Search papers, labs, and topics across Lattice.
4
0
7
3
State-of-the-art vision-language models fail to leverage visual context, leading to biased outputs, but a new training framework shows they can learn to infer concepts from image sets effectively.
Normalizing Flows can now compete with diffusion models on image generation tasks, thanks to an iterative denoising scheme that boosts performance without sacrificing likelihood-based training.
Forget generating entire videos – this method distills motion into a highly compressed latent space, letting you steer scene dynamics with text prompts at unprecedented speeds.
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.
