List your company
ResearchFormerly WizWand

Image Restoration on Urban100

45.7PSNRMTS-CSNet
15.976823.693431.4139.1266Dec 12, 2023Apr 20, 2024Aug 29, 2024Jan 7, 2025May 18, 2025Sep 26, 2025Feb 4, 2026
Updated 8mo ago

Evaluation results

MethodLinks
MTS-CSNetCompression Ratio=50%, T=48, NB=32026.0245.799.42
MTS-CSNetCompression Ratio=50%, T=12, NB=32026.0243.1299.18
MTS-CSNetCompression Ratio=50%, T=24, NB=32026.0240.8998.93
IDMCompression Ratio=50%2026.0239.1697.71
MTS-CSNetCompression Ratio=30%, T=24, NB=32026.0236.8397.19
MTS-CSNetCompression Ratio=Avg., T=48, NB=32026.0236.694.7
MTS-CSNetCompression Ratio=30%, T=12, NB=32026.0236.597.09
MTS-CSNetCompression Ratio=Avg., T=12, NB=32026.0236.0594.67
IDMCompression Ratio=30%2026.0235.7895.7
MTS-CSNetCompression Ratio=Avg., T=24, NB=32026.0235.5494.86
IDMCompression Ratio=Avg.2026.0235.2694.37
MTS-CSNetCompression Ratio=30%, T=48, NB=32026.0235.1496.14
IDMCompression Ratio=10%2026.0230.8589.7
MTS-CSNetCompression Ratio=10%, T=48, NB=32026.0228.9588.54
MTS-CSNetCompression Ratio=10%, T=24, NB=32026.0228.9188.46
MTS-CSNetCompression Ratio=10%, T=12, NB=32026.0228.5387.73
Full-ft#param=11.6M, Pre-trained model=EDT2023.1225.1160.96
AdaptIR#param=170K, Pre-trained model=EDT2023.1225.0460.57
AdaptFor.#param=162K, Pre-trained model=EDT2023.1225.0260.48
Adapter#param=168K, Pre-trained model=EDT2023.1225.0160.39
LoRA#param=155K, Pre-trained model=EDT2023.122560.38
FacT#param=174K, Pre-trained model=EDT2023.122560.41
SSF#param=117K, Pre-trained model=EDT2023.1224.9860.27
VPT#param=311K, Pre-trained model=EDT2023.1224.7559.02
PretrainPre-trained model=EDT2023.1224.0855.8
DDNMCompression Ratio=10%2026.0220.7646.82
GDPCompression Ratio=10%2026.0220.7450.75
ΠGDMCompression Ratio=10%2026.0220.0950.89
PSLDCompression Ratio=10%2026.0219.4340.54
DDRMCompression Ratio=10%2026.0219.1643.48
SR3Compression Ratio=10%2026.0218.950.23
DPSCompression Ratio=10%2026.0217.1232.7