Large-scale pretrained models such as LXMERT are becoming popular for learning cross-modal representations on text-image pairs for vision-language tasks. According to the lottery ticket hypothesis, NLP and computer vision models contain smaller subnetworks capable of being trained in isolation to full performance. In this paper, we combine these observations to evaluate whether such trainable subnetworks exist in LXMERT when fine-tuned on the VQA task. In addition, we perform a model size cost-benefit analysis by investigating how much pruning can be done without significant loss in accuracy. Our experiment results demonstrate that LXMERT can be effectively pruned by 40%-60% in size with 3% loss in accuracy.
@article{arxiv.2310.15325,
title = {LXMERT Model Compression for Visual Question Answering},
author = {Maryam Hashemi and Ghazaleh Mahmoudi and Sara Kodeiri and Hadi Sheikhi and Sauleh Eetemadi},
journal= {arXiv preprint arXiv:2310.15325},
year = {2023}
}
Comments
To appear in The Fourth Annual West Coast NLP (WeCNLP) Summit