VisRAG-Ret-Test-SlideVQA
收藏魔搭社区2025-12-05 更新2025-05-17 收录
下载链接：
https://modelscope.cn/datasets/OpenBMB/VisRAG-Ret-Test-SlideVQA
下载链接
链接失效反馈官方服务：
资源简介：
## Dataset Description
This is a VQA dataset based on Slide Decks from SlideVQA dataset from [SlideVQA](https://arxiv.org/abs/2301.04883).
### Load the dataset
```python
from datasets import load_dataset
import csv
def load_beir_qrels(qrels_file):
qrels = {}
with open(qrels_file) as f:
tsvreader = csv.DictReader(f, delimiter="\t")
for row in tsvreader:
qid = row["query-id"]
pid = row["corpus-id"]
rel = int(row["score"])
if qid in qrels:
qrels[qid][pid] = rel
else:
qrels[qid] = {pid: rel}
return qrels
corpus_ds = load_dataset("openbmb/VisRAG-Ret-Test-SlideVQA", name="corpus", split="train")
queries_ds = load_dataset("openbmb/VisRAG-Ret-Test-SlideVQA", name="queries", split="train")
qrels_path = "xxxx" # path to qrels file which can be found under qrels folder in the repo.
qrels = load_beir_qrels(qrels_path)
```
# 数据集说明
本数据集是基于[SlideVQA](https://arxiv.org/abs/2301.04883)数据集所提供的幻灯片集构建的视觉问答（Visual Question Answering, VQA）数据集。
## 数据集加载方式
python
from datasets import load_dataset
import csv
def load_beir_qrels(qrels_file):
qrels = {}
with open(qrels_file) as f:
tsvreader = csv.DictReader(f, delimiter=" ")
for row in tsvreader:
qid = row["query-id"]
pid = row["corpus-id"]
rel = int(row["score"])
if qid in qrels:
qrels[qid][pid] = rel
else:
qrels[qid] = {pid: rel}
return qrels
corpus_ds = load_dataset("openbmb/VisRAG-Ret-Test-SlideVQA", name="corpus", split="train")
queries_ds = load_dataset("openbmb/VisRAG-Ret-Test-SlideVQA", name="queries", split="train")
qrels_path = "xxxx" # 可在本仓库的qrels文件夹下获取该qrels文件的路径
qrels = load_beir_qrels(qrels_path)
提供机构：
maas
创建时间：
2025-05-15
