maheshsmc/image-caption-decoder

🖼️ Image Caption Generator with Decoding Strategies

This project is a Hugging Face + GitHub demo of an Image Caption Generator using the nlpconnect/vit-gpt2-image-captioning model.

🚀 Features

• —Upload an image and generate captions
• —Supports multiple decoding strategies:
• —Greedy
• —Beam Search
• —Top-k Sampling
• —Top-p (Nucleus) Sampling

📦 Installation

pip install -r requirements.txt

▶️ Run Locally

streamlit run app.py

🌐 Deployment

• —Hugging Face Spaces (Streamlit): Upload all files
• —GitHub: Push repo for tracking + sharing

📖 Example

Greedy: a man riding a horse in the field Beam (5): a person on a horse riding in a grassy field Top-k: a rider sits proudly on a horse in the countryside Top-p: someone is on horseback, enjoying the open field