「view this page in B3 βῆτα server」

Revisions №62179

branch: rosetta:main 「№62179」
Commited by: Jason C. Klima
GitHub commit link: 「7eecbf982a05136a」
Difference from previous tested commit:  code diff
Commit date: 2025-07-28 14:57:51
linux.clang linux.gcc linux.srlz mac.clang
debug
release
unit
mac.clang.python38.PyRosetta.unit ubuntu.clang.cxx11thread.serialization.python38.PyRosetta.unit mpi mpi.serialization build.header build.levels build.ninja_debug m1.clang.python310.build.release m1.clang.python311.build.release m1.clang.python39.build.release beautification code_quality.cppcheck code_quality.merge_size serialization code_quality.submodule_regression integration.mpi integration.release_debug integration.tensorflow integration.thread integration.torch integration performance profile release.PyRosetta.MinSizeRel scientific.FlexPepDock scientific.abinitio_RosettaNMR_pcs scientific.glycan_dock scientific.peptide_pnear_vs_ic50.debug linux.clang.score linux.gcc.score mac.clang.score scripts.pyrosetta scripts.rosetta.parse scripts.rosetta.validate scripts.rosetta.verify m1.clang.python310.unit util.apps windows.build.debug windows.build.release

Add the deterministic_flag option to the ProteinMPNNProbabilitiesMetric SimpleMetric (#485) This PR adds the `deterministic_flag` option to the `ProteinMPNNProbabilitiesMetric` PerResidueProbabilitiesMetric. After some evidence provided in #429, the `ProteinMPNNMover` already has the `deterministic_flag` option set to `false` by default, and therefore due to Torch randomness the `ProteinMPNNProbabilitiesMetric` can return non-deterministic PSSM tables, even with the `-run:constant_seed 1` Rosetta option enabled. This PR just adds some control of the `ProteinMPNNMover` from the `ProteinMPNNProbabilitiesMetric` to enable deterministic mode when calculating per-residue ProteinMPNN probabilities.

...
Test: linux.gcc.mpi.serialization.integration.mpi

Failed sub-tests (click for more details):
replica_docking revold_vhts
Test: linux.clang.integration.release_debug_no_symbols

Failed sub-tests (click for more details):
features output_schema
Test: mac.clang.python39.integration

Failed sub-tests (click for more details):
output_schema