From the course: Level up LLM applications development with LangChain and OpenAI
Unlock the full course today
Join today to access over 24,700 courses taught by industry experts.
Initialize a vector store (Chroma) and ingest documents
From the course: Level up LLM applications development with LangChain and OpenAI
Initialize a vector store (Chroma) and ingest documents
- [Instructor] So after loading and splitting the documents into chunks, the next step is to create and store the vectors that are called embeddings. So, we are exactly at this stage. After indexing the documents, we move on to the creation of the embeddings. And so, embeddings are like numerical representations of words. So, there are a list of floating-point numbers that are used to measure the distance and relatedness between two vectors. So, let's see how they look like. So, you see that this is a list of floating-point numbers. And so the smaller the distance between two vectors suggests a high relatedness between the retrieved documents and the user query. And on the opposite, large distances between vectors suggest low relatedness. And so the OpenAI text embeddings model is used to measure the relatedness of text strings to allow the similarity search from a vector database. And as a vector database, we're going to use Chroma. Chroma DB, which is an AI-native open-source vector…
Contents
-
-
-
-
-
(Locked)
RAG: Overview and architecture2m 12s
-
(Locked)
Breaking down the RAG pipeline2m 50s
-
(Locked)
Project setup3m 33s
-
(Locked)
Load and split documents into chunks5m 6s
-
(Locked)
Initialize a vector store (Chroma) and ingest documents5m 6s
-
(Locked)
Create the chain: Prompt + model + parser5m 39s
-
(Locked)
Create the chain: Add context with a retriever4m 48s
-
(Locked)
Pass data with RunnablePassthrough and query data3m 27s
-
Challenge: Create a custom agent with history3m 12s
-
Solution: Add a chain with chat history5m 19s
-
(Locked)
Solution: Context- and history-aware chatbot5m 49s
-
(Locked)
-
-
-
-
-
-
-
-