Goals:
- Create an interactive chatbot that roleplays as an ancient Greek priestess, drawing inspiration and knowledge from A.R. Ammons' poetry and writings.
- Implement a Retrieval-Augmented Generation (RAG) system to provide context-aware responses based on the content of Ammons' works, filtered through the perspective of an ancient Greek priestess.
- Utilize Streamlit for the user interface, integrating it with Hugging Face's Transformers library (version 4.29.2) for NLP tasks, and FAISS (version 1.7.4) for efficient similarity search.
- Implement the system to run in a Python 3.10 environment, using PyTorch (version 2.0.1) as the deep learning framework, with all dependencies specified in a requirements.txt file.
- Design the application to operate within the constraints of a Hugging Face Space, ensuring compatibility with their serverless environment and adhering to their file system limitations for data persistence.
rag2/ │ ├── app.py ├── requirements.txt ├── ammons_muse.txt └── data/ ├── chunks_{model_combination}.pkl ├── embeddings_{model_combination}.npy └── faiss_index_{model_combination}.bin