63fc5c5 | Aaron Meyer | 03 April 2024, 14:09:47 UTC | Merge pull request #538 from braun-steven/main Fix torch tensor creation to preserve dtype and device. | 03 April 2024, 14:09:47 UTC |
a49f9b9 | Aaron Meyer | 03 April 2024, 13:35:45 UTC | Update test_proximal.py Making a trivial change to rebuild the tests. | 03 April 2024, 13:35:45 UTC |
5a19c82 | Jean Kossaifi | 02 April 2024, 21:53:56 UTC | Merge pull request #548 from FBen3/add_missing_makefile_targets Add missing targets to .PHONY directive | 02 April 2024, 21:53:56 UTC |
11d683f | Benjamin Fockter | 28 March 2024, 23:30:24 UTC | Add missing targets to .PHONY directive | 28 March 2024, 23:30:24 UTC |
59ba570 | Jean Kossaifi | 18 March 2024, 18:21:24 UTC | Merge pull request #511 from OsmanMalik/master Add randomized sampling-based ALS method for tensor ring decomposition | 18 March 2024, 18:21:24 UTC |
768dc0a | Aaron Meyer | 12 March 2024, 03:39:50 UTC | Start testing Python 3.12 and resolve JAX deprecation (#540) * Upgrade python * Update test.yml --------- Co-authored-by: Aaron Meyer <git@asmlab.org> | 12 March 2024, 03:39:50 UTC |
cfc82a6 | Osman Asif Malik | 24 February 2024, 20:25:12 UTC | Apply updated black | 24 February 2024, 20:25:12 UTC |
2d7c27b | Osman Asif Malik | 24 February 2024, 20:22:01 UTC | Merge branch 'tensorly:main' into master | 24 February 2024, 20:22:01 UTC |
0d5af9d | Osman Asif Malik | 24 February 2024, 20:18:04 UTC | Split TR files based on main algorithm Split TR decomposition and testing files based on type of algorithm (SVD or ALS). | 24 February 2024, 20:18:04 UTC |
2a8ff56 | Aaron Meyer | 08 January 2024, 15:03:19 UTC | Simplify proximal operator code (#534) Loops could be removed through vectorized indexing. Also, no need to keep track of the passive and active indices because these are complementary. | 08 January 2024, 15:03:19 UTC |
7569688 | Aaron Meyer | 08 January 2024, 14:33:14 UTC | Merge pull request #533 from meyer-lab/rm-mxnet Remove mxnet entirely, as the project has been halted and archived. | 08 January 2024, 14:33:14 UTC |
a46f479 | Jean Kossaifi | 25 December 2023, 16:34:30 UTC | Merge pull request #539 from meyer-lab/cross-product Avoid large concatenation within PARAFAC2 upon SVD initialization | 25 December 2023, 16:34:30 UTC |
3b0499a | Aaron Meyer | 04 December 2023, 04:51:44 UTC | Add information about cross-product to the docstring | 04 December 2023, 04:51:44 UTC |
d850fe9 | Osman Asif Malik | 14 November 2023, 02:18:05 UTC | Fix bug with handling of function output | 14 November 2023, 02:18:05 UTC |
17aecca | Osman Asif Malik | 14 November 2023, 01:42:20 UTC | Merge branch 'tensorly:main' into master | 14 November 2023, 01:42:20 UTC |
3104f84 | Steven Braun | 10 November 2023, 07:52:20 UTC | Simplify torch tensor creation | 10 November 2023, 07:52:20 UTC |
e32701d | Steven Braun | 09 November 2023, 22:13:02 UTC | Remove unnecessary torch.tensor call | 09 November 2023, 22:13:02 UTC |
44d3391 | Jean Kossaifi | 09 November 2023, 19:35:44 UTC | Merge pull request #535 from meyer-lab/fix-pf2 Small fixes in NN PARAFAC/PARAFAC2 | 09 November 2023, 19:35:44 UTC |
fb516e8 | Aaron Meyer | 09 November 2023, 19:35:09 UTC | Add max rank argument to svd_compress_tensor_slices (#536) * Add max rank argument * Fix variable typo | 09 November 2023, 19:35:09 UTC |
a3721d9 | Steven Braun | 09 November 2023, 08:27:08 UTC | Fix failing test cases | 09 November 2023, 08:27:08 UTC |
259a154 | Steven Braun | 09 November 2023, 08:26:43 UTC | Format tensorflow_backend.py | 09 November 2023, 08:26:43 UTC |
43ca82b | Steven Braun | 09 November 2023, 08:26:11 UTC | Make dtype tensor init test for all backends | 09 November 2023, 08:26:11 UTC |
1a320eb | Steven Braun | 09 November 2023, 08:00:39 UTC | Fix TF dtype/device tensor init See also: #537 | 09 November 2023, 08:00:39 UTC |
90760da | Steven Braun | 09 November 2023, 07:03:37 UTC | Fix missing tensor contexts See also: #538 | 09 November 2023, 07:03:37 UTC |
9e2435f | Steven Braun | 09 November 2023, 07:02:02 UTC | Add torch tests for tensor init dtype fix See also: #538 | 09 November 2023, 07:02:02 UTC |
90cda20 | Aaron Meyer | 08 November 2023, 17:23:20 UTC | Looser tolerance for PyTorch | 08 November 2023, 17:23:20 UTC |
4b8a515 | Aaron Meyer | 08 November 2023, 15:38:18 UTC | Done | 08 November 2023, 15:38:18 UTC |
b5f81c0 | Steven Braun | 08 November 2023, 11:59:49 UTC | Fix tests that previously assumed float32 | 08 November 2023, 11:59:49 UTC |
45e8631 | Steven Braun | 08 November 2023, 11:22:16 UTC | Merge branch 'main' of github.com:braun-steven/tensorly | 08 November 2023, 11:22:16 UTC |
cc3c652 | Steven Braun | 08 November 2023, 10:55:10 UTC | Fix torch tensor creation dtype/device The previous implementation simply took the input data and forced it to be on the given device (`cpu` default) with the given dtype (`float32` default). This lead to `tl.tensor(torch.zeros(1, device="cuda:0"))` creating a float32 tensor on `cpu` while the expected behavior should be an int32 tensor on `cuda:0`. Resolves: #537 | 08 November 2023, 11:18:36 UTC |
a99232d | Steven Braun | 08 November 2023, 10:55:10 UTC | Fix torch tensor creation dtype/device The previous implementation simply took the input data and forced it to be on the given device (`cpu` default) with the given dtype (`float32` default). This lead to `tl.tensor(torch.zeros(1, device="cuda:0"))` creating a float32 tensor on `cpu` while the expected behavior should be an int32 tensor on `cuda:0`. Resolves: #537 | 08 November 2023, 10:58:15 UTC |
ad5eb92 | Aaron Meyer | 30 October 2023, 03:25:28 UTC | Fix context handling | 30 October 2023, 03:25:28 UTC |
7f5176b | Aaron Meyer | 28 October 2023, 14:49:22 UTC | Fix pf2 initialization | 28 October 2023, 14:49:22 UTC |
1fe1fc7 | Jean Kossaifi | 20 October 2023, 16:00:31 UTC | Merge pull request #532 from meyer-lab/backend-deps Fully deprecate mxnet | 20 October 2023, 16:00:31 UTC |
9c82c5e | Aaron Meyer | 19 October 2023, 13:14:11 UTC | Remove mxnet | 19 October 2023, 13:14:11 UTC |
dac438b | Aaron Meyer | 19 October 2023, 13:00:53 UTC | Deprecate mxnet | 19 October 2023, 13:00:53 UTC |
357aa66 | Aaron Meyer | 18 October 2023, 19:05:26 UTC | Merge pull request #525 from meyer-lab/pf2-linesearch Bro's line search for PARAFAC2 | 18 October 2023, 19:05:26 UTC |
6c549a5 | Aaron Meyer | 18 October 2023, 19:00:20 UTC | Fix black formatting | 18 October 2023, 19:00:20 UTC |
34e99f7 | Aaron Meyer | 18 October 2023, 18:01:50 UTC | Fix _BroThesisLineSearch import | 18 October 2023, 18:01:50 UTC |
a4b4cf7 | Aaron Meyer | 18 October 2023, 17:59:48 UTC | Last edits | 18 October 2023, 17:59:48 UTC |
c79b0ba | Aaron Meyer | 18 October 2023, 17:56:40 UTC | Update tensorly/decomposition/_parafac2.py Co-authored-by: Marie Roald <roald.marie@gmail.com> | 18 October 2023, 17:56:40 UTC |
7a99839 | Aaron Meyer | 18 October 2023, 17:49:53 UTC | Update tensorly/decomposition/tests/test_parafac2.py Co-authored-by: Marie Roald <roald.marie@gmail.com> | 18 October 2023, 17:49:53 UTC |
d3ba291 | Aaron Meyer | 14 October 2023, 16:38:36 UTC | Add second docstring | 14 October 2023, 16:38:36 UTC |
1097e20 | Aaron Meyer | 14 October 2023, 16:29:35 UTC | Black formatting | 14 October 2023, 16:29:35 UTC |
9585da4 | Aaron Meyer | 14 October 2023, 16:28:28 UTC | Add first docstring | 14 October 2023, 16:28:28 UTC |
33399bb | Aaron Meyer | 14 October 2023, 15:56:38 UTC | Merge branch 'main' into pf2-linesearch | 14 October 2023, 15:56:38 UTC |
848f87c | Aaron Meyer | 03 October 2023, 13:47:58 UTC | Apply Marie's suggestions Co-authored-by: Marie Roald <roald.marie@gmail.com> | 03 October 2023, 13:47:58 UTC |
4ba1cbe | Aaron Meyer | 27 September 2023, 16:38:49 UTC | Merge pull request #530 from MarieRoald/parafac2-compression SVD compression of PARAFAC2 | 27 September 2023, 16:38:49 UTC |
857c462 | Marie Roald | 22 September 2023, 05:06:28 UTC | Add SVD compression for PARAFAC2 Co-authored-by: Yngve Mardal Moe <yngve.m.moe@gmail.com> | 27 September 2023, 14:36:03 UTC |
b12844f | Marie Roald | 22 September 2023, 05:04:09 UTC | Fix typo and improve test parametrisation Co-authored-by: Yngve Mardal Moe <yngve.m.moe@gmail.com> | 24 September 2023, 04:06:08 UTC |
939b28c | Aaron Meyer | 23 September 2023, 21:52:09 UTC | Basic structure for separating line search | 23 September 2023, 21:52:09 UTC |
f81f97b | Aaron Meyer | 23 September 2023, 21:16:13 UTC | Update tensorly/decomposition/_parafac2.py Co-authored-by: Marie Roald <roald.marie@gmail.com> | 23 September 2023, 21:16:13 UTC |
293e581 | Aaron Meyer | 19 September 2023, 14:07:07 UTC | Fixes based on MarieRoald review | 19 September 2023, 14:07:07 UTC |
72174be | Aaron Meyer | 08 September 2023, 18:10:36 UTC | Merge pull request #517 from meyer-lab/fix-nnsvd-returns Also provide H back from make_svd_non_negative() | 08 September 2023, 18:10:36 UTC |
e22e79f | Aaron Meyer | 08 September 2023, 17:23:39 UTC | Fix mxnet context of S | 08 September 2023, 17:23:39 UTC |
2e7944f | Aaron Meyer | 08 September 2023, 15:49:42 UTC | Fix and test svd shapes | 08 September 2023, 15:49:42 UTC |
f0a6b6e | Aaron Meyer | 04 September 2023, 22:49:51 UTC | Fix typo in error comparison | 04 September 2023, 22:49:51 UTC |
84372ea | Aaron Meyer | 04 September 2023, 22:48:46 UTC | Fix that we need to reproject within the line search | 04 September 2023, 22:48:46 UTC |
bdde159 | Aaron Meyer | 04 September 2023, 22:36:13 UTC | Implemented | 04 September 2023, 22:36:13 UTC |
c1775b8 | Jean Kossaifi | 01 September 2023, 17:29:11 UTC | Merge pull request #524 from meyer-lab/pf2-avoid-reproject PARAFAC2: Avoid reprojection of X during the error calculation | 01 September 2023, 17:29:11 UTC |
1f00ea9 | Osman Asif Malik | 23 August 2023, 01:53:29 UTC | Add test of function computing leverage scores | 23 August 2023, 01:53:29 UTC |
3e59f5d | Osman Asif Malik | 23 August 2023, 01:50:04 UTC | Move function for calculating leverage scores Also: - Rename function for calculating leverage scores. - Add reference to paper on leverage scores. - Fix bug in how num_rank is computed. | 23 August 2023, 01:50:04 UTC |
554ad99 | Aaron Meyer | 21 August 2023, 19:03:36 UTC | Avoid reprojection of X during error calculation | 21 August 2023, 19:03:36 UTC |
de05e17 | Jean Kossaifi | 20 August 2023, 18:28:25 UTC | Merge pull request #505 from JeanKossaifi/main Default on math.pi | 20 August 2023, 18:28:25 UTC |
83a4c48 | Jean Kossaifi | 19 August 2023, 09:52:36 UTC | Merge pull request #519 from hello-fri-end/fixInitializeTucker Fix initialize_tucker | 19 August 2023, 09:52:36 UTC |
4f8a945 | Anwaar Khalid | 19 August 2023, 09:06:41 UTC | Fixed linting errors with black | 19 August 2023, 09:06:41 UTC |
3e33fad | Jean Kossaifi | 18 August 2023, 08:34:57 UTC | Merge pull request #514 from SauravMaheshkar/saurav/ci-pip-cache Doc: adds `pip` caching to CI | 18 August 2023, 08:34:57 UTC |
8bc7747 | Saurav Maheshkar | 15 August 2023, 07:52:40 UTC | Merge branch 'main' into saurav/ci-pip-cache | 15 August 2023, 07:52:40 UTC |
e997012 | Saurav Maheshkar | 15 August 2023, 07:51:51 UTC | refactor(ci): organise pip cache Signed-off-by: Saurav Maheshkar <sauravvmaheshkar@gmail.com> | 15 August 2023, 07:51:51 UTC |
6ec126f | hello-fri-end | 08 August 2023, 05:24:02 UTC | fixed the shape of core and factors when init==random in intialize_tucker and updated the test of partial_tucker to pass init==random when testting randomstate | 08 August 2023, 05:24:02 UTC |
a982cf9 | Aaron Meyer | 02 August 2023, 21:45:44 UTC | Merge pull request #513 from meyer-lab/fix-exports-plsr-perf Miscellaneous housekeeping improvements | 02 August 2023, 21:45:44 UTC |
b5161a7 | Aaron Meyer | 02 August 2023, 20:40:51 UTC | Break out return values | 02 August 2023, 20:40:51 UTC |
b99cb2c | Aaron Meyer | 02 August 2023, 15:23:13 UTC | Also provide H back | 02 August 2023, 15:23:13 UTC |
373c46a | Saurav Maheshkar | 25 July 2023, 16:52:18 UTC | feat: add pip caching to CI | 25 July 2023, 16:52:18 UTC |
9db2226 | Aaron Meyer | 22 July 2023, 17:51:09 UTC | Fix YAML | 22 July 2023, 17:51:09 UTC |
d132435 | Aaron Meyer | 22 July 2023, 17:43:52 UTC | Upgrade to Python 3.10 by default | 22 July 2023, 17:43:52 UTC |
72c51cb | Aaron Meyer | 22 July 2023, 17:33:00 UTC | Other fixes | 22 July 2023, 17:33:00 UTC |
d7c0c8b | Aaron Meyer | 22 July 2023, 17:21:16 UTC | Loosen tensor-train assertion | 22 July 2023, 17:21:16 UTC |
6c7ecfb | Jean Kossaifi | 21 July 2023, 18:44:02 UTC | Merge pull request #508 from Kiord/main Documentation update for svd missing values imputation | 21 July 2023, 18:44:02 UTC |
ec8b6f9 | Osman Asif Malik | 10 July 2023, 21:02:33 UTC | Add option for uniform sampling New parameter allows for uniform sampling to be used instead of leverage score sampling. | 10 July 2023, 21:02:33 UTC |
7263a6f | Osman Asif Malik | 10 July 2023, 13:56:16 UTC | Add randomized tensor ring decomposition Add the randomized sampling-based ALS method for tensor ring decomposition proposed in O. A. Malik, S. Becker, "A Sampling-Based Method for Tensor Ring Decomposition", Proceedings of the 38th International Conference on Machine Learning (ICML), PMLR 139:7400-7411, 2021. | 10 July 2023, 13:56:16 UTC |
f409de5 | Glenn Kerbiriou | 28 June 2023, 07:30:59 UTC | typo | 28 June 2023, 07:30:59 UTC |
5e4e7a5 | Glenn Kerbiriou | 28 June 2023, 07:21:14 UTC | Documentation cues for imputation with svd, and added a condition to perform matrix masking | 28 June 2023, 07:21:14 UTC |
a773116 | Aaron Meyer | 12 June 2023, 19:42:36 UTC | Fix CMTF test | 12 June 2023, 19:42:36 UTC |
bb6bf0c | Aaron Meyer | 11 June 2023, 16:26:17 UTC | Fix black linting | 11 June 2023, 16:26:17 UTC |
790c1d7 | Aaron Meyer | 11 June 2023, 16:25:23 UTC | Fix occasional torch error | 11 June 2023, 16:25:23 UTC |
e03341e | Aaron Meyer | 07 June 2023, 14:50:05 UTC | Black lint | 07 June 2023, 14:50:05 UTC |
8573328 | Aaron Meyer | 07 June 2023, 14:47:51 UTC | Testing adjustments | 07 June 2023, 14:47:51 UTC |
f660108 | Aaron Meyer | 07 June 2023, 03:13:58 UTC | Black formatting | 07 June 2023, 03:13:58 UTC |
afd8f09 | Aaron Meyer | 07 June 2023, 03:12:38 UTC | Fix pytorch | 07 June 2023, 03:12:38 UTC |
54dc5a3 | Aaron Meyer | 07 June 2023, 03:09:44 UTC | Fix Jax | 07 June 2023, 03:09:44 UTC |
035fbf3 | Aaron Meyer | 06 June 2023, 14:04:21 UTC | PLSR much faster, and pass-through of lstsq | 06 June 2023, 14:04:21 UTC |
bcc00bf | Aaron Meyer | 05 June 2023, 15:07:24 UTC | Fix exports and improve PLSR performance | 05 June 2023, 15:07:24 UTC |
94c2ff2 | Yuri Brigance | 01 June 2023, 14:04:59 UTC | Fixed svd_flip() when used on GPU w/ PyTorch backend (#504) * 503: Fixed svd_flip() when used on GPU w/ PyTorch backend * Update svd.py Added context instead of just device --------- Co-authored-by: Jean Kossaifi <jean.kossaifi@gmail.com> | 01 June 2023, 14:04:59 UTC |
886bc57 | Jean Kossaifi | 26 May 2023, 19:31:00 UTC | Use math.pi by default | 26 May 2023, 19:31:00 UTC |
423d03c | Jean Kossaifi | 03 April 2023, 01:18:16 UTC | Cosmetic | 26 May 2023, 19:30:22 UTC |
1911fd0 | Jean Kossaifi | 26 May 2023, 19:22:50 UTC | Merge pull request #501 from OsmanMalik/master Add ALS-based method for tensor ring decomposition | 26 May 2023, 19:22:50 UTC |
42218f8 | Osman Asif Malik | 26 May 2023, 01:48:14 UTC | Modify code to use matricize | 26 May 2023, 01:48:14 UTC |
e45c97e | Aaron Meyer | 17 May 2023, 13:59:52 UTC | More efficient error calculation in PARAFAC2 through inner product (#502) * More efficient error calculation * Fix transpose * Fix test * Black formatting * Add docstring --------- Co-authored-by: Aaron Meyer <git@asmlab.org> | 17 May 2023, 13:59:52 UTC |
f02b1c1 | Osman Asif Malik | 11 April 2023, 00:03:26 UTC | Add ALS-based method for tensor ring decomposition - Add function for tensor ring decomposition via ALS - Add class wrapper for said function - Add tests for new functionality | 12 April 2023, 18:18:37 UTC |