Dive into a live code review session where experts build a semantic search engine using Typescript. Explore fundamental concepts of semantic search, compare it with lexical search, and delve deep into the codebase of Pinecone's semantic search Typescript example. Learn about embeddings, CSV loaders, Transformers.js, and the indexing process. Discover techniques for batch embedding and parallelizing embeddings using workers. Follow along as the speakers demonstrate querying the index and provide insights on implementing semantic search effectively.
Overview
Syllabus
- Intro
- Basic concepts of lexical search
- What are embeddings?
- How does semantic search work?
- Comparing semantic and lexical search
- Build session intro, CSV loader
- CSV loader done, short review of utilities
- Introducing Transformers.js
- Short look at why types can be helpful
- Embedding in batches
- Ideas around using workers to parallelize embeddings
- A client helper function
- The main portion of the indexing process
- Where to find the repo
- Querying the index
Taught by
Pinecone
Related Courses
-
Supercharging Semantic Search with Pinecone and Cohere
-
The Magic of Multilingual Search with Pinecone Serverless and Inference
-
OpenAI's New GPT 3.5 Embedding Model for Semantic Search
-
NER Powered Semantic Search in Python
-
Understanding Generative AI: RAG, Semantic Search, Embeddings, and Vectors
-
Cohere AI's LLM for Semantic Search in Python
