FarSSiM: Persian Sentence Similarity Dataset

This repository hosts the Persian Sentence Similarity dataset, originally created and maintained by Mojtaba Sajjadi under the FarSSiM repository. All credits for the creation, curation, and publication of this dataset go to Mojtaba Sajjadi.

Dataset Overview

FarSSiM is a high-quality Persian dataset focused on sentence similarity tasks. The dataset contains pairs of Persian sentences along with similarity scores, making it suitable for tasks such as:

• Semantic textual similarity (STS)
• Natural language inference (NLI)
• Paraphrase detection

This dataset can be a valuable resource for researchers and practitioners working on Persian NLP or multilingual NLP tasks.

Dataset Structure

Each sample in the dataset consists of:

• sentence1: The first sentence in Persian.
• sentence2: The second sentence in Persian.
• similarity_score: A numeric score representing the similarity between the two sentences.

The similarity scores are typically normalized between 0 and 1, where 1 indicates high similarity and 0 indicates no similarity.

File Structure

The dataset is provided in CSV format and includes the following columns:

Column Name	Description
sentence1	The first Persian sentence.
sentence2	The second Persian sentence.
similarity_score	Similarity score between sentence1 and sentence2.

Installation and Usage

This dataset is available on the Hugging Face Hub and can be loaded using the datasets library:

from datasets import load_dataset

# Load the FarSSiM dataset
farssim = load_dataset("AlirezaF138/FarSSiM")

# Access the training split
print(farssim["train"][0])

Citation

If you use this dataset in your research, please cite the original repository:

@misc{sajjadi2023farssim,
  author = {Mojtaba Sajjadi},
  title = {FarSSiM: Persian Sentence Similarity Dataset},
  year = {2023},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/mojtabasajjadi/FarSSiM}}
}

License

Please refer to the original repository for licensing information. Ensure compliance with the original license terms when using this dataset.

Credits

This dataset was originally created and maintained by Mojtaba Sajjadi. This repository simply rehosts the dataset for ease of access and integration into the Hugging Face ecosystem.