Views
No views yet
1# Review Summarization (FLAN-T5)
2
3## Overview
4This model generates concise summaries from customer reviews.
5
6It helps transform long reviews into short, meaningful insights.
7
8The model is based on FLAN-T5 and fine-tuned for text summarization tasks.
9
10## Model Details
11- Base model: FLAN-T5
12- Task: Text Summarization (Text2Text Generation)
13
14## Dataset
15Dataset used:
16- Amazon Polarity Dataset
17
18A subset of reviews was used for training summarization.
19
20## Evaluation Results
21| Model | ROUGE-1 | ROUGE-2 | ROUGE-L |
22|--------|---------|---------|---------|
23| FLAN-T5 | 0.106 | 0.021 | 0.096 |
24
25## Usage
26
27```python
28from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
29
30model_name = "JerryJJJJJ/review-summarization-flan-t5"
31
32tokenizer = AutoTokenizer.from_pretrained(model_name)
33model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
34
35text = "The phone has great performance but poor battery life."
36
37inputs = tokenizer(text, return_tensors="pt", truncation=True)
38outputs = model.generate(**inputs, max_length=30)
39
40summary = tokenizer.decode(outputs[0], skip_special_tokens=True)
41
42print(summary)