Louis Serrano

We propose a novel architecture and training objective specifically designed to upsample features from foundation vision encoders at any resolution.

ACCO is a new and principled optimization techniques with provable guarantees for Sharded Distributed LLM Training

We introduce a generative masked autoregressive neural operator that models parametric PDE dynamics in latent space, enabling efficient trajectory generation, in-context learning, and few-shot generalization from irregular data.