AI Upscale stands on the work of these authors. Thank you!
Models and ffmpeg are NOT bundled: the installer downloads them from the official pages below.

AI MODELS
  Real-ESRGAN (general-x4v3, general-wdn-x4v3, x4plus, x4plus_anime_6B) - Xintao Wang et al.
      BSD-3-Clause - https://github.com/xinntao/Real-ESRGAN
  APISR RRDB 2x / RRDB 4x / DAT 4x - Boyang Wang et al. (CVPR 2024)
      GPL-3.0 - https://github.com/Kiteretsu77/APISR
  2x AnimeJaNai HD V3 / SD V1beta34 (Compact) - the-database
      CC-BY-NC-SA-4.0 (non-commercial) - https://github.com/the-database/mpv-upscale-2x_animejanai
  2x AniScale 2 ESRGAN-Lite - Sirosky
      CC-BY-NC-4.0 (non-commercial) - https://github.com/Sirosky/Upscale-Hub
  2x LiveAction V1 SPAN - jcj83429
      CC-BY-NC-SA-4.0 (non-commercial) - https://github.com/jcj83429/upscaling
  4x Nomos2 hq MoSR - Philip Hofmann (Helaman)
      CC-BY-4.0 - https://github.com/Phhofm/models
  2x CGIMaster v1 - Madiator
      GPL-3.0 - https://openmodeldb.info/models/2x-CGIMaster-v1
  RIFE 4.26 frame interpolation - Zhewei Huang (hzwer) et al.
      MIT - https://github.com/hzwer/Practical-RIFE  (weights via https://github.com/HolyWu/vs-rife)
  CLIP ViT-B/32 (content-type detection) - OpenAI
      MIT - https://huggingface.co/openai/clip-vit-base-patch32

CODE
  engine/rife/ifnet_v4_26.py, warplayer.py - from vs-rife by HolyWu, MIT (see engine/rife/LICENSE)
  spandrel (model loading) - chaiNNer-org, MIT - https://github.com/chaiNNer-org/spandrel
  PyTorch - BSD-3-Clause; TensorRT - NVIDIA (proprietary, installed from PyPI under NVIDIA's license)
  Triton for Windows - MIT; Transformers - Apache-2.0; pywebview - BSD-3-Clause
  ffmpeg (BtbN Windows build) - GPL-3.0 - https://github.com/BtbN/FFmpeg-Builds
  uv (installer) - Astral, MIT/Apache-2.0 - https://github.com/astral-sh/uv
