ai-ank / ainimalgo-chat-with-animals Goto Github PK

Welcome to "AInimal Go!", an innovative application inspired by the charm of Pokémon Go, crafted for animal lovers and technology aficionados. It showcases capabilities of MultiModal Retriever-Augmented Generation (RAG) in creating a delightful user experience. This unique app bridges the gap between the digital world and the animal kingdom, allowing users to upload or capture images of animals and engage in interactive, text-based conversations with them, powered by advanced AI technologies.

• Image Upload & Capture: Seamlessly upload an image from your gallery or use the in-app camera to snap a picture of an animal.
• ResNet Image Classification: Utilizes the ResNet model for accurate identification of various animal species from the uploaded images.
• AI-Driven Conversations: Engage in text-based chats with animals through a role-playing Large Language Model (LLM), bringing a personalized and imaginative touch to each interaction.
• User-Friendly Interface: Intuitive design and simple navigation enhance the user experience, making animal interaction accessible to everyone.

1. Capture or Upload: Begin by capturing a new image or uploading an existing one of an animal.
2. Animal Identification: The app uses the ResNet model to analyze the image and identify the animal species.
3. Engage in Conversation: The identified animal species is then used to initiate a conversation. Our LLM, equipped with conversational capabilities, role-plays as the animal, allowing for a unique chatting experience.
4. Interactive Chat Interface: Users can ask questions or chat freely with the AI-represented animal, receiving responses that are imaginative, informative, and engaging.

• LlamaIndex: Orchestrates the integration of LLMs and deep learning models, streamlining the conversational AI process.
• ResNet Model: Chosen for its speed and accuracy in image classification, it's pivotal for the initial animal identification step.
• Cohere LLM API: Powers the core of the conversational AI, enabling the app to simulate realistic and knowledgeable interactions with animals.
• Wikipedia: The app retrieves factual information from Wikipedia, enriching the conversations with educational content about the animals.
• Streamlit: Provides the framework for the app's user interface, offering an accessible and pleasant user experience.
• Hugging Face Spaces: The hosting platform that supports the app's deployment, ensuring it's readily accessible to users worldwide.

Try out the live demo here: AInimal GO!

This project is licensed under the MIT License.

"AInimal Go!" is a conceptual prototype aimed at demonstrating the potential of AI in creating interactive experiences. It's a playful and educational tool, not intended for scientific or research purposes.

We welcome contributions, suggestions, and feedback to enhance the app. Feel free to fork, star, and contribute to this project!

Developed by Harshad Suryawanshi. Connect with me on LinkedIn, Medium.