coveralls
commented
2 months ago Pull Request Test Coverage Report for Build 11409577418
Details
- 7 of 7 (100.0%) changed or added relevant lines in 1 file are covered.
- 56 unchanged lines in 2 files lost coverage.
- Overall coverage decreased (-3.7%) to 88.808%
| Files with Coverage Reduction | New Missed Lines | % | ||
|---|---|---|---|---|
| curvlinops/examples/utils.py | 10 | 31.25% | ||
| curvlinops/fisher.py | 46 | 26.47% | ||
| <!-- | Total: | 56 | --> |
| Totals | |
|---|---|
| Change from base Build 11001385384: | -3.7% |
| Covered Lines: | 1349 |
| Relevant Lines: | 1519 |
Resolves #132.
The problem was that the deterministic checks generate random numpy vectors which were then converted to the incorrect PyTorch data dtype.
Interestingly, I found that the other linear operators, such as Hessian, GGN, Fisher, EF, were not affected by this problem. I'm not 100% sure why. @runame any thoughts?