Volodymyr Kuleshov 的扩散 LLM 指南为从掩码到模型绘制了路线。
发表于7月5日，Cornell 的教授兼 Inception 联合创始人解释了掩码去噪、可变长度生成以及质量与速度之间的权衡。
By RuntimeWire Staff · Published
Primary source: Kuleshov Group
为何重要
Kuleshov 正在将多年来在 Cornell 的研究转化为面向 Inception 商业押注的开发堆栈：并行文本生成能够使 AI 代理更快、更便宜，同时不会牺牲太多质量。
Volodymyr Kuleshov (@volokuleshov), the Cornell professor and Inception co-founder, published a technical guide on July 5th that lays out how researchers can turn masked-token prediction into a working diffusion language model. The document puts a coherent development path around a field that has spent the past two years moving from academic papers into products from Inception, Google and NVIDIA.…
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.