Ryan Curtin
59494a2f3e
Start a simple history file.
2018-10-20 00:41:38 -04:00
Ryan Curtin
5515bca6d6
Remove discussion comments from paper for release.
...
(If we need them back for future revisions, we can just check out the previous
revision and work on a branch or something.)
2018-10-20 00:40:33 -04:00
Ryan Curtin
c4f843cc42
Merge pull request #35 from conradsnicta/edit-19oct2018-e
...
bump to version 1.10; proposed name
2018-10-19 23:40:25 -05:00
Ryan Curtin
558f46c2d0
Final changes. Also updates the definition of partially differentiable.
2018-10-19 22:33:32 -04:00
Ryan Curtin
ff20f9ea67
Use much cleaner way of suprressing errors!
2018-10-19 15:14:23 -04:00
conrad
f837350248
bump to version 1.10; proposed name
2018-10-20 04:12:39 +10:00
Ryan Curtin
5008cadfbd
Fix categorical optimizers to use non-mlpack types.
...
Later we'll have to fix the mlpack code that uses them.
2018-10-19 12:02:18 -04:00
Ryan Curtin
5e757febe9
Strip references to mlpack.
...
Honestly the duplicate documentation for DecomposableFunctionType should be
removed when the FunctionType documentation is ready, and simply replaced with a
link to that. So that will happen soon.
2018-10-19 12:01:44 -04:00
Ryan Curtin
e9adce346a
Remove leftover mlpack includes from tests.
2018-10-19 11:09:48 -04:00
Ryan Curtin
39d26179de
Ha, yeah, it was an abomination....
...
Turns out, an unnecessary one too.
2018-10-19 11:07:41 -04:00
Ryan Curtin
9f273b001a
Enable ParallelSGD and autodetect OpenMP.
2018-10-19 11:06:24 -04:00
Ryan Curtin
8c1d5af325
Merge pull request #34 from conradsnicta/edit-19oct2019-d
...
bug fixes: use correct include header guard; use ENS_ prefix for macros
2018-10-19 14:45:45 +00:00
Ryan Curtin
05edc0f5e9
Merge pull request #33 from conradsnicta/edit-19oct2018-c
...
use consistent license blurbs
2018-10-19 14:44:52 +00:00
Ryan Curtin
85821ff200
Merge pull request #32 from conradsnicta/edit-19oct2018-b
...
improved README
2018-10-19 14:44:15 +00:00
Ryan Curtin
ae80819c32
Build and run the tests directly in the build directory.
2018-10-19 10:42:57 -04:00
conrad
d7c1df3f52
bug fixes: use correct include header guard; use ENS_ prefix for macros to avoid potential clashes with mlpack and other software which may use same macro names
2018-10-20 00:13:40 +10:00
conrad
2db0505b85
add list of developers/contributors
2018-10-19 23:43:27 +10:00
conrad
7786ead2c4
improved README: add link to main website and list requirements
2018-10-19 23:34:00 +10:00
Ryan Curtin
c0b6a9b947
Remove unneeded functions.
2018-10-19 09:32:27 -04:00
Ryan Curtin
f6f5f5e710
Add a simple block for setting random seed.
2018-10-19 09:32:27 -04:00
Ryan Curtin
6dfce5b6ae
Suppress warnings from another test.
2018-10-19 09:32:27 -04:00
Ryan Curtin
1fff066999
Adjust tolerances significantly for occasionally failing tests.
...
I observed that the failure probabilities were too high even in the original
mlpack code.
2018-10-19 09:32:27 -04:00
Ryan Curtin
065fbc51a6
Hide scary-looking Armadillo output.
...
(Nothing is wrong with the test, it's expected behavior.)
2018-10-19 09:32:27 -04:00
Ryan Curtin
689d6dca3e
Allow a failure in SPALeRA test and retry.
2018-10-19 09:32:27 -04:00
conrad
f2a851c7e2
use consistent license blurbs
2018-10-19 23:09:06 +10:00
Ryan Curtin
23c8f87adf
Merge pull request #31 from conradsnicta/edit-19oct2018-a
...
spelling fix in abstract: typicalle -> typically
2018-10-19 04:41:08 +00:00
conrad
0d0a876a1b
spelling fix: typicalle -> typically
2018-10-19 12:29:07 +10:00
Ryan Curtin
64b97590ea
Revisions after a full readthrough. I am happy with this.
2018-10-18 21:44:31 -04:00
Marcus Edel
1eb6462271
Add timings.
2018-10-19 02:09:46 +02:00
Marcus Edel
e7f08f3d6e
Add benchmark system specs.
2018-10-18 23:01:48 +02:00
Ryan Curtin
1c56b56d1e
This test was barely failing, so increase the tolerance.
...
I've reviewed the code and I can't find anything different from mlpack, but I am
thinking that perhaps this might have to do with using different CSV readers (I
only say that because I can't think of anything else, and I'm pretty certain no
code is broken).
2018-10-18 12:25:08 -04:00
Ryan Curtin
0f4de04d24
Oops, I read the mlpack code wrong! Covariance is I not 3I.
2018-10-18 12:24:48 -04:00
Ryan Curtin
18ca77c90a
Fix test compilation; set max iterations.
...
Otherwise the tests could run indefinitely...
2018-10-18 11:58:44 -04:00
Ryan Curtin
9517cc18ef
Merge pull request #28 from conradsnicta/edit-18oct2018-a
...
fixes and improvements
2018-10-18 15:10:56 +00:00
Ryan Curtin
753def1a1c
Merge pull request #30 from conradsnicta/check-armadillo-version
...
explicitly check for suitable armadillo version in case ensmallen is used without the cmake based installation
2018-10-18 15:08:08 +00:00
Ryan Curtin
11246837d6
Start minor versions at 10.
2018-10-18 11:06:57 -04:00
Ryan Curtin
f0fb8a512f
Clarify RC versioning for #29 .
2018-10-18 11:06:07 -04:00
conrad
cfb3415e89
fix
2018-10-19 00:43:28 +10:00
conrad
c020d1fc79
explicitly check for suitable armadillo version in case ensmallen is used without the cmake based installation
2018-10-19 00:37:46 +10:00
conrad
e6137248c4
more fixes for references
2018-10-18 19:14:52 +10:00
conrad
50f153fea5
remove junk
2018-10-18 18:37:43 +10:00
conrad
99903d9fd3
fixes for references
2018-10-18 18:36:37 +10:00
conrad
5fa33dc20e
added citations to many optimisation algorithms
2018-10-18 18:13:18 +10:00
conrad
3ab590d21e
remove blank space around the plot in figure 2(b)
2018-10-18 16:26:23 +10:00
conrad
eda8844297
remove blank space around the plot in figure 2(b)
2018-10-18 16:25:57 +10:00
conrad
264840d453
better flow for mentioning mlpack
2018-10-18 16:12:46 +10:00
conrad
24cbacf52f
fit into 6 pages
2018-10-18 15:52:57 +10:00
conrad
845fd8714b
minor clarifications and fixes
2018-10-18 15:52:03 +10:00
conrad
a73173ded1
fix and clarification: avoid confusion between the name 'ensmallen' and the experiment named 'ensmallen'
2018-10-18 15:40:26 +10:00
Ryan Curtin
51dbf63a19
Some basic contribution guidelines.
2018-10-18 01:02:08 -04:00
Ryan Curtin