urt5-base-finetuned

Summarization Model (Type:T5)

Summarization: Extractive and Abstractive

• urT5 adapted from mT5 having monolingual vocabulary only; 40k tokens of Urdu.
  • Fine-tuned on https://huggingface.co/mbshr/XSUMUrdu-DW_BBC, ref to https://doi.org/10.48550/arXiv.2310.02790 for details.

Model Description

<!-- Provide a longer summary of what this model is. -->

• Model type: urT5 adapted version of mT5
• Language(s) (NLP): Urdu
• Finetuned from model: google/mt5-base

Model Sources

<!-- Provide the basic links for the model. -->

• Repository: [More Information Needed]
• Paper: https://doi.org/10.48550/arXiv.2310.02790

Uses

Summarization

Evaluation & Results

<!-- This section describes the evaluation protocols and provides the results. -->

Evaluated on https://huggingface.co/mbshr/XSUMUrdu-DW_BBC

• ROUGE-1 F Score: 40.03 combined, 46.35 BBC Urdu datapoints only and 36.91 DW Urdu datapoints only)
• BERTScore: 75.1 combined, 77.0 BBC Urdu datapoints only and 74.16 DW Urdu datapoints only

Citation [optional]

<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->

https://doi.org/10.48550/arXiv.2310.02790

Contact

• [email protected]