lmms-lab/ScienceQA
收藏资源简介:
--- dataset_info: - config_name: ScienceQA-FULL features: - name: image dtype: image - name: question dtype: string - name: choices sequence: string - name: answer dtype: int8 - name: hint dtype: string - name: task dtype: string - name: grade dtype: string - name: subject dtype: string - name: topic dtype: string - name: category dtype: string - name: skill dtype: string - name: lecture dtype: string - name: solution dtype: string splits: # - name: train # num_bytes: 422199906.182 # num_examples: 12726 - name: validation num_bytes: 140142913.699 num_examples: 4241 - name: test num_bytes: 138277282.051 num_examples: 4241 download_size: 679275875 dataset_size: 700620101.932 - config_name: ScienceQA-IMG features: - name: image dtype: image - name: question dtype: string - name: choices sequence: string - name: answer dtype: int8 - name: hint dtype: string - name: task dtype: string - name: grade dtype: string - name: subject dtype: string - name: topic dtype: string - name: category dtype: string - name: skill dtype: string - name: lecture dtype: string - name: solution dtype: string splits: # - name: train # num_bytes: 413310651.0 # num_examples: 6218 - name: validation num_bytes: 137253441.0 num_examples: 2097 - name: test num_bytes: 135188432.0 num_examples: 2017 download_size: 663306124 dataset_size: 685752524.0 configs: - config_name: ScienceQA-FULL data_files: # - split: train # path: ScienceQA-FULL/train-* - split: validation path: ScienceQA-FULL/validation-* - split: test path: ScienceQA-FULL/test-* - config_name: ScienceQA-IMG data_files: # - split: train # path: ScienceQA-IMG/train-* - split: validation path: ScienceQA-IMG/validation-* - split: test path: ScienceQA-IMG/test-* --- <p align="center" width="100%"> <img src="https://i.postimg.cc/g0QRgMVv/WX20240228-113337-2x.png" width="100%" height="80%"> </p> # Large-scale Multi-modality Models Evaluation Suite > Accelerating the development of large-scale multi-modality models (LMMs) with `lmms-eval` 🏠 [Homepage](https://lmms-lab.github.io/) | 📚 [Documentation](docs/README.md) | 🤗 [Huggingface Datasets](https://huggingface.co/lmms-lab) # This Dataset This is a formatted version of [derek-thomas/ScienceQA](https://huggingface.co/datasets/derek-thomas/ScienceQA). It is used in our `lmms-eval` pipeline to allow for one-click evaluations of large multi-modality models. ``` @inproceedings{lu2022learn, title={Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answering}, author={Lu, Pan and Mishra, Swaroop and Xia, Tony and Qiu, Liang and Chang, Kai-Wei and Zhu, Song-Chun and Tafjord, Oyvind and Clark, Peter and Ashwin Kalyan}, booktitle={The 36th Conference on Neural Information Processing Systems (NeurIPS)}, year={2022} } ```
# 大规模多模态模型评测套件 > 借助`lmms-eval`加速大规模多模态模型(Large-scale Multi-modality Models, LMMs)的研发 🏠 [主页](https://lmms-lab.github.io/) | 📚 [文档](docs/README.md) | 🤗 [Huggingface数据集](https://huggingface.co/lmms-lab) # 本数据集 本数据集是[derek-thomas/ScienceQA(科学问答数据集)](https://huggingface.co/datasets/derek-thomas/ScienceQA)的格式化版本,被应用于我们的`lmms-eval`流程中,以实现大规模多模态模型的一键式评测。 @inproceedings{lu2022learn, title={学会解释:面向科学问答任务的基于思维链的多模态推理}, author={Lu Pan、Swaroop Mishra、Tony Xia、Liang Qiu、Kai-Wei Chang、Song-Chun Zhu、Øyvind Tafjord、Peter Clark、Ashwin Kalyan}, booktitle={第36届神经信息处理系统会议(NeurIPS 2022)}, year={2022} } <p align="center" width="100%"> <img src="https://i.postimg.cc/g0QRgMVv/WX20240228-113337-2x.png" width="100%" height="80%"> </p> ## 数据集信息 - 配置名称:ScienceQA-FULL 特征列表: - 字段名:image,数据类型:图像 - 字段名:question,数据类型:字符串 - 字段名:choices,数据类型:字符串序列 - 字段名:answer,数据类型:int8(8位有符号整数) - 字段名:hint,数据类型:字符串 - 字段名:task,数据类型:字符串 - 字段名:grade,数据类型:字符串 - 字段名:subject,数据类型:字符串 - 字段名:topic,数据类型:字符串 - 字段名:category,数据类型:字符串 - 字段名:skill,数据类型:字符串 - 字段名:lecture,数据类型:字符串 - 字段名:solution,数据类型:字符串 划分集: # - 划分名称:训练集 # 字节数:422199906.182 # 样本数量:12726 - 划分名称:验证集 字节数:140142913.699 样本数量:4241 - 划分名称:测试集 字节数:138277282.051 样本数量:4241 下载大小:679275875 数据集总大小:700620101.932 - 配置名称:ScienceQA-IMG 特征列表: - 字段名:image,数据类型:图像 - 字段名:question,数据类型:字符串 - 字段名:choices,数据类型:字符串序列 - 字段名:answer,数据类型:int8(8位有符号整数) - 字段名:hint,数据类型:字符串 - 字段名:task,数据类型:字符串 - 字段名:grade,数据类型:字符串 - 字段名:subject,数据类型:字符串 - 字段名:topic,数据类型:字符串 - 字段名:category,数据类型:字符串 - 字段名:skill,数据类型:字符串 - 字段名:lecture,数据类型:字符串 - 字段名:solution,数据类型:字符串 划分集: # - 划分名称:训练集 # 字节数:413310651.0 # 样本数量:6218 - 划分名称:验证集 字节数:137253441.0 样本数量:2097 - 划分名称:测试集 字节数:135188432.0 样本数量:2017 下载大小:663306124 数据集总大小:685752524.0 配置项: - 配置名称:ScienceQA-FULL 数据文件: # - 划分:训练集 # 路径:ScienceQA-FULL/train-* - 划分:验证集 路径:ScienceQA-FULL/validation-* - 划分:测试集 路径:ScienceQA-FULL/test-* - 配置名称:ScienceQA-IMG 数据文件: # - 划分:训练集 # 路径:ScienceQA-IMG/train-* - 划分:验证集 路径:ScienceQA-IMG/validation-* - 划分:测试集 路径:ScienceQA-IMG/test-*
数据集概述
ScienceQA-FULL
- 特征:
image: 图像类型question: 字符串类型choices: 字符串序列answer: 8位整数类型hint: 字符串类型task: 字符串类型grade: 字符串类型subject: 字符串类型topic: 字符串类型category: 字符串类型skill: 字符串类型lecture: 字符串类型solution: 字符串类型
- 拆分:
validation: 140142913.699字节, 4241个样本test: 138277282.051字节, 4241个样本
- 下载大小: 679275875字节
- 数据集大小: 700620101.932字节
ScienceQA-IMG
- 特征:
image: 图像类型question: 字符串类型choices: 字符串序列answer: 8位整数类型hint: 字符串类型task: 字符串类型grade: 字符串类型subject: 字符串类型topic: 字符串类型category: 字符串类型skill: 字符串类型lecture: 字符串类型solution: 字符串类型
- 拆分:
validation: 137253441字节, 2097个样本test: 135188432字节, 2017个样本
- 下载大小: 663306124字节
- 数据集大小: 685752524字节
数据文件配置
- ScienceQA-FULL:
validation: ScienceQA-FULL/validation-*test: ScienceQA-FULL/test-*
- ScienceQA-IMG:
validation: ScienceQA-IMG/validation-*test: ScienceQA-IMG/test-*




