Vchitect-2.0: Parallel Transformer for Scaling Up Video Diffusion Models

2025-01-14

AI Summary

Existing text-to-video (T2V) diffusion models suffer from three key challenges: discontinuous long-video generation, severe memory bottlenecks during training, and scarcity of high-quality paired data. To address these, we propose a parallel Transformer architecture for T2V generation, introducing the first multimodal diffusion module that jointly optimizes text-video alignment and temporal coherence. We further design a hybrid-parallel, memory-efficient training framework to overcome GPU memory constraints in long-sequence modeling. Additionally, we construct Vchitect T2V DataVerse—a million-scale, high-fidelity T2V dataset—augmented with an automated aesthetic evaluation and annotation pipeline. Experiments demonstrate state-of-the-art performance across video fidelity, temporal consistency, training throughput, and model scalability. Our method achieves, for the first time, end-to-end generation of high-fidelity, minute-long videos.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Generation

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Web Mining and Content Analysis: Large pretrained models with web data
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

We present Vchitect-2.0, a parallel transformer architecture designed to scale up video diffusion models for large-scale text-to-video generation. The overall Vchitect-2.0 system has several key designs. (1) By introducing a novel Multimodal Diffusion Block, our approach achieves consistent alignment between text descriptions and generated video frames, while maintaining temporal coherence across sequences. (2) To overcome memory and computational bottlenecks, we propose a Memory-efficient Training framework that incorporates hybrid parallelism and other memory reduction techniques, enabling efficient training of long video sequences on distributed systems. (3) Additionally, our enhanced data processing pipeline ensures the creation of Vchitect T2V DataVerse, a high-quality million-scale training dataset through rigorous annotation and aesthetic evaluation. Extensive benchmarking demonstrates that Vchitect-2.0 outperforms existing methods in video quality, training efficiency, and scalability, serving as a suitable base for high-fidelity video generation.

Problem

• Video Generation
• Performance Bottleneck
• Dataset Scalability

Innovation

• Text-Video Alignment
• Efficient Training Methodology
• High-Quality Dataset