AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

On the Effect of Pre-training for Transformer in Different Modality on Offline Reinforcement Learning

Shiro Takagi

NeurIPS 2022 · Architectures and efficiency › Pruning and efficient tuning · approach: Efficient ViT adaptation · 1 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
Pre-train on language (GPT2) instead of random initDecision Transformer · Hopper-mediumnormalized return: 67.7 -> 79.5Improved
Pre-train on language (GPT2) instead of random initDecision Transformer · HalfCheetah-mediumnormalized return: 48.6 -> 48.4No clear effect
Pre-train on language (GPT2) instead of random initDecision Transformer · Walker 2D-mediumnormalized return: 71.0 -> 71.2No clear effect
Pre-train on images (iGPT) instead of random initDecision Transformer · Hopper-mediumnormalized return: 67.7 -> 2.6Hurt
Pre-train on images (iGPT) instead of random initDecision Transformer · HalfCheetah-mediumnormalized return: 48.6 -> 1.3Hurt
Pre-train on images (iGPT) instead of random initDecision Transformer · Walker 2D-mediumnormalized return: 71.0 -> 4.3Hurt
Remove context (K=1) (ablation)randomly initialized model · Hopper-mediumnormalized return: 67.7 -> 0.24Hurt
Remove context (K=1) for GPT2 (ablation)GPT2 Decision Transformer · Hopper-mediumnormalized return: 79.5 -> 83.73Improved
Remove context (K=1) for GPT2 (ablation)GPT2 Decision Transformer · Walker 2D-mediumnormalized return: 71.2 -> 69.0No clear effect
Remove context (K=1) for GPT2 (ablation)GPT2 Decision Transformer · HalfCheetah-mediumnormalized return: 48.4 -> 47.7No clear effect
Remove context (K=1) for random init (ablation)randomly initialized model · Walker 2D-mediumnormalized return: 71.0 -> 69.1No clear effect
Replace block 1 of random init with pre-trained GPT2 block, no context (ablation)randomly initialized model · Hopper-mediumaction error and mean returnNo clear effect

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