CLD track/cluster datasets use Key4HEP/EDM4hep production at 365 GeV. The schema version is 3.3.0, and the standard model recipe is pyg-cld-v1.
Physics samples¶
| Physics sample | TFDS dataset | Training selection |
|---|---|---|
| Top-quark pairs | cld_edm_ttbar_pf | Included in default recipe |
| Fully hadronic WW | cld_edm_ww_fullhad_pf | Included in default recipe |
| Inclusive quark pairs | cld_edm_qq_pf | Included in default recipe |
| ZZ | cld_edm_zz_pf | Additional builder; add to a custom recipe |
The recipes configure partitions 1--10; the verified public selection provides configuration 1. Use --data_config 1 for the download-based training path. The default recipe combines ttbar, WW, and qq; their ordering and relative exposure are described in Dataset mixing.
Event view¶
These views show the same simulated ttbar event (index 0) at 365 GeV in the transverse plane. The particle panel shows the Pandora reference reconstruction.
| Tracks and clusters | Pandora reconstructed particles |
|---|---|
 |  |
All-component overlay, vector input view and vector particle view are available for reuse. See event-view conventions for particle sources, sampling and scale.
Production and related inputs¶
Use shared Key4HEP production with PROD=cld and build this representation with pixi run tfds. For tracker and calorimeter measurements, use the separate CLD hit datasets.