Closed adriangb closed 2 years ago
Merging #278 (1a02a26) into master (d50e75a) will decrease coverage by
0.13%
. The diff coverage is100.00%
.
@@ Coverage Diff @@
## master #278 +/- ##
==========================================
- Coverage 98.27% 98.14% -0.14%
==========================================
Files 7 7
Lines 754 755 +1
==========================================
Hits 741 741
- Misses 13 14 +1
Impacted Files | Coverage Δ | |
---|---|---|
scikeras/wrappers.py | 97.51% <100.00%> (+<0.01%) |
:arrow_up: |
scikeras/_saving_utils.py | 96.73% <0.00%> (-1.09%) |
:arrow_down: |
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 d50e75a...1a02a26. Read the comment docs.
Spoke with the scikit-learn team w.r.t lowering precision of their estimator checks so that we don't have to do juggling like we did here with sample weights. They're working on it, but nothing concrete so far. For now, I think our best option is to move forward with this. I will think about making a PR to their repo and then clean stuff up here (the whole separation into high and low precision falls into that category as well)