Chroma Cloud Splade

Chroma provides a convenient wrapper around Chroma Cloud’s Splade sparse embedding API. This embedding function runs remotely on Chroma Cloud’s servers, and requires a Chroma API key. You can get an API key by signing up for an account at Chroma Cloud. Sparse embeddings are useful for retrieval tasks where you want to match on specific keywords or terms, rather than semantic similarity.

Python
TypeScript

This embedding function relies on the httpx python package, which you can install with pip install httpx.

from chromadb.utils.embedding_functions import ChromaCloudSpladeEmbeddingFunction, ChromaCloudSpladeEmbeddingModel
import os

os.environ["CHROMA_API_KEY"] = "YOUR_API_KEY"
splade_ef = ChromaCloudSpladeEmbeddingFunction(
    model=ChromaCloudSpladeEmbeddingModel.SPLADE_PP_EN_V1
)

texts = ["Hello, world!", "How are you?"]
sparse_embeddings = splade_ef(texts)

You can optionally pass in a model argument. By default, Chroma uses prithivida/Splade_PP_en_v1.

// npm install @chroma-core/chroma-cloud-splade

import { ChromaCloudSpladeEmbeddingFunction, ChromaCloudSpladeEmbeddingModel } from "@chroma-core/chroma-cloud-splade";

const embedder = new ChromaCloudSpladeEmbeddingFunction({
  apiKeyEnvVar: "CHROMA_API_KEY", // Or set CHROMA_API_KEY env var
  model: ChromaCloudSpladeEmbeddingModel.SPLADE_PP_EN_V1,
});

// use directly
const sparseEmbeddings = await embedder.generate(["document1", "document2"]);

Visit Chroma Cloud documentation for more information on available models and configuration.

Dense Embedding Models

Sparse Embedding Models

Frameworks