E^2-LLM: Efficient and Extreme Length Extension of Large Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866929252301012992 |
|---|---|
| author | Liu, Jiaheng Bai, Zhiqi Zhang, Yuanxing Zhang, Chenchen Zhang, Yu Zhang, Ge Wang, Jiakai Que, Haoran Chen, Yukang Su, Wenbo Ge, Tiezheng Fu, Jie Chen, Wenhu Zheng, Bo |
| author_facet | Liu, Jiaheng Bai, Zhiqi Zhang, Yuanxing Zhang, Chenchen Zhang, Yu Zhang, Ge Wang, Jiakai Que, Haoran Chen, Yukang Su, Wenbo Ge, Tiezheng Fu, Jie Chen, Wenhu Zheng, Bo |
| contents | Typically, training LLMs with long context sizes is computationally expensive, requiring extensive training hours and GPU resources. Existing long-context extension methods usually need additional training procedures to support corresponding long-context windows, where the long-context training data (e.g., 32k) is needed, and high GPU training costs are assumed. To address the aforementioned issues, we propose an Efficient and Extreme length extension method for Large Language Models, called E 2 -LLM, with only one training procedure and dramatically reduced computation cost, which also removes the need to collect long-context data. Concretely, first, the training data of our E 2 -LLM only requires a short length (e.g., 4k), which reduces the tuning cost greatly. Second, the training procedure on the short training context window is performed only once time, and we can support different evaluation context windows at inference. Third, in E 2 - LLM, based on RoPE position embeddings, we introduce two different augmentation methods on the scale and position index parameters for different samples in training. It aims to make the model more robust to the different relative differences when directly interpolating the arbitrary context length at inference. Comprehensive experimental results on multiple benchmark datasets demonstrate the effectiveness of our E 2 -LLM on challenging long-context tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_06951 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | E^2-LLM: Efficient and Extreme Length Extension of Large Language Models Liu, Jiaheng Bai, Zhiqi Zhang, Yuanxing Zhang, Chenchen Zhang, Yu Zhang, Ge Wang, Jiakai Que, Haoran Chen, Yukang Su, Wenbo Ge, Tiezheng Fu, Jie Chen, Wenhu Zheng, Bo Computation and Language Artificial Intelligence Typically, training LLMs with long context sizes is computationally expensive, requiring extensive training hours and GPU resources. Existing long-context extension methods usually need additional training procedures to support corresponding long-context windows, where the long-context training data (e.g., 32k) is needed, and high GPU training costs are assumed. To address the aforementioned issues, we propose an Efficient and Extreme length extension method for Large Language Models, called E 2 -LLM, with only one training procedure and dramatically reduced computation cost, which also removes the need to collect long-context data. Concretely, first, the training data of our E 2 -LLM only requires a short length (e.g., 4k), which reduces the tuning cost greatly. Second, the training procedure on the short training context window is performed only once time, and we can support different evaluation context windows at inference. Third, in E 2 - LLM, based on RoPE position embeddings, we introduce two different augmentation methods on the scale and position index parameters for different samples in training. It aims to make the model more robust to the different relative differences when directly interpolating the arbitrary context length at inference. Comprehensive experimental results on multiple benchmark datasets demonstrate the effectiveness of our E 2 -LLM on challenging long-context tasks. |
| title | E^2-LLM: Efficient and Extreme Length Extension of Large Language Models |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2401.06951 |