Part 2/8:
Among the features that's created a buzz is the Text-to-Video (T2V) capability, particularly demonstrated by the smallest model, T2V D 1.3B. Impressively, this model can operate on consumer-grade GPUs with a moderate 8.19 GB VRAM. This means that users with a graphics card like the Nvidia 4090 can generate a 5-second video clip at 480p within approximately four minutes. While the rendering speed may not be lightning fast just yet, it indicates a promising step towards more accessible and powerful AI video generation on personal systems.