#LLM优化

3 stories · 0 in 7 days · 1 in 30 days

DeepSeek-3.2 1 GEPA 1 GLM-5.3 1 QUILT 1 提示词工程 1 稀疏注意力 1 论文 1

Stories, products and papers tagged "LLM优化", newest first.

Nov 2

Papers Builders10:08

Chinese content研究：LLM 优化器的主要收益来自父代继承而非反馈广度

跑了 1750 次实验的研究：优化器变强靠的是保留能跑通的父代，不是看更多错误样本，还能省大笔算力。

#GEPA #提示词工程 #LLM优化 #论文

arXiv cs.AI@Shanglin Yang, Zhan ShiSource

Oct 9

Papers BuildersSelected10:21

QUILT sparse attention mechanism

QUILT addresses a fundamental inefficiency in sparse attention kernels by eliminating redundant memory operations through cross-query reuse. This approach significantly improves performance for long-context models without compromising accuracy, making it particularly valuable for developers working with large language models like GLM-5.3 and DeepSeek-3.2 that process extensive sequences.

#QUILT #GLM-5.3 #DeepSeek-3.2 #sparse attention

arXiv: DeepSeek@Zhenduo Zhao et al. (10)Source

Aug 7

Products BuildersOfficial00:33

Chinese contentAmazon SageMaker Python SDK v3新增LLM推理推荐

AWS把LLM优化搬进了SageMaker的notebook，跑完基准测试直接给你部署建议，不用再手动调参了。

#Amazon SageMaker #Python SDK #AWS #LLM优化

Primary sourceAWS Machine Learning Blog@Dan FergusonSource