ROCm/jax

Pull Requests Commits

Merge pull request #158 from hawkinsp/master

hawkinsp committed 7 years ago

Verified e05ac8d5

Merge pull request #155 from rsepassi/tfds

alexbw committed 7 years ago

Verified 58464136

Merge pull request #161 from google/add-jaxvals-batching

mattjj committed 7 years ago

Verified 58364f03

cover unimplemented add_jaxvals_p batching case

mattjj committed 7 years ago

43f2e2a7

Merge pull request #159 from hawkinsp/lr

alexbw committed 7 years ago

Verified 179f54c6

Lower learning rate in NN example notebook to fix divergence.

hawkinsp committed 7 years ago

12a36f4c

Merge pull request #1 from hawkinsp/learningrate

hawkinsp committed 7 years ago

Verified 484022bb

Lower learning rate in NN notebook to fix divergence

hawkinsp committed 7 years ago

Verified 9888b33d

Implement np.kron.

hawkinsp committed 7 years ago

0a75007c

Merge pull request #157 from hawkinsp/master

hawkinsp committed 7 years ago

Verified aae556bf

Implement np.{isposinf,isneginf,nansum,nanprod,nanmin,nanmax,nan_to_num}.

hawkinsp committed 7 years ago

bd50c5b6

Merge pull request #156 from hawkinsp/master

hawkinsp committed 7 years ago

Verified c8c8fa1a

Implement np.linalg.slogdet.

hawkinsp committed 7 years ago

b68c93d3

Merge pull request #154 from hawkinsp/master

hawkinsp committed 7 years ago

Verified 3db941f1

Add TensorFlow Datasets data-loading notebook

Ryan Sepassi committed 7 years ago

0df8a8e5

Be more defensive about old jaxlib versions and non-CPU devices in LU decomposition usage.

hawkinsp committed 7 years ago

1815f17a

Add LU decomposition implementation backed by LAPACK on the CPU platform.

hawkinsp committed 7 years ago

dfdc2e38

Merge pull request #153 from hawkinsp/numpy

hawkinsp committed 7 years ago

Verified 5cf642e3

Implement np.{diagonal,count_nonzero}.

hawkinsp committed 7 years ago

4fc73205

Merge pull request #151 from google/value-and-grad

mattjj committed 7 years ago

Verified 4124932b

add simple value_and_grad test

mattjj committed 7 years ago

565c9558

add value-and-grad fun (closes #149)

mattjj committed 7 years ago

8f1bc997

Merge pull request #150 from hawkinsp/master

hawkinsp committed 7 years ago

Verified 21e5eb13

Use lax.bitwise_ in isnan().

hawkinsp committed 7 years ago

bf73c1e2

Implement np.{full_like,isinf,isnan}. Fix np.isfinite.

hawkinsp committed 7 years ago

7f6119c7

Add implementations of np.{meshgrid,linspace,logspace,geomspace,diag_indices} that forward to the usual numpy implementation.

hawkinsp committed 7 years ago

95135377

lower tolerance on triangular-solve test for x64 mode

froystig committed 7 years ago

72bc8829

Merge pull request #147 from google/einsum

mattjj committed 7 years ago

Verified 449da4cd

disable test, py3 opt_einsum nondeterministic bug?

mattjj committed 7 years ago

6bb9609f

fix several einsum bugs

mattjj committed 7 years ago

6a138202

Newer Older