f-dangel / curvlinops / 8972983757

Builds Branch Commit Type Ran Committer Via Coverage
8972983757 fix-kfac-scale Minor review fixes Pull #110 runame github
88.3
8961239818 fix-kfac-scale Merge branch 'main' into fix-kfac-scale Pull #110 runame github
88.3
8961240104 fix-kfac-scale Merge efce74cfe into d43e884ae Pull #110 web-flow github
88.3
8961225600 main [FIX] Scaling of empirical/MC Fisher for output with more than two dimensions (#109) * Fix EF scale for >2d outputs * Fix scale of MC Fisher for >2d outputs * Increase tolerance of CG test * Increase CG precision * Increase tolerance ... push web-flow github
88.46
8961022604 fix-scale Fix tests Pull #109 runame github
88.46
8951616005 fix-kfac-scale Fix darglint and flake8 Pull #110 runame github
88.29
8951434613 fix-kfac-scale Change how _num_per_example_loss_terms is inferred and allow setting it explicitly Pull #110 runame github
88.28
8951434804 fix-kfac-scale Merge 5bfd1c2dc into 7d498c317 Pull #110 web-flow github
88.28
8916088672 fix-scale Merge 59d71649c into 7d498c317 Pull #109 web-flow github
88.45
8916001309 fix-scale Increase tolerance of KFAC (log)det tests push runame github
88.45
8840487861 main [ADD] `LSMRInverseLinearOperator` (#106) * Add tests for LSMRInverseLinearOperator * Add LSMRInverseLinearOperator * Add LSMRInverseLinearOperator and more KFAC methods to docs * Add type hints to cg and lsmr hparams + update tol to rtol... push web-flow github
88.68
8840419980 lsmr-inverse Revert tol to rtol change Pull #106 runame github
88.68
8840420299 lsmr-inverse Merge f7a35981e into 30c77b4a2 Pull #106 web-flow github
88.68
8839315115 lsmr-inverse Merge cce65ff7c into 30c77b4a2 Pull #106 web-flow github
88.68
8839248041 lsmr-inverse Add LSMRInverseLinearOperator and more KFAC methods to docs push runame github
88.68