Pytorch Docs Refactored
Createdseanmichaelmcgee
Provides a command-line interface for semantic search within PyTorch documentation, leveraging embeddings and a vector database.
About
Pytorch Docs Refactored is a semantic search prototype designed to improve the discoverability of information within the PyTorch documentation. Currently paused for redesign, the tool offers a basic command-line interface that allows users to query the documentation using natural language. It integrates ChromaDB for storing and querying embeddings, differentiates between code and text content, and offers an interactive mode for continuous queries. The project aims to enhance relevance quality, expand content coverage, and refine the chunking strategy for improved search results.
Key Features
- Distinguishes between code and text content
- Command-line interface for querying PyTorch documentation
- Interactive mode for continuous queries
- ChromaDB integration for storing and querying embeddings
- 1 GitHub stars
