ShahJahan-del/imdb-sentiment-analysis

DistilBERT Fine-Tuned on IMDb for Sentiment Analysis

This model is a fine-tuned version of distilbert-base-uncased on the IMDb movie reviews dataset. It was developed during Week 3 of the AI Internship to perform binary sentiment classification (Positive/Negative).

Model Description

• Developed by: Shah Jahan SHEIKH (AI Intern)
• Model Type: Text Classification (Sentiment Analysis)
• Language: English
• Base Model: distilbert-base-uncased

Training Results

The model was trained on a downsampled subset of 1,000 training examples and evaluated on 200 test examples over 2 epochs.

• Final Test Accuracy: 74.50%
• Training Loss: 0.4456
• Validation Loss: 0.5367

How to Use

You can easily load and use this model for inference using the Hugging Face pipeline:

from transformers import pipeline

# Load the model from the Hub
classifier = pipeline("text-classification", model="ShahJahan-del/imdb-sentiment-analysis")

# Test with custom reviews
reviews = [
    "This movie was an absolute masterpiece, the acting was phenomenal!",
    "I fell asleep after 20 minutes. Total waste of time and money."
]

results = classifier(reviews)
for review, res in zip(reviews, results):
    sentiment = "POSITIVE" if res['label'] == "LABEL_1" else "NEGATIVE"
    print(f"Review: '{review}' -> {sentiment} ({res['score']:.2%})")