Related papers: Time and Memory Trade-off of KV-Cache Compression …

Following the success of dot-product attention in Transformers, numerous approximations have been recently proposed to address its quadratic complexity with respect to the input length. While these variants are memory and compute efficient,…

The increasing size of the Key-Value (KV) cache during the Large Language Models long-context inference is the main obstacle for its balance between the deployment cost and task accuracy. To reduce the KV cache size in such scenarios, most…

Large Language models (LLMs) have become a research hotspot. To accelerate the inference of LLMs, storing computed caches in memory has become the standard technique. However, as the inference length increases, growing KV caches might lead…

Context lengths of Large Language Models (LLMs) have exploded in recent years, with 128k-token context becoming a standard and million-token context becoming a reality. Efficiently supporting long-context inference remains challenging as…

As long-context language modeling becomes increasingly important, the cost of maintaining and attending to large Key/Value (KV) caches grows rapidly, becoming a major bottleneck in both training and inference. While prior works such as…

This paper derives upper and lower bounds on the secrecy capacity-memory tradeoff of a wiretap erasure broadcast channel (BC) with Kw weak receivers and Ks strong receivers, where weak receivers, respectively strong receivers, have same…

At K=16 tokens (0.4% of a 4K context), every existing KV-cache compression method achieves 0% on credential retrieval. The failure mode is dormant tokens: credentials, API keys, and configuration values that receive near-zero attention but…

Key-value (KV) caching is widely used to accelerate transformer inference, but its memory cost grows linearly with input length, limiting long-context deployment. Existing token eviction methods reduce memory by discarding less important…

Although the Transformer has become the cornerstone of modern AI, its autoregressive inference suffers from a linearly growing KV Cache and a computational complexity of O(N^2 d), severely hindering its ability to process ultra-long…

Key-Value (KV) cache quantization has become a widely adopted optimization technique for efficient large language models (LLMs) inference by reducing KV cache memory usage and mitigating memory-bound constraints. Recent studies have…

Large Language Models (LLMs), despite their remarkable performance across a wide range of tasks, necessitate substantial GPU memory and consume significant computational resources. Beyond the memory taken up by model weights, the memory…

As the context length of current large language models (LLMs) rapidly increases, the memory demand for the Key-Value (KV) cache is becoming a bottleneck for LLM deployment and batch processing. Traditional KV cache compression methods…

The rapid expansion of context window sizes in Large Language Models~(LLMs) has enabled them to tackle increasingly complex tasks involving lengthy documents. However, this progress comes at the cost of a substantial increase in memory…

In this study, we introduce adaptive KV cache compression, a plug-and-play method that reduces the memory footprint of generative inference for Large Language Models (LLMs). Different from the conventional KV cache that retains key and…

KV caches, typically used only to speed up autoregressive decoding, encode contextual information that can be reused for downstream tasks at no extra cost. We propose treating the KV cache as a lightweight representation, eliminating the…

Transformers process images and videos by flattening space and time into long token sequences. While attention and KV caching preserve past features, their memory grows with sequence length and they lack an explicit, persistent spatial…

The Key-Value (KV) cache is crucial for efficient Large Language Models (LLMs) inference, but excessively long contexts drastically increase KV cache memory footprint. Existing KV cache compression methods typically rely on input-side…

In this paper, we exhibit the tradeoffs between the (training) sample, computation and storage complexity for the problem of supervised classification using signal subspace estimation. Our main tool is the use of tensor subspaces, i.e.…

When transformer-based language models are deployed for text generation, most of the inference time is spent in the decoding stage, where output tokens are generated sequentially. Reducing the hardware cost of each decoding step is…

Scaling language models to long contexts is often bottlenecked by the size of the key-value (KV) cache. In deployed settings, long contexts are typically managed through compaction in token space via summarization. However, summarization…