AdaFlow demonstrates a training-free method to edit more than 1,000 video frames in one inference on a single A800 GPU via adaptive attention token slimming and content-aware keyframe selection.
Please add a caption to the video in great detail
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
AdaFlow: Efficient Long Video Editing via Adaptive Attention Slimming And Keyframe Selection
AdaFlow demonstrates a training-free method to edit more than 1,000 video frames in one inference on a single A800 GPU via adaptive attention token slimming and content-aware keyframe selection.