push nomic and add sample output

[Feliren88]

Image Duplicate Finder using Nomic Vision Embeddings

This PR adds a standalone script for detecting duplicate or highly similar images within a directory. The implementation:

• Uses the powerful Nomic Vision embedding model (nomic-embed-vision-v1) to generate high-quality image representations
• Implements efficient pairwise similarity comparison with PyTorch
• Provides GPU acceleration for faster processing
• Includes a configurable similarity threshold to control matching strictness
• Outputs results in a clearly formatted CSV file

Key improvements:

• Single folder processing - finds duplicates within any given directory
• Memory-efficient design - stores embeddings on CPU while processing
• Progress tracking with tqdm
• Comprehensive documentation in README.md
• Error handling for corrupted or problematic images