jorses
commited on
Commit
ยท
1d252ef
1
Parent(s):
9d54dd4
Update README.md
Browse files
README.md
CHANGED
@@ -685,6 +685,32 @@ answering systems on tabular data, they are not large and diverse enough to eval
|
|
685 |
To this end, we provide a corpus of 65 real world datasets, with 3,269,975 and 1615 columns in total, and 1300 questions to evaluate your models for the task of QA over Tabular Data.
|
686 |
|
687 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
688 |
## ๐ Datasets
|
689 |
By clicking on each name in the table below, you will be able to explore each dataset.
|
690 |
|
@@ -797,30 +823,4 @@ If you use this resource, please use the following reference:
|
|
797 |
year = "2024",
|
798 |
address = "Turin, Italy"
|
799 |
}
|
800 |
-
```
|
801 |
-
|
802 |
-
# Usage
|
803 |
-
|
804 |
-
```python
|
805 |
-
from datasets import load_dataset
|
806 |
-
|
807 |
-
# Load all QA pairs
|
808 |
-
all_qa = load_dataset("cardiffnlp/databench", name="qa", split="full")
|
809 |
-
|
810 |
-
# Load SemEval 2025 task 8 Question-Answer splits
|
811 |
-
semeval_train_qa = load_dataset("cardiffnlp/databench", name="semeval", split="train")
|
812 |
-
semeval_dev_qa = load_dataset("cardiffnlp/databench", name="semeval", split="dev")
|
813 |
-
|
814 |
-
|
815 |
-
# is "001_Forbes", the id of the dataset where information to answer the Question is located
|
816 |
-
all_qa['dataset'][0]
|
817 |
-
|
818 |
-
# This id can be used load a specific Question-Answer pair collection from the splits
|
819 |
-
forbes_qa = load_dataset("cardiffnlp/databench", name="qa", split=all_qa['dataset'][0] )
|
820 |
-
|
821 |
-
# you can load a specific dataset containg the "answer" for a QA pair using the
|
822 |
-
forbes_full = load_dataset("cardiffnlp/databench", name=all_qa['dataset'][0] , split="full")
|
823 |
-
|
824 |
-
# or to load the databench lite equivalent dataset, to answer the "sample_answer"
|
825 |
-
forbes_sample = load_dataset("cardiffnlp/databench", name=all_qa['dataset'][0] , split="lite")
|
826 |
```
|
|
|
685 |
To this end, we provide a corpus of 65 real world datasets, with 3,269,975 and 1615 columns in total, and 1300 questions to evaluate your models for the task of QA over Tabular Data.
|
686 |
|
687 |
|
688 |
+
## Usage
|
689 |
+
|
690 |
+
```python
|
691 |
+
from datasets import load_dataset
|
692 |
+
|
693 |
+
# Load all QA pairs
|
694 |
+
all_qa = load_dataset("cardiffnlp/databench", name="qa", split="full")
|
695 |
+
|
696 |
+
# Load SemEval 2025 task 8 Question-Answer splits
|
697 |
+
semeval_train_qa = load_dataset("cardiffnlp/databench", name="semeval", split="train")
|
698 |
+
semeval_dev_qa = load_dataset("cardiffnlp/databench", name="semeval", split="dev")
|
699 |
+
|
700 |
+
|
701 |
+
# "001_Forbes", the id of the dataset where information to answer the Question is located
|
702 |
+
all_qa['dataset'][0]
|
703 |
+
|
704 |
+
# This id can be used load a specific Question-Answer pair collection from the splits
|
705 |
+
forbes_qa = load_dataset("cardiffnlp/databench", name="qa", split=all_qa['dataset'][0] )
|
706 |
+
|
707 |
+
# you can load a specific dataset containg the "answer" for a QA pair using the
|
708 |
+
forbes_full = load_dataset("cardiffnlp/databench", name=all_qa['dataset'][0] , split="full")
|
709 |
+
|
710 |
+
# or to load the databench lite equivalent dataset, to answer the "sample_answer"
|
711 |
+
forbes_sample = load_dataset("cardiffnlp/databench", name=all_qa['dataset'][0] , split="lite")
|
712 |
+
```
|
713 |
+
|
714 |
## ๐ Datasets
|
715 |
By clicking on each name in the table below, you will be able to explore each dataset.
|
716 |
|
|
|
823 |
year = "2024",
|
824 |
address = "Turin, Italy"
|
825 |
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
826 |
```
|