VSGAN Versions Save

PyTorch-based Super-Resolution and Restoration Image Processing Module for VapourSynth

v1.0.6

2 years ago

Added

  • Detect ESRGAN old/new arch models via archaic trial-and-error.

v1.0.5

2 years ago

Changed

  • Reworked code from Functional to Object-oriented Programming.
  • Improve code readability, project starting to get serious.

v1.0.4

2 years ago

Added

  • Add ability to tile the input to reduce VRAM (does not hide seams).

v1.0.3

2 years ago

Added

  • VapourSynth to requirements.

Changed

  • Convert back to original color-space after applying the model.

v1.0.2

2 years ago

Added

  • Ability to select device via argument.

v1.0.1

2 years ago

Added

  • README file with some basic information.

Changed

  • Improved RGB conversion by using mvsfunc instead of core.resize.Point.

v1.0.0

2 years ago

Initial Release.

v1.5.0

2 years ago

Added

  • Add support for ESRGAN+ models, Real-ESRGAN models (including 2x and 1x if pixel-shuffle was used), and A-ESRGAN models.
  • Add support for Newer-New-arch in ESRGAN new-to-old state dict conversion.

Changed

  • Rework model/arch file system structure to /models, /models/blocks and /models/ESRGAN.
  • Rework ESRGAN architecture as a singular class, with all ESRGAN-specific operation done within it.
  • Move ESRGAN-specific blocks within ESRGAN.py.

Removed

  • Removed some unused blocks from RRDBNet.

Fixed

  • Ensure clip parameter of VSGAN is a VapourSynth VideoNode object (a clip).
  • Move RGB clip check to the constructor of VSGAN rather than run().

v1.4.1

2 years ago

Added

  • Created new sphinx documentation, replacing the old Jekyll documentation.
  • Added HISTORY.md file for recording history (now CHANGELOG.md).

Changed

  • Reword some error/warning messages, now less opinionated and more concise.
  • Some attributes have been renamed to be more ambiguous in the hopes more Model Architectures get supported in the future.

Fixed

  • Fix model chaining. It now gets the correct model and model scale values for each FrameEval call.
  • Fixed the pytorch extra group to correctly be optional and correctly reference a dependency.
  • Some type-hinting has been corrected.

v1.4.0

2 years ago

Added

  • Added support for all RGB formats including float.

Changed

  • Heavily improved main model execution code.
  • Replace current chunk system with a seamless chunk system using overlap.
  • Add self-chaining system, calls can be made directly after another.
  • Made torch dependency optional and pointed directly to torch+cuda. This is due to conflicting kinds of torch installation methods.

Removed

  • Remove JetBrains .idea folder, added to gitignore.

Fixed

  • Only transpose C for RGB if it's 3-channels.