Summary of Catp: Cross-attention Token Pruning For Accuracy Preserved Multimodal Model Inference, by Ruqi Liao et al.
CATP: Cross-Attention Token Pruning for Accuracy Preserved Multimodal Model Inference
by Ruqi Liao, Chuqing Zhao, Jin Li, Weiqi Feng
First submitted to arxiv on: 2 Apr 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper introduces Cross-Attention Token Pruning (CATP), a method for precision-focused token pruning in large multimodal models. By leveraging cross-attention layers, CATP determines token importance and achieves up to 12.1X higher accuracy compared to existing methods, addressing the trade-off between computational efficiency and model precision. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper creates a new way to make big models smaller without losing their good qualities. It uses special attention parts in these models to figure out which words are most important. This helps make the model work better while using less computer power. |
Keywords
» Artificial intelligence » Attention » Cross attention » Precision » Pruning » Token
