Search papers, labs, and topics across Lattice.
4
1
9
10
Natural-language critiques can transform how we evaluate and optimize song generation models, leading to more human-aligned outputs.
Today's best language models can barely make sense of your messy group chats and fragmented digital life, achieving only 19% accuracy on a new benchmark of real-world reasoning.
Retrofit your VLMs with Multi-Head Latent Attention (MLA) for faster inference and smaller memory footprint, without costly pretraining, using this parameter-efficient conversion framework.
Finally, a fully open-source, reproducible system for long-form song generation is here, complete with licensed data, code, and a Qwen-based model that rivals closed-source systems.
