DeepTFUS reproduction Collection Open reproduction of the DeepTFUS architecture (Srivastav et al., 2025): base model + 5 fine-tune variants exploring focal-position losses. • 6 items • Updated about 17 hours ago