MagicPIG: LSH Sampling for Efficient LLM Generation
- URL: http://arxiv.org/abs/2410.16179v4
- Date: Wed, 18 Dec 2024 17:36:36 GMT
- Title: MagicPIG: LSH Sampling for Efficient LLM Generation
- Authors: Zhuoming Chen, Ranajoy Sadhukhan, Zihao Ye, Yang Zhou, Jianyu Zhang, Niklas Nolte, Yuandong Tian, Matthijs Douze, Leon Bottou, Zhihao Jia, Beidi Chen,
- Abstract summary: We show that TopK attention itself suffers from quality degradation in certain downstream tasks because attention is not always as sparse as expected.
We propose MagicPIG, a heterogeneous system based on Locality Sensitive Hashing (LSH)
MagicPIG significantly reduces the workload of attention while preserving high accuracy for diverse tasks.
- Score: 41.75038064509643
- License:
- Abstract: Large language models (LLMs) with long context windows have gained significant attention. However, the KV cache, stored to avoid re-computation, becomes a bottleneck. Various dynamic sparse or TopK-based attention approximation methods have been proposed to leverage the common insight that attention is sparse. In this paper, we first show that TopK attention itself suffers from quality degradation in certain downstream tasks because attention is not always as sparse as expected. Rather than selecting the keys and values with the highest attention scores, sampling with theoretical guarantees can provide a better estimation for attention output. To make the sampling-based approximation practical in LLM generation, we propose MagicPIG, a heterogeneous system based on Locality Sensitive Hashing (LSH). MagicPIG significantly reduces the workload of attention computation while preserving high accuracy for diverse tasks. MagicPIG stores the LSH hash tables and runs the attention computation on the CPU, which allows it to serve longer contexts and larger batch sizes with high approximation accuracy. MagicPIG can improve decoding throughput by up to $5\times$ across various GPU hardware and achieve 54ms decoding latency on a single RTX 4090 for Llama-3.1-8B-Instruct model with a context of 96k tokens. The code is available at https://github.com/Infini-AI-Lab/MagicPIG.
Related papers
- SparseAccelerate: Efficient Long-Context Inference for Mid-Range GPUs [0.0]
We introduce SparseAccelerate, a dynamic sparse attention method that adapts its sparsity patterns based on input characteristics.
Experimental results show that SparseAccelerate achieves up to a 1.04x reduction in Time-To-First-Token (TTTF) latency at 32K tokens.
arXiv Detail & Related papers (2024-12-09T04:27:03Z) - Squeezed Attention: Accelerating Long Context Length LLM Inference [64.11145320159126]
We propose Squeezed Attention as a mechanism to accelerate LLM applications where a large portion of the input prompt is fixed.
We use K-means clustering offline to group the keys for the fixed context based on semantic similarity and represent each cluster with a single centroid value.
We then compute exact attention using only these important keys from the fixed context, thereby reducing bandwidth and computational costs.
arXiv Detail & Related papers (2024-11-14T18:54:19Z) - Anchor Attention, Small Cache: Code Generation with Large Language Models [15.94784908771546]
Current practices in NLP often use sparse attention which may, unfortunately, lead to substantial inaccuracies, or hallucinations, in code generation tasks.
We propose a novel approach, AnchorCoder, which features token-wise anchor attention designed to extract and compress contextual information.
It can consistently achieve a significant (at least 70%) reduction in KV cache requirements, while preserving the majority of model's performance.
arXiv Detail & Related papers (2024-11-11T02:47:05Z) - Discovering the Gems in Early Layers: Accelerating Long-Context LLMs
with 1000x Input Token Reduction [47.38471103190534]
Large Language Models (LLMs) have demonstrated remarkable capabilities in handling long context inputs, but this comes at the cost of increased computational resources and latency.
Our research introduces a novel approach for the long context bottleneck to accelerate LLM inference and reduce GPU memory consumption.
We propose an algorithm that uses early layers of an LLM as filters to select and compress input tokens, significantly reducing the context length for subsequent processing.
arXiv Detail & Related papers (2024-09-25T23:14:47Z) - RetrievalAttention: Accelerating Long-Context LLM Inference via Vector Retrieval [24.472784635757016]
RetrievalAttention is a training-free approach to both accelerate attention computation and reduce GPU memory consumption.
We show that RetrievalAttention achieves near full attention accuracy while only requiring access to 1--3% of the data.
arXiv Detail & Related papers (2024-09-16T17:59:52Z) - MInference 1.0: Accelerating Pre-filling for Long-Context LLMs via Dynamic Sparse Attention [36.49445805074941]
MInference (Milliontokens Inference) is a sparse calculation method designed to accelerate pre-filling of long-sequence processing.
We demonstrate that MInference effectively reduces inference latency by up to 10x for pre-filling on an A100, while maintaining accuracy.
arXiv Detail & Related papers (2024-07-02T17:59:56Z) - A Training-free Sub-quadratic Cost Transformer Model Serving Framework With Hierarchically Pruned Attention [43.211427581302715]
We propose Hierarchically Pruned Attention (HiP) to increase context length in large language models.
HiP reduces the time complexity of the attention mechanism to $O(T log T)$ and the space complexity to $O(T)$, where $T$ is the sequence length.
We show that HiP significantly reduces both prefill and decoding latencies, as well as memory usage, while maintaining high-quality generation with minimal degradation.
arXiv Detail & Related papers (2024-06-14T08:32:45Z) - Enabling High-Sparsity Foundational Llama Models with Efficient Pretraining and Deployment [56.44025052765861]
Large language models (LLMs) have revolutionized Natural Language Processing (NLP), but their size creates computational bottlenecks.
We introduce a novel approach to create accurate, sparse foundational versions of performant LLMs.
We show a total speedup on CPUs for sparse-quantized LLaMA models of up to 8.6x.
arXiv Detail & Related papers (2024-05-06T16:03:32Z) - Sparsity May Cry: Let Us Fail (Current) Sparse Neural Networks Together! [100.19080749267316]
"Sparsity May Cry" Benchmark (SMC-Bench) is a collection of carefully-curated 4 diverse tasks with 10 datasets.
SMC-Bench is designed to favor and encourage the development of more scalable and generalizable sparse algorithms.
arXiv Detail & Related papers (2023-03-03T18:47:21Z) - SMYRF: Efficient Attention using Asymmetric Clustering [103.47647577048782]
We propose a novel type of balanced clustering algorithm to approximate attention.
SMYRF can be used as a drop-in replacement for dense attention layers without any retraining.
We show that SMYRF can be used interchangeably with dense attention before and after training.
arXiv Detail & Related papers (2020-10-11T18:49:17Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.