forked from wesselb/stheno
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtodo.tasks
117 lines (106 loc) · 6.96 KB
/
todo.tasks
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
TODO:
☐ Test WeightedUnique @high
Refactor:
Stheno 1.0:
☐ Anchestral sampling
☐ Remove duplication in handling of boxing
☐ Let SparseObs take in Obs for consistency
☐ Sparse -> Pseudo
☐ Periodic GP
☐ FixedDelta -> make use of traits
☐ RowVecs
☐ ColVecs
☐ Add eval @low
Misc:
☐ Throw warning at fallback for elwise
Optimise:
☐ SelectedKernel
README:
☐ Additive pseudo-point example
☐ B.epsilon
Features:
☐ DerivativeKernel for Torch, AutoGrad, and Jax
---- ✄ -----------------------
Features:
☐ FITC approximation
☐ DTC approximation
Multi Output Support:
☐ ICM
☐ (O)LMM
Profiling:
☐ Single sample 100-d normal maximum 20% slower.
＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿
Archive:
✓ Remove uprank: use B.uprank @done (21-11-30 19:26) @project(TODO / Refactor / Misc)
✓ Update manual, e.g. kernels @done (20-11-30 19:26) @project(TODO / README)
✓ Note: accumulation to default graph can cause memory problems @critical @done (20-11-30 19:26) @project(TODO / README)
✓ Note that kernel is given first everywhere @done (20-11-30 19:26) @project(TODO / README)
✓ Refactor tests to use PyTest: remove raises, fixtures, and parametrisation. @done (20-11-30 19:25) @project(TODO)
✓ Refactor tests @done (20-11-30 19:25) @project(TODO)
✓ GP composition framework @done (20-11-30 19:25) @project(TODO / Refactor / Stheno 1.0)
✓ Fixtures for kernel inputs @done (20-11-30 19:25) @project(TODO)
✓ Fix coverage @done (20-11-30 19:25) @project(TODO)
✓ Construct GPs and Normals both ways around @done (20-11-30 19:25) @project(TODO / Refactor / Stheno 1.0)
✓ Clean printing @done (20-11-30 19:25) @project(TODO)
x Use `mul_matched` for approximate multiplication @cancelled (20-11-30 19:25) @project(TODO / Refactor / Stheno 1.0)
✓ Remove clutter @done (19-12-29 16:22) @project(TODO)
✓ Kernel properties except for `stationary` @done (19-12-29 16:22) @project(TODO)
✓ Normal1D @done (19-12-29 16:18) @project(TODO)
✓ EIS @done (19-12-29 16:18) @project(TODO)
✓ Use ABC with six @done (19-12-29 16:17) @project(TODO / Refactor / Misc)
✓ Do not automatically load TensorFlow at startup @done (19-12-29 16:17) @project(TODO / Refactor / Misc)
✓ Remove exceptions for torch 0-dim behaviour @done (19-12-29 16:16) @project(TODO / Refactor / Misc)
✓ Fix `test_delta_evaluations` for `Unique` inputs @done (19-12-29 16:10) @project(TODO)
✓ Remove diff_approx @done (19-12-29 16:06) @project(TODO / Refactor / Stheno 1.0)
✓ Clean test random @done (19-12-29 16:05) @project(TODO)
✓ Remove Python 2 compat remains @done (19-12-29 16:04) @project(TODO)
✓ Minimise B.dense calls @done (19-12-29 16:02) @project(TODO)
✓ Ensure that cov. matrices are AMs with dispatch @done (19-12-29 16:02) @project(TODO)
✓ StruMat @done (19-12-29 15:29) @project(TODO / Refactor / Stheno 1.0 / Move to external packages)
✓ Rings @done (19-12-29 15:29) @project(TODO / Refactor / Stheno 1.0 / Move to external packages)
✓ Kronecker product, and sum thereof @done (19-12-27 17:42) @project(Structured Matrices)
x Toeplitz @cancelled (19-12-27 17:42) @project(Structured Matrices)
x Block matrix @cancelled (19-12-27 17:42) @project(Structured Matrices)
✓ Fix examples to use new WBML and LAB @done (19-11-19 16:39) @project(TODO / README)
✓ Rename tf->tensorflow, but still allow tf @critical @done (19-07-10 22:28) @project(TODO)
✓ LogKernel @done (19-05-02 16:53) @project(TODO / Test)
✓ Kernel powers @done (19-05-02 16:53) @project(TODO / Test)
✓ Multi-argument sparse conditioning @done (18-10-14 16:51) @project(Features)
✓ Simply signatures in `field.py`. @done (18-07-04 16:52) @project(Refactor)
✓ Efficient multi-output kernels: low-rank SPDs returned by multi-output kernels @done (18-06-26 15:39) @project(Features)
✓ Write documentation. @started @done (18-06-26 15:38) @project(Documentation)
✓ Sparse approximations @done (18-06-26 15:38) @project(Features)
✓ Improve docstrings. @done (18-06-26 15:38) @project(Documentation)
✓ Define terminology in the docstrings: design matrix, shape, etc. @done (18-06-26 15:38) @project(Documentation)
✓ Refactor to type union for Graph.condition once type unions are fixed. @done (18-06-19 02:10) @project(Refactor)
✓ Efficient diagonal computation: pairwise and element-wise @started @done (18-06-15 21:39) @project(Features)
✓ Write examples. @started @done (18-06-13 23:40) @project(Documentation)
✓ LMM/ICM @started @done (18-06-13 23:40) @project(Examples)
✓ What is going on with `[np.random.randn(1,1).__array_interface__['data'][0] for _ in range(2)]`? @done (18-06-11 13:49)
✓ Caching of computation in kernels @done (18-06-04 03:43) @project(Optimisations)
✓ Implement * and + for SPDs, and use these in the arithmetic of Normals. @done (18-06-02 17:49) @project(Optimisations)
✓ Write tests. @started @done (18-05-31 16:14) @project(Documentation / Other)
✓ Regression @done (18-05-31 16:13) @project(Features)
✓ Posterior decomposition @done (18-05-31 16:13) @project(Examples)
✓ Posterior decomposition (additive) @done (18-05-31 16:13) @project(Features)
✓ Make all compatible with all the backends. @done (18-05-31 16:13) @project(Examples)
✓ Extended input space. @done (18-05-31 16:13) @project(Features)
✓ Simple posterior decomposition. @started @done (18-05-31 15:41) @project(Current TODO)
✓ Extended input space GP regression. @started @done (18-05-31 12:02) @project(Current TODO)
✓ Fix examples. @done (18-05-30 20:27) @project(Current TODO)
✓ Arithmetic for GPs. @done (18-05-30 20:13) @project(Current TODO)
✓ Arithmetic for means. @done (18-05-30 19:34) @project(Current TODO)
✓ Disable caching in Plum? @done (18-05-30 18:16) @project(Current TODO)
✓ Implement sum and product as Kernels @high @done (18-05-30 18:15) @project(Current TODO / Fix kernels)
✓ Set up Sphinx. @done (18-03-29 17:10) @project(Documentation)
✓ Rename repo. @done (18-03-29 17:10) @project(Documentation / Other)
✓ Package. @done (18-03-29 17:10) @project(Documentation / Other)
✓ Name? gptools is already taken. @done (17-07-31 12:48) @project(Documentation / Other)
✓ Remove noise option from GP. @done (17-06-29 22:13) @project(Features)
✓ Optimise import statements for cleanness: `from core import np` to import the proxy. @done (17-06-29 14:58) @project(Features)
✓ Write init files. @done (17-06-28 00:12) @project(Features)
✓ More elegant TensorFlow import. @done (17-06-28 00:12) @project(Features)
✓ Learning @started(17-06-26 18:42) @done (17-06-28 00:12) @lasted(1 day, 5:30) @project(Examples)
✓ Explicitly define interfaces for modules. @done (17-06-28 00:12) @project(Features)
✓ Kernels @done (17-06-26 18:41) @project(Examples)
✓ Periodic kernel. @done (17-06-26 15:25) @project(Features)