d

Commits on Nov 15, 2018

• Small updates to LinearOperatorComposition. Noticed while writing

  Show description for f18df6b
  

  langmore

  authored and

  tensorflower-gardener

  committed
  f18df6b

  Copy full SHA for f18df6b

Commits on Nov 8, 2018

• tf.linalg.matvec: Use tf.squeeze and expand_dims rather than implicit [...,

  Show description for 6c40bb5
  

  langmore

  authored and

  tensorflower-gardener

  committed
  6c40bb5

  Copy full SHA for 6c40bb5

Commits on Nov 6, 2018

• matvec added to math_ops.py. This is pure sugar:

  Show description for 1018eda
  

  langmore

  authored and

  tensorflower-gardener

  committed
  1018eda

  Copy full SHA for 1018eda

Commits on Oct 30, 2018

• linear_operator_util: Special case some broadcasting for matmul and

  Show description for 699445d
  

  langmore

  authored and

  tensorflower-gardener

  committed
  699445d

  Copy full SHA for 699445d

Commits on Sep 17, 2018

• Remove tensorflow/contrib/linalg library. linalg remains in core.

  Show description for d5f4c3a
  

  langmore

  authored and

  tensorflower-gardener

  committed
  d5f4c3a

  Copy full SHA for d5f4c3a

Commits on Sep 14, 2018

• Point VectorDiffeomixture to tf.linalg rather than tf.contrib.linalg

  Show description for b2cb6e2
  

  langmore

  authored and

  tensorflower-gardener

  committed
  b2cb6e2

  Copy full SHA for b2cb6e2

Commits on Sep 12, 2018

• Point distributions code to tf.linalg rather than tf.contrib.linalg

  Show description for 2710a18
  

  langmore

  authored and

  tensorflower-gardener

  committed
  2710a18

  Copy full SHA for 2710a18

• Add linear_operator_addition to tensorflow/python/. A subsequent CL

  Show description for 26509bf
  

  langmore

  authored and

  tensorflower-gardener

  committed
  26509bf

  Copy full SHA for 26509bf

Commits on May 7, 2018

• Add 'optonly' directive to linear_operator_circulant tests.

  Show description for c3fef21
  

  langmore

  authored and

  tensorflower-gardener

  committed
  c3fef21

  Copy full SHA for c3fef21

Commits on Apr 24, 2018

• Move LinearOperatorCirculant to third_party.

  Show description for f62c472
  

  langmore

  authored and

  tensorflower-gardener

  committed
  f62c472

  Copy full SHA for f62c472

Commits on Apr 19, 2018

• distribution_util moved into its own BUILD target, so linear_operator can depend on it.

  Show description for 5fbd21e
  

  langmore

  authored and

  tensorflower-gardener

  committed
  5fbd21e

  Copy full SHA for 5fbd21e

Commits on Apr 3, 2018

• cholesky_solve_with_broadcast, matrix_solve_with_broadcast and matrix_triangular_solve_with_broadcast added to linear_operator_util.py

  Show description for 5d1086a
  

  langmore

  authored and

  tensorflower-gardener

  committed
  5d1086a

  Copy full SHA for 5d1086a

Commits on Mar 30, 2018

• BUGFIX: sample_stats.percentile: Fractions for interpolation were done in

  Show description for 9d8201c
  

  langmore

  authored and

  tensorflower-gardener

  committed
  9d8201c

  Copy full SHA for 9d8201c

Pagination

Previous

Next