AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Learning Visual Prior via Generative Pre-Training

Jinheng Xie, Kai Ye, Yudong Li, Yuexiang Li, Kevin Qinghong Lin, Yefeng Zheng, Linlin Shen, Mike Zheng Shou

NeurIPS 2023 · Vision-language and multimodal › Human and embodied modeling · approach: Vision-language pretraining · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add special words to coordinate tokens (ablation)VISORGPT · COCOKL divergence: 1.720 -> 1.647Improved
add special words to coordinate tokens (ablation)VISORGPT · Open ImagesKL divergence: 1.950 -> 1.895Improved
initialize from pretrained GPT-2 text weights (ablation)VISORGPT · COCOKL divergence: 1.720 -> 1.959Hurt
initialize from pretrained GPT-2 text weights (ablation)VISORGPT · Open ImagesKL divergence: 1.950 -> 2.240Hurt
increase sampled sequences from ~40 to ~120 (ablation)VISORGPT · COCOKL divergence: 1.958 -> 0.930Improved
increase sampled sequences from ~40 to ~120 (ablation)VISORGPT · Open ImagesKL divergence: 3.247 -> 2.144Improved
scale model from 117M to 762M parameters (ablation)VISORGPT · COCO (box)KL divergence: 0.850 -> 0.798Improved
scale model from 117M to 345M parameters (ablation)VISORGPT · COCO (box)KL divergence: 0.850 -> 0.836Improved
use both prompt templates Ta+Tb instead of Ta (full corpus) (ablation)VISORGPT · COCOKL divergence location: 1.212 -> 1.583Hurt
use both prompt templates Ta+Tb instead of Ta (full corpus) (ablation)VISORGPT · COCOKL divergence shape: 1.813 -> 1.710Improved
use both prompt templates Ta+Tb instead of Ta (small corpus) (ablation)VISORGPT-dagger · COCOKL divergence location: 1.133 -> 1.032Improved
use both prompt templates Ta+Tb instead of Ta (small corpus) (ablation)VISORGPT-dagger · COCOKL divergence relation: 0.452 -> 0.445Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy