Multi-query attention (MQA), which only uses a single key-value head, drastically speeds up decoder inference. However, MQA can lead to quality degradation, and moreover it may not be desirable to train a separate model just for faster inference. We (1) propose a recipe for uptraining existing multi-head language model checkpoints into models with MQA using 5% of original pre-training compute, and (2) introduce grouped-query attention (GQA), a generalization of multi-query attention which use...| papers.cool
Multi-head attention layers, as used in the Transformer neural sequence model, are a powerful alternative to RNNs for moving information across and between sequences. While training these layers is generally fast and simple, due to parallelizability across the length of the sequence, incremental inference (where such paralleization is impossible) is often slow, due to the memory-bandwidth cost of repeatedly loading the large "keys" and "values" tensors. We propose a variant called multi-query...| papers.cool
2017年中，有两篇类似同时也是笔者非常欣赏的论文，分别是FaceBook的《Convolutional Sequence to Sequence Learning》和Google的《Atten...| kexue.fm
前几天，幻方发布的DeepSeek-V2引起了大家的热烈讨论。首先，最让人哗然的是1块钱100万token的价格，普遍比现有的各种竞品API便宜了两个数量级，以至于有人调侃“这个价格哪怕它输出乱...| kexue.fm
We present DeepSeek-V2, a strong Mixture-of-Experts (MoE) language model characterized by economical training and efficient inference. It comprises 236B total parameters, of which 21B are activated for each token, and supports a context length of 128K tokens. DeepSeek-V2 adopts innovative architectures including Multi-head Latent Attention (MLA) and DeepSeekMoE. MLA guarantees efficient inference through significantly compressing the Key-Value (KV) cache into a latent vector, while DeepSeekMo...| papers.cool
事实上，除了写博客内容，在这几年里，笔者是花了相当一部分时间来做科学空间的“表面功夫”，为此还专门学了一点php、css和js。虽然不敢说精益求精，但总体来说网站的浏览体验应该比前几年要好得多。...| kexue.fm
In the process of working on calm, a minimal from-scratch fast CUDA implementation of transformer-based language model inference, a critical consideration was establishing the speed of light for the inference process, and measuring the progress relative to that speed of light. In this post we’ll cover this theoretical limit and its implications.| zeux.io