Enhance-A-Video is a project focused on improving video generation quality by adjusting the temporal attention parameters within video models to enhance consistency and visual quality between frames. This project is a collaborative effort among researchers from the National University of Singapore, Shanghai AI Laboratory, and the University of Texas at Austin. The primary advantage of Enhance-A-Video is its ability to enhance the performance of existing video models at zero cost, without the need for retraining. It introduces a temperature parameter to control inter-frame correlations, improving the temporal attention output and thereby enhancing video quality.