Super-Resolution on SR RFs = 4
25.88PSNRN-DIPA
Evaluation results
| Method | Links | |
|---|---|---|
| N-DIPALearning=true, Preconditioning model=P_θ⋆ (13)2026.05 | 25.88 | |
| L-DIPALearning=true, Preconditioning model=P⋆ (13)2026.05 | 25 | |
| Full-linearLearning=true, Preconditioning model=P_k2026.05 | 22.02 | |
| ConvolutionalLearning=true, Preconditioning model=p_k * x2026.05 | 17.55 | |
| PolynomialLearning=false, Preconditioning model=p(A⊤A)2026.05 | 11.81 | |
| HessianLearning=false, Preconditioning model=(A⊤A)⁻¹2026.05 | 11.75 | |
| Scalar stepLearning=true, Preconditioning model=p_k I_n2026.05 | 11.62 | |
| PointwiseLearning=true, Preconditioning model=p_k ⊙ x2026.05 | 11.19 | |
| BaselineLearning=false, Preconditioning model=I_n2026.05 | 11.14 |