SuganyaP/quick-distilbert-imdb

DistilBERT IMDB Sentiment Classifier

Overview

This repository contains a fine-tuned DistilBERT model for binary sentiment classification on the IMDB movie reviews dataset. The model predicts whether a given review expresses positive or negative sentiment. It is intended as a lightweight, reproducible NLP model suitable for demonstrations, small-scale applications, and experimentation.

Base Model

• Model: distilbert-base-uncased
• Framework: Hugging Face Transformers
• Task: Text Classification (Binary Sentiment)

Training Details

• Dataset: IMDB movie review dataset (train/test split)
• Objective: Binary sentiment classification
• Optimization:
  • Adam optimizer
  • Learning rate scheduling
  • Early stopping
• Regularization:
  • Dropout applied as per DistilBERT architecture
  • Gradient clipping

Evaluation Metrics

The model was evaluated using standard binary classification metrics:

• Accuracy
• Precision
• Recall
• F1-score

Inference Example

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

model_name = "SuganyaP/quick-distilbert-imdb"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

inputs = tokenizer("This movie was excellent!", return_tensors="pt")
outputs = model(**inputs)
prediction = torch.argmax(outputs.logits).item()

print("Positive" if prediction == 1 else "Negative")