Search papers, labs, and topics across Lattice.
4
6
9
2
LLMs can now optimize GPU kernels more effectively by learning from a structured memory of optimization strategies at different levels of abstraction.
Stop retraining your diffusion models for every device: OFA-Diffusion lets you extract the right-sized model in a single training run.
Reconstruction-based AI-generated image detectors, despite their promise, can be completely fooled by adding imperceptible noise, revealing a critical security flaw.
DPO's classification loss, often seen as distinct from RL, is actually deeply connected to RL algorithms like PPO, according to a new unified framework.
