VideoJaNai

• Make backend installation more robust by using HTTP/2 and supporting multiple retries before failing download

• Update to vsmlrt v15.2 (TensorRT 10.3.0)
• Fix upscaling error due to blank RIFE model selected (#26)
• Improve error handling for backend setup
• Add RIFE 4.21 and 4.22 models

• Rename project from AnimeJaNaiConverterGui to VideoJaNai. Please uninstall old AnimeJaNaiConverterGui versions manually, as they are not automatically replaced by this update due to the name change.
• Redesigned app structure to separate backend Python/VapourSynth environment and models from the GUI, which are now downloaded and set up when launching the GUI for the first time.
• New design reduces app file size and improves app startup time.
• Add workflows to allow having multiple preset settings.
• Add auto and manual TensorRT engine generation settings, to support more model types such as SwinIR and DAT
• Add all RIFE 4.x models up to 4.20
• Fix final resize factor (#11)
• Fix decimal format for non-English Windows (#12)
• Add -max_interleave_delta 0 to ffmpeg preset settings, to fix audio issues (#24)

NOTE: vsmlrt v14.test3 with TensorRT 9.2.0 is used by default for compatibility with older NVIDIA GPUs and better performance with RIFE. If you wish to use the latest version of vsmlrt with TensorRT 10, you can manually update the vsmlrt install by downloading the latest vsmlrt windows cuda package and extracting to %APPDATA%\VideoJaNai\python\vs-plugins.

• Rename project from AnimeJaNaiConverterGui to VideoJaNai. Previous versions are not automatically replaced by this update and must be uninstalled manually due to the name change.
• Redesigned app structure to separate backend Python/VapourSynth environment and models from the GUI, which are now downloaded and set up when launching the GUI for the first time.
• New design improves app startup time.
• Add workflows to allow having multiple preset settings.
• Add TensorRT engine generation settings
• Fix final resize factor (#11)
• Add -max_interleave_delta 0 to ffmpeg preset settings, to fix audio issues (#24)

• Fix failure to upscale due to missing prepackaged ffmpeg

• TensorRT: fix bug preventing generation of dynamic shapes engines (thanks @kim2091 and @sirosky)

• TensorRT: generate static shapes engine if dynamic shapes engine generation fails (#2)
• Update vsmlrt to v14.test3 with TensorRT v9.2.0
• Fix bug where videos larger then 1080p are force downscaled to 1080p (#6)
• Fix bug where RIFE interpolation causes conversion to fail, add options to select interpolation factor, RIFE model, scene detection threshold (#5)
• Update app theme with FluentAvalonia
• Update installer - migrate from Squirrel Clowd to Velopack

• Update theme
• Reorganize input and output fields
• Add status bar
• Fix performance issue when console is overloaded
• Add installable version and test auto update

Note: starting with version 0.0.6, a portable and installable version are both available. Choose one or the other. Functionally, they are equivalent, with the following differences which are typical between installable and portable apps.

AnimeJaNaiConverterGuiSetup.exe:

• Can be installed and uninstalled
• Has an option to automatically update to new releases, enabled by default
• Adds shortcuts to desktop and start menu once installed

AnimeJaNaiConverterGui-0.0.6-Portable.7z:

• Extract to any folder of your choice
• Run AnimeJaNaiConverterGui.exe inside of the folder you extracted to to run the program
• Does not install to your PC and does not add shortcuts to desktop or start menu
• Does not have any uninstall option, simply delete all of the files to remove
• Does not auto update on its own, updates must be downloaded from GitHub manually

• fix engine generation output not appearing in console
• add settings to resize video after all models
• update app icon

• add drag and drop support for input/output textboxes
• add cancel upscale button
• kill ffmpeg when closing app

Incorporate bug fixes and enhancements suggested by @Zarxrax :

• use fixed text box length to avoid issues with long file paths
• automatically append extension to output filenames
• add option to overwrite existing files
• instead of opening new console window, print console context to screen and also log to error.log file in the app directory
• add inference backend selection with TensorRT, DirectML, and NCNN options

• persist gui settings

• basic single and batch upscaling