This is your work, valued
Ph.D. student at UT Austin. Research intern at Adobe Research. Prev. @thu-ml
Self-Forcing. Official codebase for "Self Forcing: Bridging Training and Inference in Autoregressive Video Diffusion" (NeurIPS 2025 Spotlight)
3.4kSageAttention. Quantized Attention achieves speedup of 2-3x and 3-5x compared to FlashAttention and xformers, without lossing end-to-end metrics across language, image, and video models.
1