Table 8 Datasets for conditional generation
From: Probabilistic generative transformer language models for generative design of molecules
Whole set | Training set (Top 50%) | Generated samples | |
|---|---|---|---|
LogP | 1,584,662 | 792,331 | 16,748 |
tPSA | 1,584,662 | 792,331 | 16,643 |
QED | 1,584,662 | 792,331 | 17,082 |