Vchitect-2.0: Parallel Transformer for Scaling Up Video Diffusion Models
This addresses the problem of generating high-fidelity videos from text for applications in media and AI, representing a strong specific gain with incremental architectural improvements.
The paper tackles scaling up video diffusion models for text-to-video generation by introducing Vchitect-2.0, a parallel transformer architecture that improves video quality, training efficiency, and scalability, outperforming existing methods in benchmarks.
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.