Modelpublic

JerryJJJJJ/review-summarization-flan-t5

sourceHugging Faceupdated 6mo agoView on Hugging Face

0likes4downloads

Model Card

🟨 3. Review Summarization（FLAN-T5）

markdown

# Review Summarization (FLAN-T5)

## Overview
This model generates concise summaries from customer reviews.

It helps transform long reviews into short, meaningful insights.

The model is based on FLAN-T5 and fine-tuned for text summarization tasks.

## Model Details
- Base model: FLAN-T5
- Task: Text Summarization (Text2Text Generation)

## Dataset
Dataset used:
- Amazon Polarity Dataset

A subset of reviews was used for training summarization.

## Evaluation Results
| Model   | ROUGE-1 | ROUGE-2 | ROUGE-L |
|--------|---------|---------|---------|
| FLAN-T5 | 0.106   | 0.021   | 0.096   |

## Usage

from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

model_name = "JerryJJJJJ/review-summarization-flan-t5"

tokenizer = AutoTokenizer.frompretrained(modelname) model = AutoModelForSeq2SeqLM.frompretrained(modelname)

text = "The phone has great performance but poor battery life."

inputs = tokenizer(text, returntensors="pt", truncation=True) outputs = model.generate(**inputs, maxlength=30)

summary = tokenizer.decode(outputs[0], skipspecialtokens=True)

print(summary)