On the Efficacy of Eviction Policy for Key-Value Constrained Generative Language Model Inference
Fuente:
arXiv
Salvato in:
| Autori principali: | , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909110611476480 |
|---|---|
| author | Ren, Siyu Zhu, Kenny Q. |
| author_facet | Ren, Siyu Zhu, Kenny Q. |
| contents | Despite the recent success associated with Large Language Models (LLMs), they are notably cost-prohibitive to deploy in resource-constrained environments due to their excessive memory and computational demands. In addition to model parameters, the key-value cache is also stored in GPU memory, growing linearly with batch size and sequence length. As a remedy, recent works have proposed various eviction policies for maintaining the overhead of key-value cache under a given budget. This paper embarks on the efficacy of existing eviction policies in terms of importance score calculation and eviction scope construction. We identify the deficiency of prior policies in these two aspects and introduce RoCo, a robust cache omission policy based on temporal attention scores and robustness measures. Extensive experimentation spanning prefilling and auto-regressive decoding stages validates the superiority of RoCo. Finally, we release EasyKV, a versatile software package dedicated to user-friendly key-value constrained generative inference. Code available at https://github.com/DRSY/EasyKV. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_06262 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | On the Efficacy of Eviction Policy for Key-Value Constrained Generative Language Model Inference Ren, Siyu Zhu, Kenny Q. Computation and Language Artificial Intelligence Despite the recent success associated with Large Language Models (LLMs), they are notably cost-prohibitive to deploy in resource-constrained environments due to their excessive memory and computational demands. In addition to model parameters, the key-value cache is also stored in GPU memory, growing linearly with batch size and sequence length. As a remedy, recent works have proposed various eviction policies for maintaining the overhead of key-value cache under a given budget. This paper embarks on the efficacy of existing eviction policies in terms of importance score calculation and eviction scope construction. We identify the deficiency of prior policies in these two aspects and introduce RoCo, a robust cache omission policy based on temporal attention scores and robustness measures. Extensive experimentation spanning prefilling and auto-regressive decoding stages validates the superiority of RoCo. Finally, we release EasyKV, a versatile software package dedicated to user-friendly key-value constrained generative inference. Code available at https://github.com/DRSY/EasyKV. |
| title | On the Efficacy of Eviction Policy for Key-Value Constrained Generative Language Model Inference |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2402.06262 |