247917c | Jean Kossaifi | 23 December 2018, 23:37:01 UTC | TYPO | 23 December 2018, 23:37:01 UTC |
f04806e | Jean Kossaifi | 23 December 2018, 23:32:26 UTC | MXNET: fix reshape with empty shape | 23 December 2018, 23:32:26 UTC |
05b6b2f | Jean Kossaifi | 23 December 2018, 23:13:49 UTC | Adds tensor contraction | 23 December 2018, 23:13:49 UTC |
3a35360 | Jean Kossaifi | 26 November 2018, 16:25:12 UTC | DOC FIX: home page user_guide link | 26 November 2018, 16:25:12 UTC |
ec8add1 | Jean Kossaifi | 29 October 2018, 14:45:26 UTC | Random: deprecate old functions | 29 October 2018, 14:45:26 UTC |
841e7e6 | Jean Kossaifi | 29 October 2018, 14:30:25 UTC | DOC: update random module API | 29 October 2018, 14:30:25 UTC |
e782136 | Jean Kossaifi | 29 October 2018, 14:29:04 UTC | Refactored random module | 29 October 2018, 14:29:04 UTC |
c5cef8e | Jean Kossaifi | 29 October 2018, 14:28:39 UTC | Sanity checks on rank for MPS | 29 October 2018, 14:28:39 UTC |
54297a3 | Jean Kossaifi | 28 October 2018, 21:00:47 UTC | FIX: use tl.ones not np. | 28 October 2018, 21:00:47 UTC |
8104e06 | Jean Kossaifi | 28 October 2018, 20:53:28 UTC | FIX: (multi_)mode_dot with vector | 28 October 2018, 20:53:28 UTC |
9badf5d | Jean Kossaifi | 28 October 2018, 18:11:22 UTC | Do not set global random seed | 28 October 2018, 18:11:22 UTC |
d4b1854 | Jean Kossaifi | 28 October 2018, 18:02:33 UTC | DOC: add contrib to API | 28 October 2018, 18:02:33 UTC |
f15cf2e | Jean Kossaifi | 28 October 2018, 18:02:01 UTC | Contrib: add decomposition submodule | 28 October 2018, 18:02:01 UTC |
d82e08b | Jean Kossaifi | 28 October 2018, 14:21:49 UTC | Merge pull request #73 from wumming/TTcross Cross-approx algorithm for MSP/tensor-train decomposition | 28 October 2018, 14:21:49 UTC |
a7971af | Jean Kossaifi | 27 October 2018, 19:47:36 UTC | Typos | 27 October 2018, 19:47:36 UTC |
cd25045 | Jean Kossaifi | 27 October 2018, 19:46:30 UTC | Updated guide | 27 October 2018, 19:46:30 UTC |
c8095a4 | Jean Kossaifi | 27 October 2018, 17:13:53 UTC | Removed un-necessary import | 27 October 2018, 17:13:53 UTC |
3ff301c | Jean Kossaifi | 08 October 2018, 12:27:34 UTC | Edit AUTHORS list | 08 October 2018, 12:27:34 UTC |
6f85b26 | Zongyi Li | 21 September 2018, 17:16:45 UTC | move check_random_state(1) | 21 September 2018, 17:16:45 UTC |
5ef625f | Jean Kossaifi | 21 September 2018, 11:08:39 UTC | Specify SVD fun for tucker | 21 September 2018, 11:08:39 UTC |
e22fa80 | Zongyi Li | 19 September 2018, 17:49:39 UTC | delete hard coded dtype. | 19 September 2018, 17:49:39 UTC |
fa50f7a | Zongyi Li | 19 September 2018, 17:46:03 UTC | Merge branch 'master' into TTcross | 19 September 2018, 17:46:03 UTC |
47b57dd | Zongyi Li | 19 September 2018, 17:37:51 UTC | Merge branch 'TTcross' of https://github.com/wumming/tensorly into TTcross | 19 September 2018, 17:37:51 UTC |
0c33fc0 | Zongyi Li | 19 September 2018, 16:12:19 UTC | modify pytorch's tl.tensor(). delete int()/tensor_as_indices() | 19 September 2018, 17:33:45 UTC |
4e9ed27 | Zongyi Li | 19 September 2018, 07:10:38 UTC | change int() to tensor_as_indices() and minor changes | 19 September 2018, 17:33:45 UTC |
463b553 | Zongyi Li | 14 September 2018, 18:15:59 UTC | use solve instead of inverse, delete npr.seed() | 19 September 2018, 17:33:45 UTC |
975c84b | Zongyi Li | 08 September 2018, 21:28:00 UTC | Fix bugs for MxNet 1. wrapper Numpy's inverse for MxNet. 2. MxNet is not robust when vector has size 1. Add a base case for it. 3. MxNet is not robust for a/b. transpose b so that a and b are in the same shape. | 19 September 2018, 17:33:45 UTC |
f0cdb8e | Zongyi Li | 03 September 2018, 21:31:50 UTC | candecomp_parafac.py should be in tensorly/decomposition instead of tensorly/contrib | 19 September 2018, 17:33:45 UTC |
2162dda | Zongyi Li | 03 September 2018, 21:27:14 UTC | add back candecomp_parafac.py from master | 19 September 2018, 17:33:45 UTC |
e918a01 | Zongyi Li | 03 September 2018, 03:57:13 UTC | delete candecomp I didn't change candecomp. Only modified mps_cross | 19 September 2018, 17:33:45 UTC |
c0e6781 | Jean Kossaifi | 19 September 2018, 17:06:18 UTC | Import dtypes | 19 September 2018, 17:06:18 UTC |
87e50c8 | Jean Kossaifi | 19 September 2018, 16:22:45 UTC | pytorch backend: don't force default dtype | 19 September 2018, 16:22:45 UTC |
0e08f2b | Zongyi Li | 19 September 2018, 16:12:19 UTC | modify pytorch's tl.tensor(). delete int()/tensor_as_indices() | 19 September 2018, 16:12:19 UTC |
910a1c2 | Zongyi Li | 19 September 2018, 07:10:38 UTC | change int() to tensor_as_indices() and minor changes | 19 September 2018, 07:10:38 UTC |
f08b657 | Zongyi Li | 14 September 2018, 18:15:59 UTC | use solve instead of inverse, delete npr.seed() | 14 September 2018, 18:15:59 UTC |
6be1f6d | Zongyi Li | 08 September 2018, 21:28:00 UTC | Fix bugs for MxNet 1. wrapper Numpy's inverse for MxNet. 2. MxNet is not robust when vector has size 1. Add a base case for it. 3. MxNet is not robust for a/b. transpose b so that a and b are in the same shape. | 08 September 2018, 21:28:00 UTC |
4c4b32e | Zongyi Li | 03 September 2018, 21:31:50 UTC | candecomp_parafac.py should be in tensorly/decomposition instead of tensorly/contrib | 03 September 2018, 21:31:50 UTC |
bd9e00d | Zongyi Li | 03 September 2018, 21:27:14 UTC | add back candecomp_parafac.py from master | 03 September 2018, 21:27:14 UTC |
5f1dae9 | Zongyi Li | 03 September 2018, 03:57:13 UTC | delete candecomp I didn't change candecomp. Only modified mps_cross | 03 September 2018, 03:57:13 UTC |
5206d0d | Zongyi Li | 03 September 2018, 03:39:10 UTC | Merge branch 'tensorly-master' into TTcross Resolve the conflicts on pytorch and tensorflow backends. | 03 September 2018, 03:39:10 UTC |
3db755b | Zongyi Li | 03 September 2018, 03:36:48 UTC | Resolved merge conflict by incorporating both suggestions. | 03 September 2018, 03:36:48 UTC |
3cce636 | wumming | 30 August 2018, 20:45:19 UTC | clear STT | 30 August 2018, 20:45:19 UTC |
5157ba9 | wumming | 30 August 2018, 20:43:58 UTC | delete STT | 30 August 2018, 20:43:58 UTC |
4f73baf | wumming | 30 August 2018, 20:15:08 UTC | merge2 | 30 August 2018, 20:15:08 UTC |
e09b6e7 | Jean Kossaifi | 30 August 2018, 10:21:50 UTC | FIX: properly use svd fun provided | 30 August 2018, 10:21:50 UTC |
4737451 | Jean Kossaifi | 30 August 2018, 10:10:52 UTC | CPD: add svd parameter (default to 'numpy_svd') | 30 August 2018, 10:10:52 UTC |
3448786 | wumming | 30 August 2018, 06:40:08 UTC | merge | 30 August 2018, 06:40:08 UTC |
4fe5366 | Jean Kossaifi | 29 August 2018, 14:58:30 UTC | Merge pull request #70 from eriche7133/patch-1 FIX: init uses `==` instead `is` to check for init options | 29 August 2018, 14:58:30 UTC |
808daad | eriche7133 | 29 August 2018, 14:23:12 UTC | Update candecomp_parafac.py Changed "is" to "==" for comparing strings in the initialize_factors function. | 29 August 2018, 14:23:12 UTC |
f1e7e39 | wumming | 21 August 2018, 23:54:41 UTC | Doc for tensor-train decomposition | 21 August 2018, 23:54:41 UTC |
f28fd20 | Jean Kossaifi | 16 August 2018, 18:39:19 UTC | ENH: adds several options to compute truncated SVD | 16 August 2018, 18:39:19 UTC |
5551175 | Jean Kossaifi | 16 August 2018, 18:39:07 UTC | Update Makefile | 16 August 2018, 18:39:07 UTC |
4d24972 | Jean Kossaifi | 14 August 2018, 11:35:58 UTC | FIX/pytorch: to_numpy detaches variables if needed | 14 August 2018, 11:35:58 UTC |
c7b16c8 | wumming | 14 August 2018, 07:52:27 UTC | greedy maxvol | 14 August 2018, 07:52:27 UTC |
b8c1d15 | wumming | 12 August 2018, 03:39:36 UTC | np.random -> randomState | 12 August 2018, 03:39:36 UTC |
f97a04b | wumming | 09 August 2018, 18:14:09 UTC | wordings | 09 August 2018, 18:14:09 UTC |
695bb8c | wumming | 09 August 2018, 18:12:20 UTC | advantages and disadvantages | 09 August 2018, 18:12:20 UTC |
3d39041 | Jean Kossaifi | 08 August 2018, 20:31:24 UTC | Don't specify default dtype for numpy backend So as not to override passed array's existing dtype by default (type cast only if dtype is specified) | 08 August 2018, 20:31:24 UTC |
a29ffb8 | wumming | 07 August 2018, 23:11:41 UTC | tensorly standard | 07 August 2018, 23:11:41 UTC |
0022c2b | wumming | 07 August 2018, 00:21:35 UTC | clear the code | 07 August 2018, 00:21:35 UTC |
2d205e6 | wumming | 06 August 2018, 15:16:22 UTC | comment out tf | 06 August 2018, 15:16:22 UTC |
1b05564 | wumming | 05 August 2018, 22:44:28 UTC | Mxnet works | 05 August 2018, 22:44:28 UTC |
ebea465 | wumming | 27 July 2018, 23:56:33 UTC | tests | 27 July 2018, 23:56:33 UTC |
0f2e8e5 | wumming | 26 July 2018, 09:25:22 UTC | TTcross for numpy and pytorch | 26 July 2018, 09:25:22 UTC |
6649dfc | Shuo Zhou | 11 July 2018, 09:13:25 UTC | Update quickstart.rst | 11 July 2018, 10:40:04 UTC |
0adc913 | Jean Kossaifi | 03 July 2018, 15:46:08 UTC | DOC: add MPS/TTrain | 05 July 2018, 10:07:31 UTC |
3dba905 | Jean Kossaifi | 03 July 2018, 15:45:52 UTC | Optimisations and cleanup | 05 July 2018, 10:07:31 UTC |
8cabcfb | AkisKefalas | 17 June 2018, 18:38:49 UTC | Update mps_tensor.py | 05 July 2018, 10:07:31 UTC |
e8396ae | AkisKefalas | 17 June 2018, 18:36:30 UTC | Update test_mps_tensor.py | 05 July 2018, 10:07:31 UTC |
5bf5faa | Admin | 17 June 2018, 18:32:55 UTC | Moved test_mps_decomposition to correct tests folder | 05 July 2018, 10:07:31 UTC |
40db493 | Admin | 17 June 2018, 18:19:08 UTC | Added mps_tensor, mps_decomposition and their tests | 05 July 2018, 10:07:31 UTC |
0802167 | AlexStasik | 03 July 2018, 17:22:59 UTC | Update candecomp_parafac.py | 04 July 2018, 09:37:25 UTC |
d93d280 | Scott Sievert | 14 June 2018, 20:58:09 UTC | MAINT: update __init__.py | 16 June 2018, 20:25:28 UTC |
b2874c5 | Jean Kossaifi | 08 June 2018, 10:24:34 UTC | FIX: partial_svd when n_eigenvecs is None | 08 June 2018, 10:24:34 UTC |
8d876cf | Jean Kossaifi | 05 June 2018, 13:50:46 UTC | Add contrib submodule | 05 June 2018, 13:50:46 UTC |
f74e403 | Jean Kossaifi | 04 June 2018, 09:09:06 UTC | Bump up version | 04 June 2018, 09:09:06 UTC |
340d83f | Jean Kossaifi | 01 June 2018, 14:37:07 UTC | Merge branch 'master' of https://github.com/tensorly/tensorly | 01 June 2018, 14:37:07 UTC |
b93d017 | Jean Kossaifi | 01 June 2018, 14:33:12 UTC | random tensors: add orthogonal parameter | 01 June 2018, 14:33:12 UTC |
f8782f6 | Jean Kossaifi | 01 June 2018, 14:28:17 UTC | Adds orthogonalise parameter for CP decomposition | 01 June 2018, 14:28:17 UTC |
05efae2 | Jean Kossaifi | 31 May 2018, 22:49:16 UTC | RCP: tests | 31 May 2018, 22:49:16 UTC |
b07ca26 | Jean Kossaifi | 31 May 2018, 22:10:32 UTC | randomised CP: make max_stagnation optional If tol=0 and max_stagnation=0, don't compute rec error | 31 May 2018, 22:10:32 UTC |
33f1200 | Jean Kossaifi | 31 May 2018, 16:56:38 UTC | FIX: preserve tensor context | 31 May 2018, 16:56:38 UTC |
9e6d89a | Jean Kossaifi | 31 May 2018, 10:01:17 UTC | DOC: added randomised CP to API | 31 May 2018, 10:01:17 UTC |
4f30689 | Jean Kossaifi | 31 May 2018, 09:29:34 UTC | Merge pull request #53 from SamJohannes/randomised-cp Adds randomised cp via sampled ALS | 31 May 2018, 09:29:34 UTC |
809da07 | SamJohannes | 30 May 2018, 22:16:19 UTC | Re-added the max_stagnation convergence check to r-parafac | 30 May 2018, 22:16:19 UTC |
34c3178 | Jean Kossaifi | 30 May 2018, 16:29:16 UTC | FIX: test_randomized_cp: rank of source tensor | 30 May 2018, 16:29:16 UTC |
3b72820 | Jean Kossaifi | 30 May 2018, 16:23:50 UTC | Refactored randomised_cp Mainly: - sample only needed elements from the tensor rather than unfold + select - rename with better names - avoid conversion to and from numpy - add convergence test | 30 May 2018, 16:23:50 UTC |
0202177 | SamJohannes | 16 May 2018, 22:46:06 UTC | Removed obselete variables from random-cp | 16 May 2018, 22:46:06 UTC |
fb0b390 | SamJohannes | 15 May 2018, 07:08:16 UTC | Changed the min/max for arrays in testing | 15 May 2018, 07:08:16 UTC |
f32784a | SamJohannes | 15 May 2018, 06:37:26 UTC | Merge branch 'randomised-cp' of https://github.com/SamJohannes/tensorly into randomised-cp | 15 May 2018, 06:37:26 UTC |
05bbc23 | SamJohannes | 15 May 2018, 06:33:38 UTC | Made tests multi-backend compatible | 15 May 2018, 06:33:38 UTC |
3fba4e1 | SamJohannes | 15 May 2018, 06:22:06 UTC | Made tests multi-backend compatible | 15 May 2018, 06:22:06 UTC |
dcf04af | SamJohannes | 15 May 2018, 06:06:22 UTC | Merge branch 'master' into randomised-cp | 15 May 2018, 06:06:22 UTC |
55f2086 | SamJohannes | 15 May 2018, 06:02:39 UTC | Cleaned up backend utilisation for randomised parafac | 15 May 2018, 06:02:39 UTC |
15beeef | SamJohannes | 14 May 2018, 03:32:20 UTC | Removed debugging statements | 14 May 2018, 03:32:20 UTC |
56826f0 | Curtis Hampton | 10 May 2018, 03:48:15 UTC | include license in wheels | 11 May 2018, 10:22:05 UTC |
af0700a | Jean Kossaifi | 10 May 2018, 16:00:31 UTC | Backend, partial_svd: full_matrices=false when sensible | 10 May 2018, 16:00:31 UTC |
9cc2280 | Jean Kossaifi | 09 May 2018, 16:28:15 UTC | CI: install pytorch for cpu only | 09 May 2018, 16:28:15 UTC |
07d3dbb | Jean Kossaifi | 09 May 2018, 16:15:24 UTC | DOC: citation + pytorch 0.4.0+ | 09 May 2018, 16:15:24 UTC |
1295ccb | Jean Kossaifi | 08 May 2018, 21:04:53 UTC | FIX README + bump version | 08 May 2018, 22:15:23 UTC |