view article
Figure 6
Performance of the different algorithms (Table 1[link]: rows 3, 8, 9, 12, 17 and 18) on mixed crystal context. Algorithms resulting from both primary (G50 or H50, panels A and D, respectively) and incremental learning trainings (G50 > H50, panels B and C; or H50 > G50, panels E and F) were tested on images containing both targets (even coexisting in the same cell). Secondary trainings were performed using 10 epochs and learning rates of 1 × 10−3 (panels B and E) or 1 × 10−4 (panels C and F). The H50 > G50 algorithm seems to require higher learning rates for adapting to the secondary target (losing in turn adaptation to its primary target). Performance indicator values are shown in Table 1[link].

Journal logoJOURNAL OF
APPLIED
CRYSTALLOGRAPHY
ISSN: 1600-5767
Volume 57| Part 2| April 2024| Pages 266-275
https://doi.org/10.1107/S1600576724000682
Follow J. Appl. Cryst.
Sign up for e-alerts
E-alerts
Follow J. Appl. Cryst. on Twitter
Twitter
Follow us on facebook
Facebook
Sign up for RSS feeds
RSS