Search papers, labs, and topics across Lattice.
KAIST
6
0
7
Fixed-point flows enable a leap in performance for language models, outperforming state-of-the-art methods in one- and few-step generation tasks.
Achieving high-fidelity language generation with 32x fewer function evaluations could revolutionize real-time applications of language models.
Merging tokens rather than truncating them allows diffusion models to maintain semantic alignment across variable lengths, drastically improving image generation quality.
Straightening the path to better image reconstruction, SelFix leverages trajectory straightness to significantly enhance fixed-point inversion quality.
Flow-based generative models can be improved *after* training, simply by steering intermediate states away from regions where the velocity field is highly divergent.
Forget slow, multi-step diffusion: this work achieves state-of-the-art text generation quality with a *single* denoising step using flow-based language models.
