This application combines comprehensive identification features of a multimodal input app with the quick and broad description handling capabilities of a Retrieval-Augmented Generation (RAG) system. The goal is to create a versatile and efficient Pokémon Identification App that can handle text, image, and combined queries, providing detailed and accurate information about Pokémon.

• Pokémon Dataset by Rounak Banik
• Pokémon Image Dataset by hlrhegemony

1. Image Dataset: Contains images and names of Pokémon.
2. General Info Dataset: Contains names, types, HP levels, attack levels, defense levels, and detailed descriptions.

1. CLIP Model: For embedding text and image inputs.
2. BERT/GPT Model: For embedding and handling broad text descriptions.
3. Retrieval-Augmented Generation (RAG): To enhance broad description handling and provide relevant information.

1. Relational Database: Stores Pokémon metadata, image URLs, and detailed descriptions.
2. Vector Database: Stores embeddings for efficient similarity searches.

1. Data Ingestion: Processes and stores data and embeddings.
2. Search API: Handles text, image, and combined queries to return matching Pokémon data.
3. RAG Integration: Enhances search capabilities by handling broad descriptions.

1. Query Input: Allows users to input text descriptions or upload images.
2. Results Display: Shows the Pokémon image and general information.

• raw/
• processed/
• combined/

• app/
  • init.py
  • main.py
  • api/
    • init.py
    • endpoints.py
  • models/
    • init.py
    • clip_model.py
    • bert_model.py
  • database/
    • init.py
    • setup.py
    • queries.py
• requirements.txt
• README.md

• public/
• src/
  • components/
    • SearchBar.js
    • ResultsDisplay.js
  • App.js
  • index.js
• package.json
• README.md

• download_images.py
• preprocess_images.py
• preprocess_csv.py
• combine_datasets.py