Flame: an open source framework for model development, hosting, and usage in production environments

Table 5 Computation time for series of diverse size

Series	Original series size	Compounds removed^a	Final series size	Time (s)^b
A	2685	11	2674	20 s
B	5832	0	5832	32 s
C	126,368	114	126,254	600 s (10 min)
D	480,000	249	479,751	2160 s (36 min)

Computation tasks involving structure normalization, computation of RDKit descriptors, generation of a XGBoost model, and validation using fivefold validation (for series A and B) and twofold validation (for series C and D).
^aCompounds removed from the computation because RDKit was unable to compute molecular descriptors
^bWall clock times, in a desktop PC with Windows 10 professional 64 bits, 32b Gb RAM and an AMD Ryzen 5 3600(6 cores) CPU

ISSN: 1758-2946