PKU-YuanGroup/MagicTime
A diffusion-based text-to-video model that generates time-lapse videos by simulating metamorphic physical changes over time.
Not currently ranked — collecting fresh signals.
star history
MagicTime is a time-lapse video generation model published in IEEE TPAMI 2025. It extends text-to-video diffusion models to simulate metamorphic physical transformations, such as objects melting, rusting, or decaying over time. The model processes temporal deformation to produce photorealistic long-form time-lapse sequences from textual descriptions, and includes a dedicated dataset called ChronoMagic for training.
Frequently asked
- What is PKU-YuanGroup/MagicTime?
- A diffusion-based text-to-video model that generates time-lapse videos by simulating metamorphic physical changes over time.
- Is MagicTime open source?
- Yes — PKU-YuanGroup/MagicTime is open source, released under the Apache-2.0 license.
- What language is MagicTime written in?
- PKU-YuanGroup/MagicTime is primarily written in Python.
- How popular is MagicTime?
- PKU-YuanGroup/MagicTime has 1.3k stars on GitHub.
- Where can I find MagicTime?
- PKU-YuanGroup/MagicTime is on GitHub at https://github.com/PKU-YuanGroup/MagicTime.