Dataset · yingss
MixLoRA multimodal eval dataset
yingss/mixlora-eval-data
Multimodal evaluation dataset and MME benchmark from the MixLoRA paper (ACL 2024).
What's inside
Contains a held-out multimodal evaluation set of 9 instruction tasks (Vision-Flan) and the MME benchmark for independent evaluation. Dataset size ranges from 1K to 10K English examples.
What you can do with it
- Evaluating multimodal language models and LoRA adapters
- Benchmarking Vision-Language models on visual question-answering tasks
- Tasks
- Visual question answering, Image classification
- Languages
- English
- Size
- 1K–10K rows
- Downloads this month
- 21.6K
- Likes
- 0
How to load it
from datasets import load_dataset
ds = load_dataset("yingss/mixlora-eval-data")