README.md

Welcome to my repository 🤗

Technologies Used:

• Web Application: Built with Django
• Databases: Utilized Django ORM
• AI Integration: Leveraged mistralai/Mistral-7B-Instruct-v0.1 | mistralai/Mixtral-8x7B-Instruct-v0.1 HuggingFace 🤗 models
• Embeddings Model: Employed the intfloat/e5-large-v2 HuggingFace 🤗 model

Guide:

• Dataset:

  • The Original Dataset was initially scraped and transformed into TextDataset and VectorDataset models using custom commands.

• Backend:

  • The backend encompasses various functions.

• Architecture:

  • The architectural concept follows this flow:
    • Incoming message ➔
    • Simplification by retaining only keywords ➔
    • Structuring and vectorization of the simplified incoming message ➔
    • Retrieval of top-k most cosine-similar embeddings from VectorDataset ➔
    • Acquisition of the IDs of these top-k VectorDataset objects, enabling retrieval of corresponding text from TextDataset ➔
    • Feeding the text context to an OPENAI API to obtain responses.