Tootfinder

Enabling Efficient Batch Serving for LMaaS via Generation Length Prediction
Ke Cheng, Wen Hu, Zhi Wang, Peng Du, Jianguo Li, Sheng Zhang
https://arxiv.org/abs/2406.04785

Enabling Efficient Batch Serving for LMaaS via Generation Length Prediction
Nowadays, large language models (LLMs) are published as a service and can be accessed by various applications via APIs, also known as language-model-as-a-service (LMaaS). Without knowing the generation length of requests, existing serving systems serve requests in a first-come, first-served (FCFS) manner with a fixed batch size, which leads to two problems that affect batch serving efficiency. First, the generation lengths of requests in a batch vary, and requests with short generation lengths …

Algorithmic methods of finite discrete structures. The Four Color Theorem. Theory, methods, algorithms
Sergey Kurapov, Maxim Davidovsky
https://arxiv.org/abs/2405.05270

Algorithmic methods of finite discrete structures. The Four Color Theorem. Theory, methods, algorithms
The Four color problem is closely related to other branches of mathematics and practical applications. More than 20 of its reformulations are known, which connect this problem with problems of algebra, statistical mechanics and planning. And this is also typical for mathematics: the solution to a problem studied out of pure curiosity turns out to be useful in representing real objects and processes that are completely different in nature. Despite the published machine methods for combinatorial …