Skip to main content
. 2023 Sep 19;31(1):89–97. doi: 10.1093/jamia/ocad190

Figure 2.

Figure 2.

Dataset averaged F1 scores of FTsource + target, DAPT + FTsource + target, and DAPT + TAPT + FTsource + target with 10, 20, 30, 40, and 50 target domain samples included in fine-tuning.