Conversational Fashion Outfit Generator
- Developed Conversational Fashion Outfit Generator with Gen AI, finetuned using social
media trends and user data.
- Integrated web crawlers to source data from Instagram, Pinterest, and Flipkart for outfit
recommendations.
- Utilized pre-trained SDXL model for accurate fashion recommendations based on
individual style and preferences.
- GitHub Link
Adversarial Attack on CNN Image Classification
- Utilizing a Convolutional Neural Network (CNN) with 32 dense layers.
- Activation function: ReLU.
- Employing the MNIST dataset to evaluate the impact of the FGSM attack.
- Model accuracy after attack: 86%, indicating confusion.
- GitHub Link
Twitter Sentiment Analysis
- Real-time extraction of Twitter data for analysis.
- Utilizing PySpark to efficiently handle large datasets.
- Employing SVM for sentiment classification.
- GitHub Link