README.md

metadata

title: IMDB Sentiment Classifier
emoji: 🎬
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 4.0.0
app_file: app.py
pinned: false
license: mit

🎬 IMDB Sentiment Classifier

A binary sentiment classifier fine-tuned on the IMDB movie reviews dataset using LoRA (Low-Rank Adaptation) with DistilBERT (≤ 110M parameters).

🔍 About the Model

• Model Type: Binary Sentiment Classification
• Base Model: distilbert-base-uncased
• Training Dataset: IMDB Dataset (50K movie reviews)
• Task: Classify text as Positive or Negative sentiment
• Fine-tuning Method: LoRA (Low-Rank Adaptation)
• Parameters: ~66M total, ~890K trainable with LoRA

🚀 Features

• Single Review Analysis: Enter any movie review text
• Batch Processing: Analyze multiple reviews at once
• File Upload: Upload .txt or .csv files for analysis
• Confidence Scores: See how confident the model is
• Visual Feedback: Color-coded results with emojis

📊 Example Results

Review	Sentiment	Confidence
"This movie was absolutely fantastic!"	Positive 😊	98%
"Terrible film, waste of time."	Negative 😞	95%
"Average movie, nothing special."	Negative 😞	72%

🛠️ Technical Details

Model Architecture

Model: DistilBERT + LoRA
- Base: distilbert-base-uncased
- LoRA Rank: 8
- Trainable Parameters: ~890K (1.35% of total)
- Max Sequence Length: 256 tokens