Retrieval Langchain Plugin - Combining the chatGPT plugin and langchain to bootstrap the Question Answering (Retreival based mechanisms) in just a template away!
PROJECT - Langchain Retrieval Template
Brainstorm:
A comprehensive guide for creating a Question Answering system should include information on the following:
-
Data Model: Different methods of loading documents such as S3, EveryNote, and Website should be discussed, along with the required metadata for each method.
-
Vector Database: The guide should provide details on various vector databases, including Pinecone, Weaviate, Zilliz, Milvus, Qdrant, and Redis,
-
Embeddings: The documentation should cover the various options available for embeddings such as Cohere, OpenAI, and Hugging Face Hosted LLM.
-
Docker Container: The guide should explain the usage of Fast API or Flask for building the container.
-
API Authentication: The documentation should outline the different methods available for API authentication.
-
Cloud Deployment: The guide should provide instructions for deploying the backend API in the cloud.
-
Utility Services: Preprocessing and post-processing scripts should be discussed in the guide.
-
Testing: The guide should cover the testing phase.
-
API Swagger: For each template, the guide should include API Swagger documentation.
-
Memory Management: The documentation should provide information on how to store and retrieve conversations.
-
Personalization: The guide should cover personalization and customization in prompts, prompt templates
-
CI/CD Pipeline: The documentation should cover the CI/CD pipeline. Webhooks: The guide should discuss webhooks for external integration with messaging platforms such as Slack, as well as an Iframe embed approach for website integration.
Authors: Sangeetha Venkatesan, Misbah Syed