mt5-billsum-summarizer

mT5 Congressional Bill Summarizer

This model is a fine-tuned version of google/mt5-small trained on the BillSum dataset. It is designed to take complex, lengthy US Congressional Bills and legal text, and condense them into highly readable 2-3 sentence summaries.

Model Details

• Base Model: google/mt5-small
• Task: Text Summarization (Sequence-to-Sequence)
• Dataset: BillSum (~22,000 U.S. Congressional Bills)
• Language: English

Intended Use

This model is specifically trained to handle the dense, specialized vocabulary of legal and government documents. It is best used for quickly extracting the core legislative intent from long political texts.

How to Use

You can easily use this model in your own applications via the Hugging Face transformers library:

from transformers import pipeline

# 1. Initialize the summarization pipeline
summarizer = pipeline("summarization", model="Ramsha-Anwar/mt5-billsum-summarizer")

# 2. Paste your legal text
bill_text = """
To amend the Internal Revenue Code of 1986 to provide a tax credit for 
expenses for household and elder care services necessary for gainful employment.
"""

# 3. Generate the summary
summary = summarizer(bill_text, max_length=128, min_length=30)
print(summary[0]['summary_text'])