v7.66.0

SciMLSensitivity v7.66.0

Diff since v7.65.0

Merged pull requests:

• Fake solver output type to please Enzyme (#1098) (@ChrisRackauckas)
• Fix null parameters with ForwardDiffSensitivity (#1104) (@ChrisRackauckas)

v7.65.0

SciMLSensitivity v7.65.0

Diff since v7.64.0

Merged pull requests:

• Check p while choosing default sensealg (#1079) (@DhairyaLGandhi)
• Update Forward Mode to use SciMLStructures (#1085) (@DhairyaLGandhi)
• CompatHelper: bump compat for DataInterpolations to 6 for package docs, (keep existing compat) (#1086) (@github-actions[bot])
• Remove duplicate callbacks in ForwardDiff and ReverseDiff adjoints (#1087) (@jClugstor)

Closed issues:

• Including callback in problem vs solve call changes gradient for ForwardDiffSensitivity (#1081)

v7.64.0

SciMLSensitivity v7.64.0

Diff since v7.63.1

Merged pull requests:

• add tests for Enzyme frontend for nonlinear problem (#1066) (@m-bossart)
• docs: Make example code consistent with text (#1080) (@moble)

v7.63.1

SciMLSensitivity v7.63.1

Diff since v7.63.0

Merged pull requests:

• Revert breakage in GPU (#1078) (@ChrisRackauckas)

Closed issues:

• SciMLSensitivity 7.63 leads to scalar indexing on CUDA (#1077)

v7.63.0

SciMLSensitivity v7.63.0

Diff since v7.62.0

Merged pull requests:

• fix: fix incorrect usage of getu resulting in flattened arrays (#1076) (@AayushSabharwal)

Closed issues:

• GaussAdjoint fails on a problem with discrete callbacks (#1072)

v7.62.0

SciMLSensitivity v7.62.0

Diff since v7.61.1

Merged pull requests:

• Use the SciMLStructures Tunable interface (#1010) (@ChrisRackauckas)
• Update to using SII and SciMLStructures (#1057) (@DhairyaLGandhi)

Closed issues:

• Julia crashes when trying to optimize Graph NODE (#1070)

v7.61.1

SciMLSensitivity v7.61.1

Diff since v7.61.0

Merged pull requests:

• Hotfix GaussAdjoint enzyme on non-vector (#1069) (@ChrisRackauckas)

v7.61.0

SciMLSensitivity v7.61.0

Diff since v7.60.1

Merged pull requests:

• Update default ODE solver tests to use new OrdinaryDiffEq infrastructure (#1063) (@ChrisRackauckas)
• Do not require gradient for vjp choice (#1064) (@ChrisRackauckas)
• Properly zero memory (#1065) (@ChrisRackauckas)

Closed issues:

• BoundsError for polyalgorithm (#1035)

v7.60.1

SciMLSensitivity v7.60.1

Diff since v7.60.0

Merged pull requests:

• Fence definition of literal_getproperty adjoint (#1059) (@DhairyaLGandhi)

Closed issues:

• precompile fails with v7.60.0 (#1058)

v7.60.0

SciMLSensitivity v7.60.0

Diff since v7.59.0

Merged pull requests:

• chore: move literal_getproperty adjoint to SciMLSensitivity (#1056) (@DhairyaLGandhi)