Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Domain Prompt Learning for Efficiently Adapting CLIP to Unseen Domains

About

Domain generalization (DG) is a difficult transfer learning problem aiming to learn a generalizable model for unseen domains. Recent foundation models (FMs) are robust to many distribution shifts and, therefore, should substantially improve the performance of DG. In this work, we study generic ways to adopt CLIP, a Visual-Language Foundation Model, for DG problems in image classification. While ERM greatly improves the accuracy with bigger backbones and training datasets using standard DG benchmarks, fine-tuning FMs is not practical in many real-world situations. We propose Domain Prompt Learning (DPL) as a novel approach for domain inference in the form of conditional prompt generation. DPL achieved a significant accuracy improvement with only training a lightweight prompt generator (a three-layer MLP), whose parameter is of equivalent scale to the classification projector in the previous DG literature. Combining \dplshort~with CLIP provides surprising performance, raising the accuracy of zero-shot CLIP from 73.7% to 79.3% on several standard datasets, namely PACS, VLCS, OfficeHome, and TerraIncognita. We hope the simplicity and success of our approach lead to broader adoption and analysis of foundation models in the domain generalization field. Our code is available at https://github.com/shogi880/DPLCLIP.

Xin Zhang, Shixiang Shane Gu, Yutaka Matsuo, Yusuke Iwasawa• 2021

Related benchmarks

TaskDatasetResultRank
Domain GeneralizationVLCS
Accuracy80.9
238
Domain GeneralizationPACS (test)
Average Accuracy97.3
225
Domain GeneralizationOfficeHome
Accuracy83
182
Domain GeneralizationDomainBed
Average Accuracy75
127
Domain GeneralizationDomainNet
Accuracy59.5
113
Domain GeneralizationDomainBed (test)
VLCS Accuracy84.3
110
Domain GeneralizationOffice-Home (test)
Average Accuracy84.2
106
Image ClassificationPACS
Accuracy97.07
100
Domain GeneralizationTerraIncognita
Accuracy46.6
81
Image ClassificationVLCS
Accuracy83.99
76
Showing 10 of 18 rows

Other info

Code

Follow for update