Skip to main content
. 2024 Mar 26;5(5):100964. doi: 10.1016/j.patter.2024.100964

Table 2.

Result of ESS-MW with various thresholds on House100K environment

Training stage
Pretext task
Downstream ImageNet classification
Threshold Training loss ↓ Training loss ↓ Test loss ↓ Test accuracy (%) ↑
(0.8, 12) 3.92 ± 0.003 4.62 ± 0.001 4.67 ± 0.004 18.39 ± 0.082
(0.4, 6) 3.77 ± 0.004 4.98 ± 0.326 4.67 ± 0.024 18.03 ± 0.114
(1.6, 24) 4.26 ± 0.003 4.72 ± 0.011 4.82 ± 0.013 17.61 ± 0.135