相关论文: LaneRoPE: Positional Encoding for Collaborative Pa…

Linear sequence modeling approaches, such as linear attention, provide advantages like linear-time training and constant-memory inference over sequence lengths. However, existing sequence parallelism (SP) methods are either not optimized…

Relative positional encoding is widely used in vanilla and linear transformers to represent positional information. However, existing encoding methods of a vanilla transformer are not always directly applicable to a linear transformer,…

Looping, reusing a block of layers across depth, and depth growing, training shallow-to-deep models by duplicating middle layers, have both been linked to stronger reasoning, but their relationship remains unclear. We provide a mechanistic…

There is a growing interest in Universal Multimodal Embeddings (UME), where models are required to generate task-specific representations. While recent studies show that Multimodal Large Language Models (MLLMs) perform well on such tasks,…

Large reasoning models (LRMs) excel at complex reasoning tasks but typically generate lengthy sequential chains-of-thought, resulting in long inference times before arriving at the final answer. To address this challenge, we introduce…

Breakthroughs in the generative AI domain have fueled an explosion of large language model (LLM)-powered applications, whose workloads fundamentally consist of sequences of inferences through transformer architectures. Within this rapidly…

Efficiently training LLMs with long sequences is important yet challenged by the massive computation and memory requirements. Sequence parallelism has been proposed to tackle these problems, but existing methods suffer from scalability or…

Numerical reasoning over documents, which demands both contextual understanding and logical inference, is challenging for low-capacity local models deployed on computation-constrained devices. Although such complex reasoning queries could…

Model ensembling is a well-established technique for improving the performance of machine learning models. Conventionally, this involves averaging the output distributions of multiple models and selecting the most probable label. This idea…

Inference of Large Language Models (LLMs) across computer clusters has become a focal point of research in recent times, with many acceleration techniques taking inspiration from CPU speculative execution. These techniques reduce…

Traditional recommender systems such as matrix factorization methods have primarily focused on learning a shared dense embedding space to represent both items and user preferences. Subsequently, sequence models such as RNN, GRUs, and,…

Extending large language models (LLMs) to process longer inputs is crucial for a wide range of applications. However, the substantial computational cost of transformers and limited generalization of positional encoding restrict the size of…

The attention mechanism is a critical component of Large Language Models (LLMs) that allows tokens in a sequence to interact with each other, but is order-invariant. Incorporating position encoding (PE) makes it possible to address by…

This research aims to accelerate the inference speed of large language models (LLMs) with billions of parameters. We propose \textbf{S}mart \textbf{P}arallel \textbf{A}uto-\textbf{C}orrect d\textbf{E}coding (SPACE), an innovative approach…

Rotary Position Embeddings (RoPE) have become a standard for encoding sequence order in Large Language Models (LLMs) by applying rotations to query and key vectors in the complex plane. Standard implementations, however, utilize only the…

Sequence parallelism (SP) serves as a prevalent strategy to handle long sequences that exceed the memory limit of a single device. However, for linear sequence modeling methods like linear attention, existing SP approaches do not take…

Several methods exist today to accelerate Machine Learning(ML) or Deep-Learning(DL) model performance for training and inference. However, modern techniques that rely on various graph and operator parallelism methodologies rely on search…

Positional encoding is essential for large language models (LLMs) to represent sequence order, yet recent studies show that Rotary Position Embedding (RoPE) can induce massive activation. We investigate the source of these instabilities via…

Large Language Models (LLMs) are known to have limited extrapolation ability beyond their pre-trained context window, constraining their application in downstream tasks with lengthy inputs. Recent studies have sought to extend LLMs' context…

Recent works have shown that Large Language Models (LLMs) could empower traditional neuro-symbolic models via programming capabilities to translate language into module descriptions, thus achieving strong visual reasoning results while…