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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Multi-view 3D detection based on BEV (bird-eye-view) has recently achievedsignificant improvements. However, the huge memory consumption ofstate-of-the-art models makes it hard to deploy them on vehicles, and thenon-trivial latency will affect the real-time perception of streamingapplications. Despite the wide application of quantization to lighten models,we show in our paper that directly applying quantization in BEV tasks will 1)make the training unstable, and 2) lead to intolerable performance degradation.To solve these issues, our method QD-BEV enables a novel view-guideddistillation (VGD) objective, which can stabilize the quantization-awaretraining (QAT) while enhancing the model performance by leveraging both imagefeatures and BEV features. Our experiments show that QD-BEV achieves similar oreven better accuracy than previous methods with significant efficiency gains.On the nuScenes datasets, the 4-bit weight and 6-bit activation quantizedQD-BEV-Tiny model achieves 37.2% NDS with only 15.8 MB model size,outperforming BevFormer-Tiny by 1.8% with an 8x model compression. On the Smalland Base variants, QD-BEV models also perform superbly and achieve 47.9% NDS(28.2 MB) and 50.9% NDS (32.9 MB), respectively.

QD-BEV : Quantization-aware View-guided Distillation for Multi-view 3D Object Detection