Beyond Accuracy Optimization: Computer Vision Losses for Large Language Model Fine-Tuning
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_ | 1866917867355635712 |
|---|---|
| author | Cambrin, Daniele Rege Gallipoli, Giuseppe Benedetto, Irene Cagliero, Luca Garza, Paolo |
| author_facet | Cambrin, Daniele Rege Gallipoli, Giuseppe Benedetto, Irene Cagliero, Luca Garza, Paolo |
| contents | Large Language Models (LLMs) have demonstrated impressive performance across various tasks. However, current training approaches combine standard cross-entropy loss with extensive data, human feedback, or ad hoc methods to enhance performance. These solutions are often not scalable or feasible due to their associated costs, complexity, or resource requirements. This study investigates the use of established semantic segmentation loss functions in natural language generation to create a versatile, practical, and scalable solution for fine-tuning different architectures. We evaluate their effectiveness in solving Math Word Problems and question answering across different models of varying sizes. For the analyzed tasks, we found that the traditional Cross-Entropy loss represents a sub-optimal choice, while models trained to minimize alternative (task-dependent) losses, such as Focal or Lovász, achieve a mean improvement of +42% on exact match without requiring additional data or human feedback. These findings suggest a promising pathway for more efficient and accessible training processes. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_13641 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Beyond Accuracy Optimization: Computer Vision Losses for Large Language Model Fine-Tuning Cambrin, Daniele Rege Gallipoli, Giuseppe Benedetto, Irene Cagliero, Luca Garza, Paolo Computation and Language Computer Vision and Pattern Recognition Large Language Models (LLMs) have demonstrated impressive performance across various tasks. However, current training approaches combine standard cross-entropy loss with extensive data, human feedback, or ad hoc methods to enhance performance. These solutions are often not scalable or feasible due to their associated costs, complexity, or resource requirements. This study investigates the use of established semantic segmentation loss functions in natural language generation to create a versatile, practical, and scalable solution for fine-tuning different architectures. We evaluate their effectiveness in solving Math Word Problems and question answering across different models of varying sizes. For the analyzed tasks, we found that the traditional Cross-Entropy loss represents a sub-optimal choice, while models trained to minimize alternative (task-dependent) losses, such as Focal or Lovász, achieve a mean improvement of +42% on exact match without requiring additional data or human feedback. These findings suggest a promising pathway for more efficient and accessible training processes. |
| title | Beyond Accuracy Optimization: Computer Vision Losses for Large Language Model Fine-Tuning |
| topic | Computation and Language Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2409.13641 |