Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Dec 2, 2021

• Add tfp.math.minimize_stateless as a purely-functional analogue of tfp.math.minimize.

  Show description for e895795
  

  davmre

  authored and

  tensorflower-gardener

  committed
  e895795

• Ensure owens_t special function gracefully handles partially dynamic shapes.

  Show description for c2299ff
  

  csuter

  authored and

  tensorflower-gardener

  committed
  c2299ff

Commits on Dec 1, 2021

• Mark initialization_test as size medium.

  Show description for b5f0348
  

  ursk

  authored and

  tensorflower-gardener

  committed
  b5f0348

• Enhance precision of truncated normal for large truncate values

  Show description for 224dd2c
  

  fehiepsi

  authored and

  tensorflower-gardener

  committed
  224dd2c

Commits on Nov 30, 2021

• Fix a bug in hager_zhang when some batch members fail to bracket.

  Show description for d5956d6
  

  jburnim

  authored and

  tensorflower-gardener

  committed
  d5956d6

• Always return Kalman filter results as a FilterResults namedtuple.

  Show description for f5541be
  

  davmre

  authored and

  tensorflower-gardener

  committed
  f5541be

Commits on Nov 29, 2021

• Improves error message in dtype_util.common_dtype.

  Show description for 9990d61
  

  jburnim

  authored and

  tensorflower-gardener

  committed
  9990d61

Commits on Nov 23, 2021

• Hack convert_to_nested_tensor to forward Keras tensors untouched.

  Show description for ae7117f
  

  axch

  authored and

  tensorflower-gardener

  committed
  ae7117f

• Add SNAPERHamiltonianMonteCarlo.

  Show description for 5f4d7b9
  

  SiegeLordEx

  authored and

  tensorflower-gardener

  committed
  5f4d7b9

Commits on Nov 22, 2021

• Restrain Hypothesis from generating absurd limit arguments to TruncatedNormal.

  Show description for b305c9c
  

  axch

  authored and

  tensorflower-gardener

  committed
  b305c9c

Commits on Nov 20, 2021

• Add SNAPER-HMC colab for the paper.

  Show description for 8ac3ac1
  

  SiegeLordEx

  authored and

  tensorflower-gardener

  committed
  8ac3ac1

Commits on Nov 19, 2021

• Re-narrow numerical tolerances perturbed by Hypothesis upgrade from 3.x.

  Show description for 16c382e
  

  axch

  authored and

  tensorflower-gardener

  committed
  16c382e

Commits on Nov 18, 2021

• Relax precision of truncated normal in hypothesis test.

  Show description for 8e03d0c
  

  fehiepsi

  authored and

  tensorflower-gardener

  committed
  8e03d0c

• Enable/test the use of pmax under TensorFlow in some MCMC algorithms.

  Show description for 40b1807
  

  SiegeLordEx

  authored and

  tensorflower-gardener

  committed
  40b1807

• Add an all_gather-based fallback for TF implementation of pmin/pmax.

  Show description for 059d789
  

  SiegeLordEx

  authored and

  tensorflower-gardener

  committed
  059d789

Commits on Nov 17, 2021

• Increment TF version requirement to 2.8 (nightly).

  Show description for c73714b
  

  emilyfertig

  authored and

  tensorflower-gardener

  committed
  c73714b

• Increment minor version number; TFP 0.15 release is in progress.

  Show description for 563ae8a
  

  emilyfertig

  authored and

  tensorflower-gardener

  committed
  563ae8a

Commits on Nov 16, 2021

• Replace value in simple_step_size_adaptation_test to work around a test failure caused by different numerics in Numpy 1.21.4/Python 3.9.

  Show description for bb15d65
  

  emilyfertig

  authored and

  tensorflower-gardener

  committed
  bb15d65

• Fix some documentation typos.

  Show description for fe99538
  

  axch

  authored and

  tensorflower-gardener

  committed
  fe99538

• Add SNAPER criterion.

  Show description for 058b3f2
  

  SiegeLordEx

  authored and

  tensorflower-gardener

  committed
  058b3f2

• A reproduction for b/206660667.

  Show description for f5a370f
  

  axch

  authored and

  tensorflower-gardener

  committed
  f5a370f

Commits on Nov 15, 2021

• Fix NaNs when running the spike-and-slab sampler under XLA.

  Show description for 76eaa7e
  

  davmre

  authored and

  tensorflower-gardener

  committed
  76eaa7e

Commits on Nov 12, 2021

• Set Softplus.dtype when hinge_softness or low is specified.

  Show description for 8870950
  

  jburnim

  authored and

  tensorflower-gardener

  committed
  8870950

• Simplify DeferredModule interface by combining base_class and args_fn into build_fn.

  Show description for 2ad13b1
  

  davmre

  authored and

  tensorflower-gardener

  committed
  2ad13b1

Commits on Nov 11, 2021

• Centralize generic loop utilities under internal/loop_util.

  Show description for 6e8ef0e
  

  davmre

  authored and

  tensorflower-gardener

  committed
  6e8ef0e

Commits on Nov 10, 2021

• Support sparse regression with spike-and-slab priors in the STS Gibbs sampler.

  Show description for 8944243
  

  davmre

  authored and

  tensorflower-gardener

  committed
  8944243

Commits on Nov 9, 2021

• Use float64 instead of float32 in tfd.Beta and tfd.Binomial tests to work around a breakage in OSS following scipy 1.7.2 release. Unpin the scipy version.

  Show description for cd35db9
  

  emilyfertig

  authored and

  tensorflower-gardener

  committed
  cd35db9

• Merge JointDistribution mixins (vmap and sample-path) into the base class.

  Show description for ed5f42e
  

  davmre

  authored and

  tensorflower-gardener

  committed
  ed5f42e

• Add a spike and slab sampler for sparse regression models.

  Show description for 7fa7441
  

  davmre

  authored and

  tensorflower-gardener

  committed
  7fa7441

• Pin scipy version in OSS tests to avoid breakages following the release of scipy 1.7.2.

  Show description for c572a77
  

  emilyfertig

  authored and

  tensorflower-gardener

  committed
  c572a77

Commits on Nov 5, 2021

• Documentation tyop.

  Show description for 60e60da
  

  axch

  authored and

  tensorflower-gardener

  committed
  60e60da

• Nitpick documentation of reproducibility and add a few reproducibility tests.

  Show description for caf14d1
  

  axch

  authored and

  tensorflower-gardener

  committed
  caf14d1

Commits on Nov 4, 2021

• Update code in preparation to upgrade hypothesis to a newer version.

  Show description for 3e5a446
  

  yilei

  authored and

  tensorflower-gardener

  committed
  3e5a446

• Allow EMG to disable parameter checking when validate_args=False.

  Show description for 9e09c66
  

  jereliu

  authored and

  tensorflower-gardener

  committed
  9e09c66

• Bump tolerances of two tests to fix the TFP build.

  Show description for 57ccb0f
  

  emilyfertig

  authored and

  tensorflower-gardener

  committed
  57ccb0f

Pagination

PreviousNext