VQGAN-LC (16x16)

Reported on 4 benchmarks across 1 task · 1 paper · 1 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Medical4 results

Image ReconstructiononImageNet
PSNR· 2024-06-17
23.8
best: 27.57 (OptVQ (16x16x8))
SOTA
Scaling the Codebook Size of VQGAN to 100,000 with a Utilization Rate of 99%arXiv:2406.11837
Image ReconstructiononImageNet
FID· 2024-06-17
2.62
best: 0.49 (MGVQ (16x16x8))
Scaling the Codebook Size of VQGAN to 100,000 with a Utilization Rate of 99%arXiv:2406.11837
Image ReconstructiononImageNet
LPIPS· 2024-06-17
0.12
best: 0.066 (OptVQ (16x16x8))
Scaling the Codebook Size of VQGAN to 100,000 with a Utilization Rate of 99%arXiv:2406.11837
Image ReconstructiononImageNet
SSIM· 2024-06-17
0.589
best: 0.787 (MGVQ (16x16x8))
Scaling the Codebook Size of VQGAN to 100,000 with a Utilization Rate of 99%arXiv:2406.11837