Back to Module Dashboard
Multimodal Classification: Zero-shot vs. Few-shot
🏆 Project Summaries & Achievements
Author: Nguyen Viet Hoang
Full Report: Report on Multimodal Classification.pdf
Abstract
Apply the **CLIP** model with classification head added to accurately classify image-text pairs belong to a class. Compare zero-shot performance against few-shot fine-tuning.
Key Achievements
- Write here ........................................................
- And here ........................................................
Methodology and Technical Highlights
- Write here ........................................................
- And here ........................................................
Limitations
- Write here ........................................................
- And here ........................................................
Live Demo
Live Inference App
Note: The model might take a few seconds to warm up on the cloud server.
📊 Exploratory Data Analysis (EDA)
Exploratory Data Analysis on (Write here)...
Data Distribution
Covarriance matrix
<<<<<<< HEAD