LLaVAC: Fine-tuning LLaVA as a Multimodal Sentiment Classifier

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Chay-intr, T., Chen, Y., Viriyayudhakorn, K., Theeramunkong, T.
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915138643165184
author Chay-intr, T.
Chen, Y.
Viriyayudhakorn, K.
Theeramunkong, T.
author_facet Chay-intr, T.
Chen, Y.
Viriyayudhakorn, K.
Theeramunkong, T.
contents We present LLaVAC, a method for constructing a classifier for multimodal sentiment analysis. This method leverages fine-tuning of the Large Language and Vision Assistant (LLaVA) to predict sentiment labels across both image and text modalities. Our approach involves designing a structured prompt that incorporates both unimodal and multimodal labels to fine-tune LLaVA, enabling it to perform sentiment classification effectively. Experiments on the MVSA-Single dataset demonstrate that LLaVAC outperforms existing methods in multimodal sentiment analysis across three data processing procedures. The implementation of LLaVAC is publicly available at https://github.com/tchayintr/llavac.
format Preprint
id arxiv_https___arxiv_org_abs_2502_02938
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle LLaVAC: Fine-tuning LLaVA as a Multimodal Sentiment Classifier
Chay-intr, T.
Chen, Y.
Viriyayudhakorn, K.
Theeramunkong, T.
Computation and Language
We present LLaVAC, a method for constructing a classifier for multimodal sentiment analysis. This method leverages fine-tuning of the Large Language and Vision Assistant (LLaVA) to predict sentiment labels across both image and text modalities. Our approach involves designing a structured prompt that incorporates both unimodal and multimodal labels to fine-tune LLaVA, enabling it to perform sentiment classification effectively. Experiments on the MVSA-Single dataset demonstrate that LLaVAC outperforms existing methods in multimodal sentiment analysis across three data processing procedures. The implementation of LLaVAC is publicly available at https://github.com/tchayintr/llavac.
title LLaVAC: Fine-tuning LLaVA as a Multimodal Sentiment Classifier
topic Computation and Language
url https://arxiv.org/abs/2502.02938