VertexAI Embeddings

This component in Nappai uses Google’s Vertex AI to transform your text into numerical codes called “embeddings.” Think of it like translating words into a secret code that computers can easily understand and compare. This allows Nappai to perform tasks like finding similar documents or organizing information based on meaning.

Relationship with Google Vertex AI

This component directly interacts with Google’s Vertex AI service to generate these embeddings. It requires you to provide access credentials so Nappai can securely connect to your Google Cloud project.

Inputs

Credentials: A JSON file containing your Google Cloud credentials. Nappai will try to use your environment settings if you leave this blank. (Advanced users only need to change this.)
Location: The geographic location of the Google Cloud service (e.g., “us-central1”). Leave this as the default unless you have a specific reason to change it. (Advanced)
Project: Your Google Cloud project ID. Leave this as the default unless you have a specific reason to change it. (Advanced)
Model Name: The specific Vertex AI model to use for generating embeddings. The default is usually fine.
Max Output Tokens: The maximum number of tokens the model can process. (Advanced)
Max Retries: The number of times Nappai will try again if there’s a problem. (Advanced)
N: The number of embeddings to generate. (Advanced)
Request Parallelism: How many requests Nappai sends to Google Cloud at once. (Advanced)
Stop: Sequences of words that tell the model when to stop generating embeddings. (Advanced)
Streaming: Whether to use streaming for embedding generation. (Advanced)
Temperature: Controls how random the embeddings are. A lower value (default 0.0) makes them more consistent.
Top K: The number of tokens to consider when generating embeddings. (Advanced)
Top P: The cumulative probability to consider when generating embeddings. (Advanced)

Outputs

The component produces a set of embeddings. These are numerical representations of your text data that are then used by other Nappai components, such as vector databases, to perform tasks like semantic search and similarity analysis.

Usage Example

Imagine you have a collection of customer reviews. You can use this component to generate embeddings for each review. Then, you can use a vector database component (like Pinecone or Weaviate) to store and search these embeddings. This allows you to quickly find reviews that are similar in meaning, even if they don’t use the exact same words.

Templates

[List of templates where the component can be seen and its configuration - This section will be populated based on the actual templates used.]

Semantic Text Splitter: Breaks down large texts into smaller, meaningful chunks before embedding.
Couchbase, Upstash, Chroma DB, Weaviate, Vectara, Redis, PGVector, FAISS, Astra DB, Qdrant, Pinecone, MongoDB Atlas, Milvus, Supabase, Cassandra: These are vector databases that store and search the embeddings generated by this component.
Text Embedder: Another embedding component (potentially using a different model).

Tips and Best Practices

Start with the default settings. Only adjust advanced settings if you have a specific need and understand their impact.
Ensure you have properly configured your Google Cloud credentials.
For large datasets, consider using the semantic text splitter to improve performance.

Security Considerations

Protect your Google Cloud credentials. Do not share your credentials file.
Use strong passwords and follow Google Cloud’s security best practices.