Commits · master · Quantum Tinkerer / adaptive

07 Feb, 2019 1 commit
- DEVELOPMENT IS ON GITHUB · fcba32e5
  Bas Nijholt authored Feb 07, 2019
  
  fcba32e5
17 Dec, 2018 3 commits
- Merge branch 'stable-0.7' · 0207eef4
  Joseph Weston authored Dec 17, 2018
  
  0207eef4
- take fewer points to speed up the test for the LearnerND · 21a836e8
  Bas Nijholt authored Dec 09, 2018 and Joseph Weston committed Dec 17, 2018
  
  21a836e8
- change the simplex_queue to a SortedKeyList · 3c901037
  Bas Nijholt authored Dec 09, 2018 and Joseph Weston committed Dec 17, 2018
  
  3c901037
14 Dec, 2018 1 commit
- ensure that points are never outside of the domain, closes #132 · e17a27d6
  Bas Nijholt authored Dec 14, 2018
  
  e17a27d6
11 Dec, 2018 2 commits
- update dependency versions · 23534389
  Joseph Weston authored Dec 11, 2018
```
We use sortedcontainers 2.0 API.
```
  23534389
- make methods private in the LearnerND, closes #85 · 7c340516
  Bas Nijholt authored Dec 11, 2018
  
  7c340516
07 Dec, 2018 9 commits
- change url in setup.py to the documentation · 190248ce
  Bas Nijholt authored Dec 07, 2018
  
  190248ce
- start development for v0.8 · 76593013
  Bas Nijholt authored Dec 07, 2018
  
  76593013
- Merge branch '78-scale-output-values' into 'master' · a90d2579
  Bas Nijholt authored Dec 07, 2018
```
LearnerND scale output values before computing loss

Closes #78

See merge request !128
```
  a90d2579
- LearnerND scale output values before computing loss · 24ac01af
  Jorn Hoofwijk authored Dec 07, 2018 and Bas Nijholt committed Dec 07, 2018
  
  24ac01af
- fix plotting resolution for larger scaled learnerND · fa4696ed
  Jorn Hoofwijk authored Dec 07, 2018 and Bas Nijholt committed Dec 07, 2018
  
  fa4696ed
- fix example in doc-string of BalancingLearner.save · a0d5e699
  Bas Nijholt authored Dec 07, 2018
  
  a0d5e699
- implement own version of 'round' that speeds up the Learner1D by 30% · 31bebf0d
  Bas Nijholt authored Dec 07, 2018
  
  31bebf0d
- Merge branch '126--speed-up-learner1D' into 'master' · 669cf077
  Bas Nijholt authored Dec 07, 2018
```
Resolve "(Learner1D) improve time complexity"

Closes #126 and #104

See merge request !139
```
  669cf077
- Resolve "(Learner1D) improve time complexity" · 4381a9e5
  Jorn Hoofwijk authored Dec 07, 2018 and Bas Nijholt committed Dec 07, 2018
  
  4381a9e5
05 Dec, 2018 1 commit
- fix typo in 'uniform_loss' doc-string, closes #127 · c0012a9a
  Bas Nijholt authored Dec 05, 2018
  
  c0012a9a
28 Nov, 2018 1 commit
- actually compute the dx by subtracting two DIFFERENT numbers · f38d20fe
  Jorn Hoofwijk authored Nov 28, 2018 and Bas Nijholt committed Nov 28, 2018
  
  f38d20fe
27 Nov, 2018 1 commit
- fix bug with saving of the BalancingLearner · 9c266692
  Bas Nijholt authored Nov 27, 2018
  
  9c266692
26 Nov, 2018 10 commits
- adhere to PEP008 by using absolute imports · 06b9200f
  Bas Nijholt authored Nov 26, 2018
```
See https://www.python.org/dev/peps/pep-0008/#imports
```
  06b9200f
- Merge branch 'test_loss_functions' into 'master' · 2f2e80d0
  Bas Nijholt authored Nov 26, 2018
```
test all the different loss functions in each test

See merge request !135
```
  2f2e80d0
- use math.isclose · d2d955e0
  Bas Nijholt authored Nov 23, 2018
  
  d2d955e0
- run fewer points in 'test_learner_performance_is_invariant_under_scaling' · 4da759fb
  Bas Nijholt authored Nov 23, 2018
  
  4da759fb
- add '_recompute_losses_factor' such that we can set it in the test · 38ff5e26
  Bas Nijholt authored Nov 23, 2018
```
and have fewer xfailing tests
```
  38ff5e26
- add 'with_all_loss_functions' to 'run_with' · d6b2187e
  Bas Nijholt authored Nov 23, 2018
  
  d6b2187e
- round the losses to 12 digits to make them equal · be74ad62
  Bas Nijholt authored Nov 22, 2018
```
Then equal losses will be sorted on x-coordinates.
```
  be74ad62
- refactor adding all loss functions to tests involving a learner · c589a2f6
  Joseph Weston authored Nov 22, 2018 and Bas Nijholt committed Nov 26, 2018
  
  c589a2f6
- test all the different loss functions in each test · ef227b84
  Bas Nijholt authored Nov 22, 2018
  
  ef227b84
- Merge branch 'feature/fname' into 'master' · d82fbae7
  Joseph Weston authored Nov 26, 2018
```
make 'fname' a parameter to 'save' and 'load' only

Closes #122

See merge request !133
```
  d82fbae7
23 Nov, 2018 4 commits
- install git in the testing Docker image · 353bebb8
  Bas Nijholt authored Nov 23, 2018
  
  353bebb8
- add an option to use a list of filenames when saving a BalancingLearner · 329f447b
  Bas Nijholt authored Nov 23, 2018
  
  329f447b
- make 'fname' a parameter to 'save' and 'load' only · fcdb0532
  Joseph Weston authored Nov 19, 2018 and Bas Nijholt committed Nov 23, 2018
```
This simplifies the API by making sure that the filenames are
only provided in one place (the calls to save and load).

Closes #122
```
  fcdb0532
- build the Dockerimage used in CI · 65ea8401
  Bas Nijholt authored Nov 23, 2018
  
  65ea8401
22 Nov, 2018 7 commits
- fix several documentation mistakes · b00f266f
  Bas Nijholt authored Nov 22, 2018
```
* add 'curvature_loss_function' to the 'tutorial.custom_loss.rst'
* fix header styling
* fix doc-string
```
  b00f266f
- change resolution_loss to a factory function · d5a63a3f
  Bas Nijholt authored Nov 22, 2018
  
  d5a63a3f
- add newline · b76ac11c
  Bas Nijholt authored Nov 22, 2018
  
  b76ac11c
- update 'uniform_sampling_1d' example in the docs · 84cba2ee
  Bas Nijholt authored Nov 22, 2018
  
  84cba2ee
- Merge branch '119-add-second-order-loss-to-adaptive' into 'master' · 7408ed57
  Bas Nijholt authored Nov 22, 2018
```
Resolve "(Learner1D) add possibility to use the direct neighbors in the loss"

Closes #119

See merge request !131
```
  7408ed57
- change loss function signature · fde1774b
  Jorn Hoofwijk authored Nov 19, 2018 and Bas Nijholt committed Nov 22, 2018
  
  fde1774b
- fix broken tests and use the new API · 27331280
  Jorn Hoofwijk authored Nov 22, 2018 and Bas Nijholt committed Nov 22, 2018
  
  27331280