Summary of Crome: Cross-modal Adapters For Efficient Multimodal Llm, by Sayna Ebrahimi et al.
CROME: Cross-Modal Adapters for Efficient Multimodal LLM
by Sayna Ebrahimi, Sercan O. Arik, Tejas Nama, Tomas Pfister
First submitted to arxiv on: 13 Aug 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Computation and Language (cs.CL); Machine Learning (cs.LG)
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 The proposed framework, CROME, is an efficient vision-language instruction tuning method that leverages a novel gated cross-modal adapter to combine visual and textual representations prior to input into a frozen Large Language Model (LLM). This approach enables cost-effective training and adaptation of Multimodal Large Language Models (MLLMs) for tasks such as visual question answering and instruction-following. The framework demonstrates superior zero-shot performance on standard benchmarks, while also achieving exceptional parameter efficiency during fine-tuning, competing with task-specific specialist state-of-the-art methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary CROME is a new way to make computers understand images and words together. It’s like a special bridge that helps machines learn from pictures and text without needing to retrain the whole computer program. This makes it faster and cheaper to use these powerful models for tasks like recognizing objects in photos or following instructions. The results are impressive, with CROME doing as well as other top-performing methods but using fewer resources. |
Keywords
» Artificial intelligence » Fine tuning » Instruction tuning » Large language model » Question answering » Zero shot
