Magic 1-For-1: Generating One Minute Video Clips within One Minute

TL;DR

Magic 1-For-1 introduces an efficient method for generating one-minute video clips in under a minute by decomposing the task into simpler steps and applying various optimization techniques to reduce computational costs.

Contribution

The paper proposes a novel two-step diffusion-based approach for fast text-to-video generation, optimizing memory and inference speed, and demonstrating high-quality, long-duration video synthesis.

Findings

Generated 5-second videos in 3 seconds.

Produced one-minute videos within one minute.

Achieved improved visual quality and motion dynamics.

Abstract

In this technical report, we present Magic 1-For-1 (Magic141), an efficient video generation model with optimized memory consumption and inference latency. The key idea is simple: factorize the text-to-video generation task into two separate easier tasks for diffusion step distillation, namely text-to-image generation and image-to-video generation. We verify that with the same optimization algorithm, the image-to-video task is indeed easier to converge over the text-to-video task. We also explore a bag of optimization tricks to reduce the computational cost of training the image-to-video (I2V) models from three aspects: 1) model convergence speedup by using a multi-modal prior condition injection; 2) inference latency speed up by applying an adversarial step distillation, and 3) inference memory cost optimization with parameter sparsification. With those techniques, we are able to generate 5-second video clips within 3 seconds. By applying a test time sliding window, we are able to generate a minute-long video within one minute with significantly improved visual quality and motion dynamics, spending less than 1 second for generating 1 second video clips on average. We conduct a series of preliminary explorations to find out the optimal tradeoff between computational cost and video quality during diffusion step distillation and hope this could be a good foundation model for open-source explorations. The code and the model weights are available at https://github.com/DA-Group-PKU/Magic-1-For-1.