Open Access

Table B.1

Additional accuracy diagnostics.

Dataset Arch mLAE (dex) LAE99 (dex) LAE100 (dex) Temp. LAE (dex)
Primordial MON 0.343 2.47 6.24 0.126
FCNN 0.094 0.693 3.47 0.030
LNODE 0.661 3.04 6.88 0.203
LP 0.843 3.35 7.07 0.27

Primordial MON 0.613 5.96 18.5 0.126
Parametric FCNN 0.259 1.74 14.5 0.060
LNODE 0.693 3.32 14.8 0.149
LP 0.743 3.43 14.2 0.17

Cloud MON 0.034 0.326 6.36 0.008
FCNN 0.035 0.384 4.33 0.005
LNODE 1.92 4.55 8.64 0.993
LP 1.64 4.70 8.36 0.497

Cloud MON 0.121 0.831 9.43 0.017
Parametric FCNN 0.244 2.80 16.9 0.019
LNODE 1.51 4.80 14.7 0.253
LP 1.37 4.51 13.5 0.274

Notes. In addition to mLAE and LAE99, we report the maximum log-space absolute error (LAE100) and the temperature-only absolute error (Temp. LAE). Best values per dataset are shown in bold.

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.