game generation

GameFactory is an innovative general world model that focuses on learning from a small amount of Minecraft game video data and leveraging the prior knowledge of a pre-trained video diffusion model to generate new game content. The core advantage of this technology lies in its open field generation ability, which can generate diverse game scenes and interactive experiences based on text prompts and operating instructions input by users. It not only demonstrates powerful scene generation capabilities, but also achieves high-quality interactive video generation through multi-stage training strategies and pluggable motion control modules. This technology has broad application prospects in fields such as game development, virtual reality and creative content generation, but its price and commercialization positioning have not yet been clarified.

RAIN is a real-time animation unlimited video streaming technology that enables high-quality, low-latency real-time animation on consumer-grade devices. It generates video frames at faster speeds and with lower latency while maintaining video stream coherence by efficiently computing attention on frame markers at different noise levels and long intervals while denoising more frame markers than previous streaming methods. RAIN only introduces a small number of additional 1D attention blocks, which puts less burden on the system. This technology is expected to be combined with CG in fields such as game rendering, live broadcast and virtual reality, using the generalization capabilities of AI to render countless new scenes and objects, and provide a more interactive way of participation.

Oasis is the first playable, real-time, open-world AI model developed by Decart AI. It is an interactive video game generated end-to-end by Transformer on a frame-by-frame basis. Oasis can receive user keyboard and mouse input, generate gameplay in real time, and internally simulate physics, game rules and graphics. The model learns from direct observation of gameplay, allowing users to move, jump, pick up items, break blocks, and more. Oasis is seen as the first step in researching a basic model for a more complex interactive world, which may replace traditional game engines in the future. The implementation of Oasis requires improvements in model architecture and breakthroughs in model reasoning technology to achieve real-time interaction between users and models. Decart AI adopts the latest diffusion training and Transformer model methods, and combines large language models (LLMs) to train an autoregressive model that can generate videos based on the user's instant actions. In addition, Decart AI has developed a proprietary inference framework to deliver peak utilization of NVIDIA H100 Tensor Core GPUs and support Etched's upcoming Sohu chips.