Yotta Labs Academic
Research Support Program

A low-rank–enhanced 4-bit quantization method that preserves image quality while reducing memory and boosting inference speed for large diffusion models.

A unified inference and post-training framework that boosts large video diffusion model performance, enabling faster generation an defficient training pipelines.

A research lab exploring use-inspired Al-reinforcement learning generative models and optimization-for real-world domains including health, environment and physical sciences.