Distilbert Base Uncased Helpful Amazon

🚀 Review Helpfulness Prediction Model

This model uses data from a specific source to predict the helpfulness of product reviews, fine - tuned from the distilbert - base - uncased model.

🚀 Quick Start

This model was trained with data from https://registry.opendata.aws/helpful-sentences-from-reviews/ to predict how "helpful" a review is. It was fine - tuned from the distilbert-base-uncased model.

Labels

• LABEL_0 - Not helpful
• LABEL_1 - Helpful

💻 Usage Examples

Basic Usage

from transformers import (
    AutoTokenizer,
    AutoModelForSequenceClassification,
    TextClassificationPipeline,
)

tokenizer = AutoTokenizer.from_pretrained("banjtheman/distilbert-base-uncased-helpful-amazon")

model = AutoModelForSequenceClassification.from_pretrained(
    "banjtheman/distilbert-base-uncased-helpful-amazon"
)

pipe = TextClassificationPipeline(model=model, tokenizer=tokenizer)

result = pipe("This was a Christmas gift for my grandson.")

print(result)

#[{'label': 'LABEL_0', 'score': 0.998775064945221}]
# This is NOT A HELPFUL comment

📄 License

This project is licensed under the Apache - 2.0 license.