saharM/bart-base-finetuned-samsum

BART Base Fine-Tuned on SAMSum for Dialogue Summarization

This model is a fine-tuned version of facebook/bart-base trained on the SAMSum dataset for the task of dialogue summarization.

---

Model Details

• Architecture: BART-base (Sequence-to-Sequence Transformer)
• Fine-tuned on: SAMSum dataset (dialogue conversations with human-written summaries)
• Use case: Summarizing chat/dialogue conversations into concise summaries

---

Usage

from transformers import BartForConditionalGeneration, BartTokenizer

model_name = "your-username/bart-finetuned-samsum"  # replace with your model repo name

tokenizer = BartTokenizer.from_pretrained(model_name)
model = BartForConditionalGeneration.from_pretrained(model_name)

dialogue = """
Speaker 1: Hey, are you coming to the party tonight?
Speaker 2: I’m not sure yet, maybe. What time does it start?
Speaker 1: Around 8 PM. Let me know!
"""

inputs = tokenizer(dialogue, return_tensors="pt")
summary_ids = model.generate(**inputs)
summary = tokenizer.decode(summary_ids[0], skip_special_tokens=True)

print("Summary:", summary)

Dataset

The SAMSum dataset is a collection of messenger-like conversations with corresponding summaries, designed to train and evaluate dialogue summarization models.

• Dataset link

---

License

This model and code are licensed under the MIT License.

---

Citation

If you use this model in your research, please cite:

@inproceedings{gliwa2019samsum,
  title={SAMSum Corpus: A Human-Annotated Dialogue Dataset for Abstractive Summarization},
  author={Gliwa, Bogdan and Wójcik, Tomasz and Biega, Agnieszka and Marasek, Konrad},
  booktitle={Proceedings of the 2nd Workshop on New Frontiers in Summarization},
  year={2019}
}