Many preview images on this site are currently broken due to their original hosts being indefinitely offline. Work is underway to recover the affected images and rehost them somewhere reliable.

SR
LR
BeforeAfter
100%
Scroll to zoom · Drag to pan
Thumbnail
1 / 2

Book Compact

1xCompactby asterixcool

Good for

RestorationText

Purpose: "binarization" for scanned books.

A compact, lightweight version trained for general-purpose cleanup. While it performs well, RealPLSKR delivers superior results for more complex cases.

  • Strengths (Handles Well): ✔ Color bleed ✔ Spots & small speckles ✔ Mould stains ✔ Fine hair & scratches – Good at thin, faint lines. ✔ Light pencil marks ✔ Dividers & text preservation
  • Weaknesses (Struggles With): ✖ Large speckles ✖ Thick hair/clumps – May leave remnants or blurring. ✖ Photographs – Bad lighting; noisy/dark images degrade quality.
Training details (6)
Date
2025-06-04
Dataset
My own collection.
Dataset size
13150
Training iterations
399000
Training batch size
16
Training HR size
95
How to upscale →Understanding licenses →
Download (2.3 MB)
PyTorch .safetensors file / Hosted by GitHub

Model

Architecture
Compact
Scale
1x
Size
2,27mb
Color Mode

Rights

CC-BY-SA-4.0
Private use
Commercial use
Distribution
Modifications
Credit required
Same License
State Changes
No Liability & Warranty
Disclaimer

Similar models