search

skinniderlab / CLM

MIT License
0 stars 0 forks source link

% novel/% valid #168

Open skinnider opened 4 months ago

skinnider commented 4 months ago

calculate_outcomes is being run on the output of collect_tabulated_molecules, which means % novel/% valid are constant and 1. We should either (1) provide separately as input a single sample of 500k "raw" SMILES (without removing invalid ones and canonicalizing) or (2) just remove these plots

vineetbansal commented 4 months ago

Note: --max_molecules 500000 in rule prep_outcome_freq