相关论文: QMoE: Practical Sub-1-Bit Compression of Trillion-…

With the widespread adoption of Large Language Models (LLMs), many deep learning practitioners are looking for strategies of running these models more efficiently. One such strategy is to use sparse Mixture-of-Experts (MoE) - a type of…

Mixture-of-Experts (MoE) models have recently demonstrated exceptional performance across a diverse range of applications. The principle of sparse activation in MoE models facilitates an offloading strategy, wherein active experts are…

Sparse Mixture-of-Experts (MoE) models can outperform dense large language models at similar computation by activating only a small set of experts per token. However, stacking many expert modules introduces substantial parameter memory,…

Mixture-of-Experts (MoE) has become a practical architecture for scaling LLM capacity while keeping per-token compute modest, but deploying MoE models on a single, memory-limited GPU remains difficult because expert weights dominate the HBM…

The scaling of large language models (LLMs) has revolutionized their capabilities in various tasks, yet this growth must be matched with efficient computational strategies. The Mixture-of-Experts (MoE) architecture stands out for its…

Mixture-of-Experts (MoE) models facilitate edge deployment by decoupling model capacity from active computation, yet their large memory footprint drives the need for GPU systems with near-data processing (NDP) capabilities that offload…

Mixture of Experts (MoE) has become a mainstream architecture for building Large Language Models (LLMs) by reducing per-token computation while enabling model scaling. It can be viewed as partitioning a large Feed-Forward Network (FFN) at…

The emergence of large-scale Mixture of Experts (MoE) models represents a significant advancement in artificial intelligence, offering enhanced model capacity and computational efficiency through conditional computation. However, deploying…

As giant dense models advance quality but require large amounts of GPU budgets for training, the sparsely gated Mixture-of-Experts (MoE), a kind of conditional computation architecture, is proposed to scale models while keeping their…

This paper presents MoE-Gen, a high-throughput MoE inference system optimized for single-GPU execution. Existing inference systems rely on model-based or continuous batching strategies, originally designed for interactive inference, which…

Mixture-of-Experts (MoE) activates only a subset of experts during inference, allowing the model to maintain low inference FLOPs and latency even as the parameter count scales up. However, since MoE dynamically selects the experts, all the…

Mixture-of-Experts (MoE) architectures in large language models (LLMs) achieve exceptional performance, but face prohibitive storage and memory requirements. To address these challenges, we present $D^2$-MoE, a new delta decompression…

Mixture-of-Experts (MoE) has garnered significant attention for its ability to scale up neural networks while utilizing the same or even fewer active parameters. However, MoE does not alleviate the massive memory requirements of networks,…

Mixture-of-Experts (MoE) layers activate a subset of model weights, dubbed experts, to improve model performance. MoE is particularly promising for deployment on process-in-memory (PIM) architectures, because PIM can naturally fit experts…

The deployment of mixture-of-experts (MoE) large language models (LLMs) presents significant challenges due to their high memory demands. These challenges become even more pronounced in multi-tenant environments, where shared resources must…

KV cache memory is the dominant bottleneck for long-context LLM inference. Existing compression methods each act on a single axis of the four-dimensional KV tensor -- token eviction (sequence), quantization (precision), low-rank projection…

Mixture-of-Experts (MoE) models improve the scalability of large language models (LLMs) by activating only a small subset of relevant experts per input. However, the sheer number of expert networks in an MoE model introduces a significant…

Large Language Models (LLMs) with Mixture-of-Expert (MoE) architectures achieve superior model performance with reduced computation costs, but at the cost of high memory capacity and bandwidth requirements. Near-Memory Processing (NMP)…

The expansion of large language models is increasingly limited by the constrained memory capacity of modern GPUs. To mitigate this, Mixture-of-Experts (MoE) architectures activate only a small portion of parameters during inference,…

The generation quality of large language models (LLMs) is often improved by utilizing inference-time sequence-level scaling methods (e.g., Chain-of-Thought). We introduce hyper-parallel scaling, a complementary framework that improves…