Datasets:
To have QAs belonging to a subject in a separate subset
Hi there!
Thank you for this dataset! I propose to put QAs of each subject e.g. شیمی or دینی into a separate subset. This way, evaluation per subject would be easier.
Hi!
Thanks for your suggestion.
We avoided adding various category labels to the data to minimize redundancy. You can obtain these categories by using our code. Different mappings are available in our code(github.com/raia-center/khayyam-challenge/blob/main/all_dict.py), and you can perform this conversion using the provided code.
from all_dict import *
import pandas as pd
df = pd.read_csv('khayyam_challenge.csv')
df['Year_en'] = df['Year'].map(years_map)
df['final_category_en'] = df['final_category_fa'].map(final_category_en)
df['Education Period_en'] = df['Education Period'].map(edu_map)
df['Level_en'] = df['Level'].map(level_map)
df['unleveled_final_category'] = df['final_category_fa'].map(unleveled_final_category)
df['group_unleveledـfinal_category'] = df['unleveled_final_category'].map(group_unleveledـfinal_category)
df['group_unleveledـfinal_category_en'] = df['group_unleveledـfinal_category'].map(group_unleveledـfinal_category_en)
df['final_category_en'] = df['final_category_fa'].map(final_category_en)
df['Education Period_en'] = df['Education Period'].map(edu_map)
regards.
I think the original question was asking about a HuggingFace dataset (using datasets library). HuggingFace datasets provide many benefits over CSV+Pandas and integrate easily with the rest of the platform. One of these benefits are the easy subset creation which was asked in the original question here.
You may read the documentation for more information about how to convert your CSV data into HuggingFace Dataset format.
https://huggingface.co./docs/datasets/en/create_dataset