vishwakftw
20332b2604
LAPACK 3.6 and below requires more workspace that the newer versions
2019-01-11 18:49:03 +05:30
vishwakftw
e21cb1a99f
Add complex support for SVD
2019-01-10 23:13:56 +05:30
Peter Hawkins
9e91fb42cb
Make LU JVP work for rectangular matrices.
...
Fix bug in BLAS TRSM kernel if left_side=False.
2019-01-09 15:52:34 -05:00
vishwakftw
170e5ab0a3
Minor nits
2019-01-08 23:20:16 +05:30
vishwakftw
c509d48b08
Address comments
...
- Fix acronym
- Include test for compute_uv=False
- Remove unnecessary imports
2019-01-08 21:48:15 +05:30
vishwakftw
ff6d110975
Added tests for SVD, and fixed some errors in computation
2019-01-08 15:51:30 +05:30
vishwakftw
e16f31e94f
Merge branch 'master' of https://github.com/google/jax into svd
2019-01-08 09:30:05 +05:30
vishwakftw
954b047bea
Address review comments
2019-01-08 09:24:48 +05:30
Peter Hawkins
7a114c8f6a
Remove unused malloc/free imports.
2019-01-07 19:20:42 -05:00
Peter Hawkins
1c2cff15d2
Finish implementation of symmetric eigendecomposition on CPU:
...
* add test case.
* add double and complex64 implementations.
Also add logic to all linalg methods to both coerce arguments to arrays, and to promote to an inexact (float or complex) type if the argument is not inexact.
2019-01-07 19:20:42 -05:00
Peter Hawkins
d3348b9cdd
Start work on an implementation of np.linalg.eigh.
2019-01-07 19:20:42 -05:00
vishwakftw
484db1e15f
Add SVD for float and double types
2019-01-05 11:13:08 +05:30
Peter Hawkins
06135fa6f5
Implement numpy.linalg.solve and scipy.linalg.solve.
...
Make Cholesky and TriangularSolve work for complex numbers on CPU. The HLO implementations are broken for complex numbers on GPU/TPU, so no tests enabled for these yet.
2018-12-21 16:29:45 -05:00
Peter Hawkins
a4386457e2
Fix test failures due to type mismatches in linear algebra tests.
...
Minor code cleanups.
2018-12-21 15:18:34 -05:00
Peter Hawkins
b68c93d37f
Implement np.linalg.slogdet.
...
Change implementation of np.linalg.logdet to call np.linalg.slogdet.
Add support for complex64 LU decomposition.
2018-12-20 22:18:20 -05:00
Peter Hawkins
dfdc2e3806
Add LU decomposition implementation backed by LAPACK on the CPU platform.
...
Implement np.linalg.det, and scipy.linalg.{lu,lu_factor,det}.
Add missing abstractification to loop arguments.
Implement XLA abstractification rules for AbstractTuple, ConcreteArray, and ShapedArray.
2018-12-20 18:45:34 -05:00
Peter Hawkins
20935448b9
Whitespace fix to jaxlib/BUILD.
2018-12-19 14:51:46 -05:00
Peter Hawkins
eac96ac239
Fix bugs with bool argument passing; pass PRED values as int32s instead.
2018-12-17 17:17:47 -05:00
Peter Hawkins
0333a98bab
Add triangular solve BLAS implementation.
2018-12-17 16:39:19 -05:00
Peter Hawkins
e630fda9ee
Add missing copyright notice.
...
Remove unused BUILD load() statement.
2018-12-17 16:30:27 -05:00
Peter Hawkins
3c388b98f1
Add support for calling LAPACK primitives from SciPy from JAX linalg.
2018-12-17 16:30:27 -05:00
