Closed maho3 closed 10 months ago
Attention: 22 lines
in your changes are missing coverage. Please review.
Comparison is base (
f1e8b80
) 97.27% compared to head (42208b1
) 96.49%. Report is 1 commits behind head on main.
Files | Patch % | Lines |
---|---|---|
ili/utils/ndes_pt.py | 93.33% | 9 Missing :warning: |
ili/dataloaders/loaders.py | 73.68% | 5 Missing :warning: |
ili/inference/runner_lampe.py | 97.48% | 4 Missing :warning: |
ili/inference/runner_pydelfi.py | 33.33% | 2 Missing :warning: |
ili/utils/samplers.py | 80.00% | 2 Missing :warning: |
:exclamation: Your organization needs to install the Codecov GitHub app to enable full functionality.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
This PR addresses #124 . It consists of the following changes:
LampeRunner
in ili/inference/runner_lampe.py, with a configuration similar to the pydelfi and sbi runners. It only does NPE (no SNPE), but can use a wide variety of flows from zuko.LampeNPE
andLampeEnsemble
to ili/utils/ndes_pt.py, to mirror similar NDE implementations in sbi (with sampling methods and weighting etc.)TorchLoader
to ili/dataloaders. This allows one to simply give LampeRunner custom train/val torch Dataloaders, which is necessary in the GNN example.Still todo:
TorchLoader