Search papers, labs, and topics across Lattice.
2
0
3
3
A novel $\mathrm{SemiCat}$ architecture achieves discrete diffusion-level performance on TinyGSM and outperforms one-hot flows at scale by addressing self-conditioning and embedding geometry in continuous diffusion language models.
Tri-modal masked diffusion models can now be trained from scratch, achieving strong results in text generation, text-to-image, and text-to-speech, thanks to a systematic exploration of the design space and a novel SDE-based batch size reparameterization.
