Tag
This paper proposes a lightweight multi-agent framework using AutoGen for automated concrete barrier design, achieving over 98% accuracy and showing that smaller models can outperform massive ones in this domain.
Light-WAM is a lightweight world action model for efficient robot manipulation that uses a compact video backbone and downsampled latent space for future-video supervision, achieving high performance with low inference latency.
Tencent open-sourced Hy-MT2, a family of multilingual translation models in sizes 1.8B, 7B, and 30B-A3B (MoE), supporting 33 languages and achieving state-of-the-art results, with the lightweight 1.8B model outperforming mainstream commercial APIs from Microsoft and Doubao.
本文提出一个轻量级、乐器无关的神经网络，用于多音符转录和复音估计，联合预测帧级起音、复音和音符激活，在保持接近最先进系统精度的同时大幅降低资源占用（对应 Spotify 的 Basic Pitch 模型）。