Awesome

Multimodal Needle in a Haystack (MMNeedle)

This repo contains the code and data for our benchmark paper:

Multimodal Needle in a Haystack: Benchmarking Long-Context Capability of Multimodal LLMs<br> H. Wang, H. Shi, S. Tan, W. Qin, W. Wang, T. Zhang, A. Nambi, T. Ganu, H. Wang<br> [Paper] [MMNeedle Dataset]

To use this benchmark, please download the MMNeedle dataset at this link. Alternatively, you could also construct your version of MMNeedle by following the instructions below.

News

[2024-06-27] New project page set up for MMNeedle.

[2024-06-24] We released the leaderboard for Multimodal Long Context Understanding on paper-with-code!

[2024-06-17] We released the paper, code, and data for Multimodal Needle in a Haystack (MMNeedle) benchmark!

Overview

MMNeedle Evaluation Overview. Correct answers are marked with checkmark ($\checkmark$), while the incorrect answers are marked with <span style="color: red;">cross ($\times$)</span>. Our evaluation setup involves the following key components: (a) Needle Sub-Image: The needle sub-image to be retrieved based on the given caption. (b) Haystack Image Inputs: The long-context visual inputs consist of M images, each stitched from <span style="color: red;">N $\times$ N</span> sub-images. (c) Text Inputs (Instructions and Caption): Detailed instructions to MLLMs, followed by a <span style="color: green;">caption</span> describing the needle, i.e., <span style="color: green;">sub-image 20</span>. (d) LLM Outputs: The answers from different MLLMs, indicating their ability to accurately locate the needle in the haystack based on the given caption. The expected output is composed of the model's identification of the index, row, and column of the matching sub-image. The results showcase the comparative performance of various models: GPT-4o correctly predicts the exact location of the needle; Gemini Pro 1.5 only correctly predicts the image index of the needle; other API models predict incorrect locations; open-source models often output with wrong formats.

MMNeedle Evaluation Performance Comparison (Claude-3 refers to Claude 3 Opus, and Gemini-1.0/1.5 refers to Gemini Pro 1.0/1.5). The x-axis shows the results of different models, and the y-axis shows the results on various input image number M and stitching size N. For each row, i.e., setting (M,N), we show the average accuracy (%) of each model. For each stitched image, the color of row r, column c indicates the accuracy of predicting the exact position for samples with the "needle" sub-image in position (r,c) of the stitched image. For the M=10 setting, we show the average accuracy of each location (r,c) over 10 images. A <span style="color: red;"><em>redder</em></span> cell indicates lower accuracy, while a <span style="color: green;"><em>greener</em></span> cell indicates higher accuracy. The best result for each row is marked with <underline>underlining</underline>.

Step 1: Setting Up the Environment

conda env create -f context.yml

Step 2: Constructing the Dataset (Optional)

Preparing the Dataset

Download MS COCO

put val2014, annotations_trainval dir to current directory.

python ./annotations_trainval/file_to_caption.py 

Sampling Images

python sample_images.py
python sample_stitched_images.py  

Sampling Needles

python sample_single_needles.py
python sample_multiple_needles.py

Step 3: Testing a Specific Model in Different Settings

export BEGIN=0
export N_SEQ=1000
export N_NEEDLES=1 
export MODEL_PROVIDER='Gemini'
bash test.sh

Step 4: Collecting the Results

export BEGIN=0
export N_SEQ=1000
python evaluate.py
python evaluate_multi.py

Reference

@misc{wang2024multimodal,
  title={Multimodal Needle in a Haystack: Benchmarking Long-Context Capability of Multimodal Large Language Models},
  author={Hengyi Wang and
          Haizhou Shi and 
          Shiwei Tan and
          Weiyi Qin and
          Wenyuan Wang and
          Tuny Zhang and
          Akshay Nambi and
          Tanuja Ganu and
          Hao Wang},
  year={2024},
  eprint={2406.11230},
  archivePrefix={arXiv},
  primaryClass={cs.LG}
}