Skip to main content
. 2024 May 5;11(26):2400829. doi: 10.1002/advs.202400829

Table 4.

Performance of generative neural network (the up‐arrows denote that the higher scores are considered better, the down‐arrows mean that lower scores are better).

Metrics Pre‐training W/O Self‐assembling Small Molecule W/O Pre‐training HydrogelFinder‐GPT
FCD (↓) 33.7017 7.1341 5.3607 4.9586
SNN (↑) 0.2153 0.5936 0.4730 0.5149
Frag (↑) 0.6705 0.9952 0.9947 0.9957
Scaff (↑) 0.1388 0.5576 0.4866 0.4821