English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Large-scale Pre-Training Vision-Language Model such as CLIP has demonstratedoutstanding performance in zero-shot classification, e.g. achieving 76.3% top-1accuracy on ImageNet without seeing any example, which leads to potentialbenefits to many tasks that have no labeled data. However, while applying CLIPto a downstream target domain, the presence of visual and text domain gaps andcross-modality misalignment can greatly impact the model performance. Toaddress such challenges, we propose ReCLIP, the first source-free domainadaptation method for vision-language models, which does not require any sourcedata or target labeled data. ReCLIP first learns a projection space to mitigatethe misaligned visual-text embeddings and learns pseudo labels, and thendeploys cross-modality self-training with the pseudo labels, to update visualand text encoders, refine labels and reduce domain gaps and misalignmentsiteratively. With extensive experiments, we demonstrate ReCLIP reduces theaverage error rate of CLIP from 30.17% to 25.06% on 22 image classificationbenchmarks. Code available at https://github.com/michiganleon/ReCLIP_WACV.

ReCLIP: Refine Contrastive Language Image Pre-Training with Source Free Domain Adaptation