metadata
language: en
license: mit
tags:
- t5
- t5-small
- summarization
- text-summarization
- pytorch
- transformers
pipeline_tag: summarization
library_name: transformers
T5 Dialogue Summarizer
A fine-tuned T5-small model for text and dialogue summarization.
Model Details
- Base model: T5-small
- Task: Text summarization
- Framework: PyTorch
- Tokenizer: T5Tokenizer (max_length: 512)
- Decoding: Beam search (num_beams=4, max_length=150, early_stopping=True)
Usage
Using Pipeline
from transformers import pipeline
summarizer = pipeline("summarization", model="unnat17/t5-dialogue-summarizer")
result = summarizer("Your text here...")
print(result[0]["summary_text"])
Direct Loading
from transformers import T5ForConditionalGeneration, T5Tokenizer
model = T5ForConditionalGeneration.from_pretrained("unnat17/t5-dialogue-summarizer")
tokenizer = T5Tokenizer.from_pretrained("unnat17/t5-dialogue-summarizer")
input_text = "summarize: " + "Your text here..."
inputs = tokenizer(input_text, return_tensors="pt", max_length=512, truncation=True)
output = model.generate(**inputs, num_beams=4, max_length=150, early_stopping=True)
print(tokenizer.decode(output[0], skip_special_tokens=True))
Web Application
A full-stack web application using this model is available at: github.com/unnat-git/Text-Summarizer