| 41417ee | jax authors | 11 October 2022, 18:12:26 UTC | Merge pull request #12752 from skye:workspace PiperOrigin-RevId: 480405935 | 11 October 2022, 18:12:26 UTC |
| 63be220 | Skye Wanderman-Milne | 11 October 2022, 17:55:05 UTC | Update WORKSPACE and setup.py for jaxlib 0.3.22 release | 11 October 2022, 17:55:05 UTC |
| af98d8b | jax authors | 11 October 2022, 16:21:28 UTC | Merge pull request #12746 from jakevdp:fix-changelog PiperOrigin-RevId: 480375907 | 11 October 2022, 16:21:28 UTC |
| 05faf0f | Jake VanderPlas | 11 October 2022, 15:15:47 UTC | Remove deprecated functionality from jax.test_util PiperOrigin-RevId: 480360504 | 11 October 2022, 15:16:34 UTC |
| 9c99498 | Adam Paszke | 11 October 2022, 13:46:59 UTC | Support MANUAL collectives in top-level xmaps It's a bit of a weird use-case, since MANUAL mode is meant for xmaps that are nested inside pjits, but it doesn't hurt us to support it. PiperOrigin-RevId: 480342531 | 11 October 2022, 13:47:30 UTC |
| bf78697 | jax authors | 11 October 2022, 13:34:13 UTC | Merge pull request #12727 from jakevdp:scipy-linalg-types PiperOrigin-RevId: 480340339 | 11 October 2022, 13:34:13 UTC |
| 9431294 | Jake VanderPlas | 11 October 2022, 13:24:38 UTC | changelog: add missing github commit links | 11 October 2022, 13:24:38 UTC |
| e45df46 | Rishabh Kabra | 11 October 2022, 11:40:36 UTC | Clarify docs for `fori_loop`, noting that negative or custom increments are not supported. PiperOrigin-RevId: 480317277 | 11 October 2022, 11:41:10 UTC |
| eade368 | jax authors | 11 October 2022, 11:01:12 UTC | Merge pull request #12324 from gnecula:tf_pjit PiperOrigin-RevId: 480310826 | 11 October 2022, 11:01:12 UTC |
| 9c879ad | George Necula | 06 September 2022, 06:32:45 UTC | [jax2tf] Implement jax2tf(pjit) for experimental_native_lowering This implementation is for the case jax2tf.convert(pjit(f_jax)), that is, the `pjit` appears at the top-level of the function to be lowered. | 11 October 2022, 07:45:07 UTC |
| ff17d3d | Yash Katariya | 11 October 2022, 05:08:06 UTC | Add support for calculating the device_assignment when there are no inputs to `jit` and `pjit`. Also look at the shardings inside the jaxpr for `sharding_constraint_p` and `pjit_p` primitives since with `jax.Array`, each `with_sharding_constraint`/`pjit` inside a computation can contain a different sharding (so we need to check if the device_assignment is the same). Also the output is `committed` if there are jaxpr shardings inside the computation via `with_sharding_constraint`/`pjit` or if any of the inputs are committed or `output_sharding` is specified. Co-authored-by: Matthew Johnson <mattjj@google.com> PiperOrigin-RevId: 480256796 | 11 October 2022, 05:08:42 UTC |
| df5f7cb | Matthew Johnson | 11 October 2022, 01:51:04 UTC | Rolling forward https://github.com/google/jax/pull/12707 after rollback, due to changes in relatively trivial jax.numpy shape validation code failed in some downstream user tests. PiperOrigin-RevId: 480229237 | 11 October 2022, 01:51:37 UTC |
| 9b3e864 | Yash Katariya | 11 October 2022, 01:10:46 UTC | Add weak_type attribute to `Array` since it exists on DA (but doesn't exist on SDA). PiperOrigin-RevId: 480223116 | 11 October 2022, 01:11:11 UTC |
| afe74b4 | Jake VanderPlas | 10 October 2022, 23:54:29 UTC | [typing] add type annotations to jax.scipy.linalg | 10 October 2022, 23:54:29 UTC |
| ec5b1c9 | Kuangyuan Chen | 10 October 2022, 22:00:28 UTC | Turn on cpp pjit py default PiperOrigin-RevId: 480185387 | 10 October 2022, 22:01:04 UTC |
| 76d8c08 | Yash Katariya | 10 October 2022, 21:44:28 UTC | Fix the type annotation of return type of `device_buffer` and `device_buffers` which return `ArrayImpl` instead of DeviceArray. PiperOrigin-RevId: 480181798 | 10 October 2022, 21:45:12 UTC |
| a3a2206 | Peter Hawkins | 10 October 2022, 21:24:17 UTC | Fix compilation failure in lapack kernel under msan. a_size wasn't defined, but it would only be caught under memory sanitizer. PiperOrigin-RevId: 480176934 | 10 October 2022, 21:24:59 UTC |
| 2246887 | Peter Hawkins | 10 October 2022, 19:53:19 UTC | Add input-output aliasing annotations for LAPACK calls on CPU. PiperOrigin-RevId: 480156067 | 10 October 2022, 19:57:29 UTC |
| 752c3ff | Yash Katariya | 10 October 2022, 19:50:41 UTC | Lift `lambda x: x` to the top level so that we don't recompile on every invocation of `process_allgather`. PiperOrigin-RevId: 480155482 | 10 October 2022, 19:51:17 UTC |
| 90e9abe | jax authors | 10 October 2022, 19:22:52 UTC | Merge pull request #12722 from hawkinsp:tests PiperOrigin-RevId: 480149233 | 10 October 2022, 19:22:52 UTC |
| 2ba0396 | Peter Hawkins | 10 October 2022, 19:11:58 UTC | Add changes accidentally omitted from https://github.com/google/jax/pull/12717 | 10 October 2022, 19:11:58 UTC |
| 34eb6ce | Tianjian Lu | 10 October 2022, 18:53:45 UTC | [sparse] BCSR fromdense and todense. PiperOrigin-RevId: 480141918 | 10 October 2022, 18:54:22 UTC |
| c657449 | Peter Hawkins | 10 October 2022, 18:34:43 UTC | Copybara import of the project: -- d39bdefb33a19e407c352df27fb04127f4fe8a1d by Peter Hawkins <phawkins@google.com>: Migrate more tests from jtu.cases_from_list to jtu.sample_product. COPYBARA_INTEGRATE_REVIEW=https://github.com/google/jax/pull/12717 from hawkinsp:sampletest d39bdefb33a19e407c352df27fb04127f4fe8a1d PiperOrigin-RevId: 480136538 | 10 October 2022, 18:35:32 UTC |
| 22cd505 | Peter Hawkins | 10 October 2022, 18:27:51 UTC | Reapply: Use input-output aliasing for jaxlib GPU custom calls. Previously we had no way to tell XLA that inputs and outputs of GPU custom calls must alias. This now works in XLA:GPU so we can just ask XLA to enforce the aliasing we need. It turns out some users are relying on the API contract of the custom calls within serialized HLO remaining stable. For the moment, we reapply only the Python changes. The C++ code is already tolerant of both aliased and unaliased outputs, and this gets us all the benefit of saving a copy. We can break backwards compatibility on the serialized HLO after users upgrade their saved HLO to the aliased version. PiperOrigin-RevId: 480134780 | 10 October 2022, 18:29:18 UTC |
| 707b07c | jax authors | 10 October 2022, 18:22:35 UTC | Merge pull request #12697 from jakevdp:lax-slicing-types PiperOrigin-RevId: 480131675 | 10 October 2022, 18:22:35 UTC |
| ad4dcc4 | jax authors | 10 October 2022, 18:15:46 UTC | Merge pull request #12700 from jakevdp:scipy-types PiperOrigin-RevId: 480131410 | 10 October 2022, 18:15:46 UTC |
| 124021d | Jake VanderPlas | 10 October 2022, 16:11:13 UTC | [typing] add annotations to jax.scipy.ndimage | 10 October 2022, 16:11:13 UTC |
| 76e4a1d | Jake VanderPlas | 09 October 2022, 12:18:45 UTC | [typing] add annotations to jax.scipy.fft | 09 October 2022, 12:18:45 UTC |
| ae9f8ee | Jake VanderPlas | 09 October 2022, 11:20:46 UTC | [typing] annotate lax.slicing | 09 October 2022, 11:20:46 UTC |
| 9cabd22 | jax authors | 09 October 2022, 08:25:18 UTC | Copybara import of the project: -- 6d2aaac2454117d54997243714c1a009827707ca by Matthew Johnson <mattjj@google.com>: implement bint arrays (opaque dtypes), add padding rules Co-authored-by: Sharad Vikram <sharad.vikram@gmail.com> PiperOrigin-RevId: 479883102 | 09 October 2022, 08:25:50 UTC |
| 25c6ef7 | jax authors | 09 October 2022, 07:23:39 UTC | Merge pull request #12707 from mattjj:djax-slice-sick4 PiperOrigin-RevId: 479876971 | 09 October 2022, 07:23:39 UTC |
| 6d2aaac | Matthew Johnson | 04 October 2022, 21:08:43 UTC | implement bint arrays (opaque dtypes), add padding rules Co-authored-by: Sharad Vikram <sharad.vikram@gmail.com> | 09 October 2022, 05:57:29 UTC |
| 75b2d05 | Yash Katariya | 09 October 2022, 02:23:32 UTC | Make `is_fully_replicated` and `is_fully_addressble` a property rather than a method. Why? 1. Because it's easy to cache a property than a method with only the `self` argument. (See below for article) 2. There's no harm in making them a property because both of them return a bool without any side-effects and are cached (so its fast). Why cache `is_fully_addressable`? Because its very expensive to calculate when you have 1000s of devices. PiperOrigin-RevId: 479850850 | 09 October 2022, 02:24:12 UTC |
| 674038c | jax authors | 08 October 2022, 18:39:05 UTC | Merge pull request #12705 from mattjj:fix-prng-key-array-device-put PiperOrigin-RevId: 479813689 | 08 October 2022, 18:39:05 UTC |
| 0a0f492 | Matthew Johnson | 07 October 2022, 23:48:34 UTC | make device_put(prngkeyarray, sharding) for Array Co-authored-by: Yash Katariya <yashkatariya@google.com> Co-authored-by: Roy Frostig <frostig@google.com> | 07 October 2022, 23:50:16 UTC |
| 2693afa | Peter Hawkins | 07 October 2022, 21:35:41 UTC | Revert: Use input-output aliasing for jaxlib GPU custom calls. Previously we had no way to tell XLA that inputs and outputs of GPU custom calls must alias. This now works in XLA:GPU so we can just ask XLA to enforce the aliasing we need. This seems to be causing some test failures downstream, so reverting this for the moment until I can debug them. PiperOrigin-RevId: 479670565 | 07 October 2022, 21:36:24 UTC |
| e8ba61d | jax authors | 07 October 2022, 21:28:51 UTC | Merge pull request #12677 from mattjj:jit-pjit-lower-sharding PiperOrigin-RevId: 479669125 | 07 October 2022, 21:28:51 UTC |
| 58cd837 | jax authors | 07 October 2022, 20:49:57 UTC | Merge pull request #12675 from mattjj:device-put2 PiperOrigin-RevId: 479660808 | 07 October 2022, 20:49:57 UTC |
| 33f18fe | jax authors | 07 October 2022, 20:12:57 UTC | Merge pull request #12703 from ROCmSoftwarePlatform:rocm-ci-update2 PiperOrigin-RevId: 479652964 | 07 October 2022, 20:12:57 UTC |
| 34f6646 | Jason Furmanek | 07 October 2022, 19:57:53 UTC | Add default setting for TENSORFLOW_ROCM_COMMIT | 07 October 2022, 19:57:53 UTC |
| 93b839a | Peter Hawkins | 07 October 2022, 19:21:36 UTC | Use input-output aliasing for jaxlib GPU custom calls. Previously we had no way to tell XLA that inputs and outputs of GPU custom calls must alias. This now works in XLA:GPU so we can just ask XLA to enforce the aliasing we need. PiperOrigin-RevId: 479642543 | 07 October 2022, 19:22:04 UTC |
| 0995f79 | jax authors | 07 October 2022, 17:29:18 UTC | Merge pull request #12698 from jakevdp:scipy-fft-todo PiperOrigin-RevId: 479615134 | 07 October 2022, 17:29:18 UTC |
| 730f993 | Jake VanderPlas | 07 October 2022, 16:57:39 UTC | Update scipy version in jax.scipy.fft | 07 October 2022, 16:57:39 UTC |
| fde444e | jax authors | 07 October 2022, 16:38:45 UTC | Merge pull request #12695 from gnecula:tf_dimension PiperOrigin-RevId: 479602588 | 07 October 2022, 16:38:45 UTC |
| fb2141f | George Necula | 07 October 2022, 06:45:12 UTC | [jax2tf] Allow the use of DimPolynomial with jnp.array and binary operations Prior to this the user had to explicitly call core.dimension_as_value whenever using a potentially polymorphic shape in the computation, e.g., x + core.dimension_as_value(x.shape[0]). Furthermore, jnp.array(x.shape[0]) would fail. Now, these operations are allowed implicitly, and the user can call `jnp.array(x.shape[0])`. This uses an internal extensibility mechanism called __jax_array__ that is experimental and probably not fully implemented. | 07 October 2022, 14:58:41 UTC |
| 363cc12 | jax authors | 07 October 2022, 13:36:11 UTC | Merge pull request #12197 from ROCmSoftwarePlatform:fixedRocmUnitTestsSkip PiperOrigin-RevId: 479566021 | 07 October 2022, 13:36:11 UTC |
| 7c7c94c | George Necula | 07 October 2022, 11:13:20 UTC | Expand support for __jax_array__ in jnp.array. This relates to the long discussion in #4725 and #10065. | 07 October 2022, 11:25:07 UTC |
| 6c70e4d | jax authors | 07 October 2022, 07:10:18 UTC | Merge pull request #12691 from mattjj:issue12688 PiperOrigin-RevId: 479507232 | 07 October 2022, 07:10:18 UTC |
| 076a734 | Matthew Johnson | 07 October 2022, 06:15:22 UTC | fix -O / PYTHONOPTIMIZE bug fixes #12688 I'm not sure how to write test cases for PYTHONOPTIMIZE=1 (without growing our whole test matrix), so I'm leaving this untested... | 07 October 2022, 06:15:22 UTC |
| 7a82536 | Tianjian Lu | 07 October 2022, 00:28:16 UTC | [sparse] Bug fix in _validate_bcsr. PiperOrigin-RevId: 479452053 | 07 October 2022, 00:28:52 UTC |
| f465aee | jax authors | 06 October 2022, 23:48:45 UTC | Merge pull request #12673 from jakevdp:lax-linalg-types PiperOrigin-RevId: 479444281 | 06 October 2022, 23:48:45 UTC |
| bcca6fb | Matthew Johnson | 06 October 2022, 23:12:20 UTC | add test, small fixes Co-authored-by: Yash Katariya <yashkatariya@google.com> | 06 October 2022, 23:45:34 UTC |
| 9104536 | Jake VanderPlas | 06 October 2022, 17:46:12 UTC | [typing] add type annotations to lax.linalg functions | 06 October 2022, 23:19:00 UTC |
| ce95eba | Matthew Johnson | 05 October 2022, 22:17:29 UTC | make device_put work with Sharding 2nd arg Co-authored-by: Yash Katariya <yashkatariya@google.com> | 06 October 2022, 23:14:15 UTC |
| f49d3d4 | Parker Schuh | 06 October 2022, 22:13:58 UTC | Rename Executable to LoadedExecutable within jax. PiperOrigin-RevId: 479423951 | 06 October 2022, 22:14:33 UTC |
| df4f399 | jax authors | 06 October 2022, 21:38:13 UTC | Merge pull request #12678 from jakevdp:average-axis-tuple PiperOrigin-RevId: 479414002 | 06 October 2022, 21:38:13 UTC |
| ba2d803 | jax authors | 06 October 2022, 21:31:33 UTC | Merge pull request #12674 from jakevdp:svd-type PiperOrigin-RevId: 479413950 | 06 October 2022, 21:31:33 UTC |
| 1f3048b | jax authors | 06 October 2022, 18:33:41 UTC | Merge pull request #12573 from ROCmSoftwarePlatform:rocm-ci-update PiperOrigin-RevId: 479369638 | 06 October 2022, 18:33:41 UTC |
| 824903b | jax authors | 06 October 2022, 17:51:56 UTC | Merge pull request #12686 from hawkinsp:sample PiperOrigin-RevId: 479354990 | 06 October 2022, 17:51:56 UTC |
| 8107e36 | Peter Hawkins | 06 October 2022, 15:53:18 UTC | Switch lax_numpy_indexing_test to use jtu.sample_product. | 06 October 2022, 17:44:17 UTC |
| d174b3d | Yash Katariya | 06 October 2022, 17:20:14 UTC | Take `shardings` as a parameter to `deserialize` and `run_deserialization` instead of `mesh` and `pspecs`. PiperOrigin-RevId: 479346552 | 06 October 2022, 17:20:49 UTC |
| 32ef3ba | Jake VanderPlas | 06 October 2022, 17:20:46 UTC | jnp.average: support tuple axis | 06 October 2022, 17:20:46 UTC |
| d94327c | Jake VanderPlas | 06 October 2022, 17:20:26 UTC | Move promote_like_jnp to jax.test_util | 06 October 2022, 17:20:26 UTC |
| 219c574 | jax authors | 06 October 2022, 15:15:22 UTC | Merge pull request #12676 from jakevdp:fix-concat-test PiperOrigin-RevId: 479318171 | 06 October 2022, 15:15:22 UTC |
| 1b5e9e4 | Roy Frostig | 06 October 2022, 00:30:56 UTC | add input/output sharding to executable protocol | 06 October 2022, 00:30:56 UTC |
| d65145f | jax authors | 06 October 2022, 00:23:30 UTC | Merge pull request #12532 from jakevdp:reduction-dtype PiperOrigin-RevId: 479180617 | 06 October 2022, 00:23:30 UTC |
| 3f08f85 | jax authors | 06 October 2022, 00:17:06 UTC | Merge pull request #12666 from jakevdp:polynomial-typing PiperOrigin-RevId: 479180604 | 06 October 2022, 00:17:06 UTC |
| 2525aa6 | jax authors | 05 October 2022, 23:54:11 UTC | Merge pull request #12642 from RissyRan:platform PiperOrigin-RevId: 479176113 | 05 October 2022, 23:54:11 UTC |
| e8dc6d1 | Matthew Johnson | 05 October 2022, 22:47:59 UTC | improve jit(f).lower(duck_args) and pjit(f).lower(duck_args) Co-authored-by: Yash Katariya <yashkatariya@google.com> | 05 October 2022, 22:47:59 UTC |
| ff08109 | Jake VanderPlas | 05 October 2022, 22:29:15 UTC | test: fix LaxNumpyTest:testConcatenate | 05 October 2022, 22:29:15 UTC |
| 2de91d2 | Artem Belevich | 05 October 2022, 21:47:52 UTC | Handle FP8 types. PiperOrigin-RevId: 479148993 | 05 October 2022, 21:48:30 UTC |
| c1328a6 | Jake VanderPlas | 05 October 2022, 21:29:00 UTC | [typing] overloads for jnp.linalg.svd & jnp.linalg.qr | 05 October 2022, 21:29:00 UTC |
| 9d12e21 | Yash Katariya | 05 October 2022, 19:34:43 UTC | Add `addressable_shards` to SDA and DA as a compatibility API to match with `jax.Array`. This will aid in transition to `jax.Array`. PiperOrigin-RevId: 479115126 | 05 October 2022, 19:35:17 UTC |
| 167004d | jax authors | 05 October 2022, 19:00:10 UTC | Merge pull request #12654 from jakevdp:fft-typing PiperOrigin-RevId: 479106034 | 05 October 2022, 19:00:10 UTC |
| 3ec5e45 | jax authors | 05 October 2022, 18:33:55 UTC | Merge pull request #12660 from hawkinsp:testing PiperOrigin-RevId: 479098519 | 05 October 2022, 18:33:55 UTC |
| 6a348f9 | Jake VanderPlas | 05 October 2022, 18:23:45 UTC | [typing] add types for jax.numpy.polynomial | 05 October 2022, 18:23:45 UTC |
| f3ded0f | Ran Ran | 05 October 2022, 18:16:49 UTC | Address comments for change log | 05 October 2022, 18:16:49 UTC |
| dd0a455 | Yash Katariya | 05 October 2022, 16:58:04 UTC | Add sharding to `DeviceArray` and `ShardedDeviceArray` as a compatibility change to rollout `jax.Array`. Also expose `device_replica_id_map` since that is important API for checkpointing to find all the unique shards of an Array. You can also use this to calculate the unique indices in a sharding (which is what `sda.one_replica_buffer_indices` does) PiperOrigin-RevId: 479072520 | 05 October 2022, 16:58:41 UTC |
| 4da72cf | Yash Katariya | 05 October 2022, 14:01:26 UTC | Add `host_local_array_to_global_array` and `global_array_to_host_local_array` for enabling transition to jax.Array. Also support `FROM_GDA` for `jax.Array` as a backwards compatible change so that users can continue to use that until they transition to jax.Array. Its currently required because of usage like `in_axis_resources = (FROM_GDA, FROM_GDA, P('data'), None)` and changing this on users side will require input from users so we as JAX can just support it as a temporary thing since GDA and Array act similarly in pjit. PiperOrigin-RevId: 479035326 | 05 October 2022, 14:01:58 UTC |
| 2c946b3 | Peter Hawkins | 05 October 2022, 01:52:41 UTC | Migrate api_test, lax_numpy_test, and lax_vmap_test to jtu.sample_product. Gives a ~2x improvement in pytest --collect-only timing for lax_numpy_test. | 05 October 2022, 13:46:19 UTC |
| 4870fd3 | Ran Ran | 05 October 2022, 01:01:31 UTC | Update message and change log | 05 October 2022, 04:39:04 UTC |
| 22e1547 | Ran Ran | 04 October 2022, 23:53:07 UTC | Update set up message | 05 October 2022, 04:39:04 UTC |
| ef55953 | Ran Ran | 04 October 2022, 23:37:37 UTC | Add set up message for JAX_PLATFORMS | 05 October 2022, 04:39:04 UTC |
| 78a7e16 | Ran Ran | 03 October 2022, 22:52:03 UTC | Set JAX_PLATFORMS=tpu,cpu on TPUs | 05 October 2022, 04:39:04 UTC |
| 015a12e | jax authors | 05 October 2022, 02:05:44 UTC | Merge pull request #12657 from jakevdp:linalg-typing PiperOrigin-RevId: 478937233 | 05 October 2022, 02:05:44 UTC |
| 9646199 | Jake VanderPlas | 04 October 2022, 22:52:52 UTC | [typing] add annotations to numpy.fft | 04 October 2022, 22:52:54 UTC |
| 78ed03c | Jake VanderPlas | 04 October 2022, 22:50:29 UTC | [typing] add annotations to jax.numpy.linalg | 04 October 2022, 22:50:29 UTC |
| 1206e10 | jax authors | 04 October 2022, 22:09:49 UTC | Merge pull request #12658 from sharadmv:fix-profiler PiperOrigin-RevId: 478893609 | 04 October 2022, 22:09:49 UTC |
| 83ef7d0 | Sharad Vikram | 04 October 2022, 21:31:17 UTC | Fix collect_profile _src import | 04 October 2022, 21:31:17 UTC |
| 2b0542d | jax authors | 04 October 2022, 19:51:04 UTC | Merge pull request #12640 from mattjj:djax-slice-sick PiperOrigin-RevId: 478860162 | 04 October 2022, 19:51:04 UTC |
| 06a2c85 | Matthew Johnson | 29 September 2022, 21:00:47 UTC | [dynamic-shapes] small fix to einsum (and indexing) | 04 October 2022, 19:29:41 UTC |
| a60ca9f | Sharad Vikram | 04 October 2022, 19:14:13 UTC | Test that array layout is preserved in Python callbacks PiperOrigin-RevId: 478852392 | 04 October 2022, 19:14:47 UTC |
| 3f663e4 | jax authors | 04 October 2022, 19:07:38 UTC | Merge pull request #12639 from skye:version PiperOrigin-RevId: 478850508 | 04 October 2022, 19:07:38 UTC |
| 62774f5 | jax authors | 04 October 2022, 17:53:12 UTC | Merge pull request #12641 from jakevdp:asarray-type PiperOrigin-RevId: 478829647 | 04 October 2022, 17:53:12 UTC |
| 9e0b8ca | jax authors | 04 October 2022, 17:47:01 UTC | Merge pull request #12653 from jakevdp:jacobian-wraps PiperOrigin-RevId: 478829535 | 04 October 2022, 17:47:01 UTC |
| 0d93679 | Jake VanderPlas | 04 October 2022, 17:21:54 UTC | jax.jacobian: propagate function signature to transformed function | 04 October 2022, 17:21:54 UTC |
| 3be2087 | Jake VanderPlas | 30 September 2022, 22:39:10 UTC | jnp.prod & jnp.sum: promote to default integer type rather than int64/uint64 | 04 October 2022, 17:08:30 UTC |
| 069866e | Jake VanderPlas | 03 October 2022, 22:55:56 UTC | Add types to jax/_src/numpy/util.py | 04 October 2022, 17:07:38 UTC |
| ae49d2e | Tianjian Lu | 04 October 2022, 16:59:26 UTC | [sparse] Add conversions between BCSR and BCOO. PiperOrigin-RevId: 478816413 | 04 October 2022, 17:00:16 UTC |
| 37f9db7 | Yash Katariya | 04 October 2022, 05:28:26 UTC | Create `Array`s from `__getitem__` and `__iter__`. This is done by `device_put`ting from the host to default device which is suboptimal. But there is a TODO to fix this! PiperOrigin-RevId: 478691051 | 04 October 2022, 05:29:03 UTC |
