AMD-Hummingbird: Towards an Efficient Text-to-Video Model
Takashi Isobe, He Cui, Dong Zhou, Mengmeng Ge, Dong Li, Emad Barsoum
2025-03-25
Summary
This paper is about creating an AI model that can generate videos from text descriptions, but in a way that's faster and uses less computing power, so it can run on devices like phones.
What's the problem?
Existing AI models for making videos from text are often too big and slow to use on smaller devices.
What's the solution?
The researchers developed a new AI model called Hummingbird that is smaller and faster, and they also improved the way the model learns to create better quality videos.
Why it matters?
This work matters because it makes AI video generation more accessible and practical for everyday use on a wider range of devices.
Abstract
Text-to-Video (T2V) generation has attracted significant attention for its ability to synthesize realistic videos from textual descriptions. However, existing models struggle to balance computational efficiency and high visual quality, particularly on resource-limited devices, e.g.,iGPUs and mobile phones. Most prior work prioritizes visual fidelity while overlooking the need for smaller, more efficient models suitable for real-world deployment. To address this challenge, we propose a lightweight T2V framework, termed Hummingbird, which prunes existing models and enhances visual quality through visual feedback learning. Our approach reduces the size of the U-Net from 1.4 billion to 0.7 billion parameters, significantly improving efficiency while preserving high-quality video generation. Additionally, we introduce a novel data processing pipeline that leverages Large Language Models (LLMs) and Video Quality Assessment (VQA) models to enhance the quality of both text prompts and video data. To support user-driven training and style customization, we publicly release the full training code, including data processing and model training. Extensive experiments show that our method achieves a 31X speedup compared to state-of-the-art models such as VideoCrafter2, while also attaining the highest overall score on VBench. Moreover, our method supports the generation of videos with up to 26 frames, addressing the limitations of existing U-Net-based methods in long video generation. Notably, the entire training process requires only four GPUs, yet delivers performance competitive with existing leading methods. Hummingbird presents a practical and efficient solution for T2V generation, combining high performance, scalability, and flexibility for real-world applications.