search

LabeliaLabs / distributed-learning-contributivity

Simulate collaborative ML scenarios, experiment multi-partner learning approaches and measure respective contributions of different datasets to model performance.
https://www.labelia.org
Apache License 2.0
56 stars 12 forks source link

LFlip refacto in S-model #281

Closed Thomas-Galtier closed 3 years ago

Thomas-Galtier commented 3 years ago

S-model is a method like LFlip but it is more robust to random noise

bowni commented 3 years ago

Good for me, apart from a secondary question I raised ☝️ . Also, just saw that it requires some conflicts to be fixed!

codecov-io commented 3 years ago

Codecov Report

Merging #281 (8251441) into master (b15cf88) will decrease coverage by 0.40%. The diff coverage is 37.08%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master     #281      +/-   ##
==========================================
- Coverage   43.61%   43.20%   -0.41%     
==========================================
  Files          11       12       +1     
  Lines        2380     2465      +85     
==========================================
+ Hits         1038     1065      +27     
- Misses       1342     1400      +58
Impacted Files Coverage Δ
mplc/constants.py 100.00% <ø> (ø)
mplc/mpl_utils.py 36.75% <0.00%> (ø)
mplc/contributivity.py 6.68% <6.25%> (+0.42%) :arrow_up:
mplc/multi_partner_learning.py 34.94% <23.52%> (+1.27%) :arrow_up:
mplc/models.py 36.61% <36.61%> (ø)
mplc/partner.py 80.55% <50.00%> (-6.12%) :arrow_down:
mplc/scenario.py 68.09% <66.66%> (+0.07%) :arrow_up:
mplc/dataset.py 82.88% <100.00%> (+6.30%) :arrow_up:

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update b15cf88...8251441. Read the comment docs.