Dataset Viewer
The dataset viewer is not available for this dataset.
Unexpected token '<', "<html>
<h"... is not valid JSON
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Dataset Card: Jordanian-Dialect-Instruct-QA
Overview
This dataset contains 1500 academic Q&A pairs for Jordanian universities, written in the Jordanian Arabic dialect. It is used to fine-tune models to provide natural, localized responses to academic inquiries and more.
This dataset was used to fine-tune Qwen2.5-7B-Instruct-Jordanian. You can visit the model page to download the GGUF weights and the custom Ollama Modelfile.
Dataset Structure
Each sample contains:
instruction: The query in Jordanian dialect.
context: The formal university policy.
response: The friendly, dialect-specific answer.
Licensing
This dataset is released under the Apache-2.0 license.
- Downloads last month
- 14