Skip to content

Commit

Permalink
Merge in Victor's experiments (#10)
Browse files Browse the repository at this point in the history
* Batch v2 + summary resusts candidates in README + results as json

* WIP development of searching campaign

* Tweak campaign

* Add developments from train journey

* Some more refactoring and adding first tests

* More refactoring and linting

* Install optimade from github

* Update tests

* Download naccarato dataset in CI

* smaller kpoints grid

* Some acquisition functions + v4 dielectric calculation v1 candidates

* Add gitignore

* Removed unecessary metadata line

* Add first wave of results from human guided and fake active learning

---------

Co-authored-by: Victor Trinquet <victor.trinquet@uclouvain.be>
  • Loading branch information
ml-evs and VicTrqt authored Feb 9, 2024
1 parent f9673e5 commit 6a13d6a
Show file tree
Hide file tree
Showing 221 changed files with 2,488,410 additions and 10 deletions.
7 changes: 7 additions & 0 deletions .gitignore
Original file line number Diff line number Diff line change
Expand Up @@ -58,3 +58,10 @@ venv/
ENV/
.matplotlibrc
matplotlibrc

# data

*pkl*
*mdd*
*mod.*
*json*
1,209 changes: 1,209 additions & 0 deletions artificial/.ipynb_checkpoints/v0-checkpoint.ipynb

Large diffs are not rendered by default.

2,458 changes: 2,458 additions & 0 deletions artificial/.ipynb_checkpoints/v1-checkpoint.ipynb

Large diffs are not rendered by default.

6 changes: 6 additions & 0 deletions artificial/acquilib.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,6 @@
# Exploration - highest uncertainties
def exploration(predictions, uncertainties, **kwargs):
unc = uncertainties.copy()
target = unc.columns.values
unc['score'] = unc[target].rank(pct=True)
return unc
6 changes: 6 additions & 0 deletions artificial/dvlpt/acquilib.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,6 @@
# Exploration - highest uncertainties
def exploration(predictions, uncertainties, **kwargs):
unc = uncertainties.copy()
target = unc.columns.values
unc['score'] = unc[target].rank(pct=True)
return unc
Loading

0 comments on commit 6a13d6a

Please sign in to comment.