TroL: Traversal of Layers for Large Language and Vision Models

Byung-Kwan Lee, Sangyun Chung, Chae Won Kim, Beomchan Park, Yong Man Ro


Abstract
Large language and vision models (LLVMs) have been driven by the generalization power of large language models (LLMs) and the advent of visual instruction tuning. Along with scaling them up directly, these models enable LLVMs to showcase powerful vision language (VL) performances by covering diverse tasks via natural language instructions. However, existing open-source LLVMs that perform comparably to closed-source LLVMs such as GPT-4V are often considered too large (e.g., 26B, 34B, and 110B parameters), having a larger number of layers. These large models demand costly, high-end resources for both training and inference. To address this issue, we present a new efficient LLVM family with 1.8B, 3.8B, and 7B LLM model sizes, Traversal of Layers (TroL), which enables the reuse of layers in a token-wise manner. This layer traversing technique simulates the effect of looking back and retracing the answering stream while increasing the number of forward propagation layers without physically adding more layers. We demonstrate that TroL employs a simple layer traversing approach yet efficiently outperforms the open-source LLVMs with larger model sizes and rivals the performances of the closed-source LLVMs with substantial sizes.
Anthology ID:
2024.emnlp-main.633
Volume:
Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing
Month:
November
Year:
2024
Address:
Miami, Florida, USA
Editors:
Yaser Al-Onaizan, Mohit Bansal, Yun-Nung Chen
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
11314–11342
Language:
URL:
https://aclanthology.org/2024.emnlp-main.633
DOI:
Bibkey:
Cite (ACL):
Byung-Kwan Lee, Sangyun Chung, Chae Won Kim, Beomchan Park, and Yong Man Ro. 2024. TroL: Traversal of Layers for Large Language and Vision Models. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 11314–11342, Miami, Florida, USA. Association for Computational Linguistics.
Cite (Informal):
TroL: Traversal of Layers for Large Language and Vision Models (Lee et al., EMNLP 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.emnlp-main.633.pdf