Releases: DeepLearnPhysics/lartpc_mlreco3d
Releases · DeepLearnPhysics/lartpc_mlreco3d
v2.4
- Support doublets in charge rescaling
- Store rescaled charge in output
- Depositions changed to rescaled charge in analysis tools
v2.3
- Dae Heun modified the edge accuracy definition for GraphSpice. (see PR #101)
- GraphSpice does not rely on
eval
configuration parameter anymore (see PR #102)
- Switch to Github Actions for CI (see PR #106)
- GrapPA primary prediction high purity fix from Francois (see PR #107)
- Script in
bin
folder to check validity of new data samples (see PR #108 )
- New parser
parse_sparse3d_ghost
from Francois (see PR #108 )
- New metric script
doublets_metrics
to assess deghosting performance with doublets (see PR #108 )
v2.2
- Updates to stopping muon / through going muon / Michel selections
concat_result
and unwrap
no longer required in configurations
parse_meta3d
new parser to access meta information
- Documentation effort including for analysis tools
- Small changes to analysis tools
v2.1
Most notable updates include:
- fix for
adapt_labels
memory issue (Dae Heun)
- update to analysis tools with comments made during Icarus ML meeting (Dae Heun)
- through-going and stopping muons selection examples using analysis tools (Laura)
v2.0 (ME)
Tagging first working and stable state of the codebase using ME library before we add more features in the near future.
v1.0-scn
Keeping a tab on latest SCN codebase before switching to ME in case we ever need it in the future.