google/spiqa
Question AnsweringEnglishBenchmarkcc-by-4.0
Google/spiqa is a question answering-focused benchmark dataset in English distributed in Parquet format. It is distributed under the cc-by-4.0 license and falls in the n<1K size category, and has been downloaded 371 times.
📊 This dataset is used as an LLM benchmark. See model leaderboards →
About google/spiqa
SPIQA Dataset Card
Dataset Details
Dataset Name: SPIQA (Scientific Paper Image Question Answering)
Paper: SPIQA: A Dataset for Multimodal Question Answering on Scientific Papers
Github: SPIQA eval and metrics code repo
Dataset Summar...
Details
- Task
- Question Answering
- Language
- English
- Format
- Parquet
- Rows / instances
- N/A
- Size
- n<1K
- Creator
- Year
- 2024
- License
- cc-by-4.0
- Downloads
- 371
- Likes
- 46