@misc{indiciae16fdbd010645,
  title = {VideoFlexTok: Flexible-Length Coarse-to-Fine Video Tokenization},
  author = {Andrei Atanov and Jesse Allardice and Roman Bachmann and Oğuzhan Fatih Kar and R Devon Hjelm and David Griffiths and Peter Fu and Afshin Dehghan and Amir Zamir},
  year = {2026},
  url = {https://arxiv.org/abs/2604.12887},
  note = {Source identifier: 2604.12887}
}