Search papers, labs, and topics across Lattice.
2
0
6
By explicitly modeling speech, SAVE leapfrogs existing audio-visual methods for video-text retrieval, achieving substantial gains over the state-of-the-art.
Uniform-state diffusion models, often overlooked in favor of masked diffusion, surprisingly outperform autoregressive and masked diffusion models on GSM8K when scaled to 1.7B parameters, despite worse perplexity.
