Large Language Models (LLMs) for Generative AI have achieved remarkable progress, evolving into sophisticated and versatile tools widely adopted across various domains and applications. However, the substantial memory overhead caused by their vast number of parameters, combined with the high computational demands of the attention mechanism, poses significant challenges in achieving low latency and high throughput for LLM inference services. Recent advancements, driven by groundbreaking research, have significantly accelerated progress in this field. This paper provides a comprehensive survey of these methods, covering fundamental instance-level approaches, in-depth cluster-level strategies, and emerging scenarios. At the instance level, we review model placement, request scheduling, decoding length prediction, storage management, and the disaggregation paradigm. At the cluster level, we explore GPU cluster deployment, multi-instance load balancing, and cloud service solutions. Additionally, we discuss specific tasks, modules, and auxiliary methods in emerging scenarios. Finally, we outline potential research directions to further advance the field of LLM inference serving.

Full-parameter fine-tuning (FPFT) has become the go-to choice for adapting language models (LMs) to downstream tasks due to its excellent performance. As LMs grow in size, fine-tuning the full parameters of LMs requires a prohibitively large amount of GPU memory. Existing approaches utilize zeroth-order optimizer to conserve GPU memory, which potentially compromises the performance of LMs as non-zero order optimizers tend to converge more readily on most downstream tasks. We propose a novel, memory-efficient, optimizer-independent, end-to-end hierarchical fine-tuning strategy, HiFT, which only updates a subset of parameters at each training step. HiFT significantly reduces the amount of gradients and optimizer state parameters residing in GPU memory at the same time, thereby reducing GPU memory usage. Our results demonstrate that: (1) HiFT achieves comparable performance with parameter-efficient fine-tuning and standard FPFT. (2) Results on six models show that HiFT reduces the number of trainable parameters by about 89.18% on average compared to FPFT. (3) HiFT supports FPFT of 7B models for 24G GPU memory devices under mixed precision without using any memory saving techniques. (4) HiFT supports various optimizers including AdamW, AdaGrad, SGD, etc. The source code link is https://github.com/misonsky/HiFT.

A wide body of evidence shows that human language processing difficulty is predicted by the information-theoretic measure surprisal, a word’s negative log probability in context. However, it is still unclear how to best estimate these probabilities needed for predicting human processing difficulty – while a long-standing belief held that models with lower perplexity would provide more accurate estimates of word predictability, and therefore lead to better reading time predictions, recent work has shown that for very large models, psycholinguistic predictive power decreases. One reason could be that language models might be more confident of their predictions than humans, because they have had exposure to several magnitudes more data. In this paper, we test what effect temperature-scaling of large language model (LLM) predictions has on surprisal estimates and their predictive power of reading times of English texts. Firstly, we show that calibration of large language models typically improves with model size, i.e. poorer calibration cannot account for poorer fit to reading times. Secondly, we find that temperature-scaling probabilities lead to a systematically better fit to reading times (up to 89% improvement in delta log likelihood), across several reading time corpora. Finally, we show that this improvement in fit is chiefly driven by words that are composed of multiple subword tokens.

Parallelism is an important rhetorical device. We propose a machine learning approach for automated sentence parallelism identification in student essays. We build an essay dataset with sentence level parallelism annotated. We derive features by combining generalized word alignment strategies and the alignment measures between word sequences. The experimental results show that sentence parallelism can be effectively identified with a F1 score of 82% at pair-wise level and 72% at parallelism chunk level. Based on this approach, we automatically identify sentence parallelism in more than 2000 student essays and study the correlation between the use of sentence parallelism and the types and quality of essays.